Cathie Wood

As of 2026-07-22, Cathie Wood remains ARK Investment Management's founder, chief executive officer, chief investment officer, controlling owner, and primary portfolio manager. ARK's March 2026 Form ADV says she has final accountability for investment selection and approval. This chapter therefore treats the philosophy as Wood-led but organizational: analysts, theme developers, portfolio staff, external specialists, and public feedback contribute research; neither an ARK forecast nor an ETF result is automatically Wood's personal work or return (ARK Form ADV, 2026, pp. 4–5, 20–22, 94–95).

Guiding Research Questions

1. What does Wood mean by disruptive innovation, and what mechanism should turn technological progress into shareholder value?
2. Why does ARK believe public markets misprice that progress?
3. How do broad technology maps become company models, entries, position weights, and exits?
4. Which disclosed controls constrain conviction, concentration, liquidity, and forecast error?
5. In which regimes is the process likely to thrive or fail?
6. Which apparent contradictions are genuine, and which reflect the difference between a five-year thesis and active trading around it?
7. What evidence supports repeatable skill rather than growth-factor exposure, favorable timing, or a few outliers?

Core Worldview

Wood's worldview begins with disruptive innovation as an economic force, not a sector label. ARK defines a qualifying technology as one capable of major cost decline and elastic demand, reach across sectors and geographies, and further innovation built on top of it. The 2026 research map groups 13 ideas across artificial intelligence, robotics, energy storage, public blockchains, space, and biology, emphasizing convergence among platforms rather than isolated products. That is ARK's institutional opportunity map, not evidence that every idea belongs in every fund (ARK thematic process, pp. 2–3; Big Ideas 2026, pp. 3–4).

The causal chain is cost decline -> adoption -> expanding unit demand -> shifting industry economics -> company revenue and cash flow. Wright's Law and related learning curves help ARK estimate how cumulative production can reduce unit cost; demand elasticity estimates how lower prices may enlarge the market. Convergence matters because one platform can accelerate another—for example, artificial intelligence can improve robotics or drug discovery. The investment problem is not merely to identify a useful technology. It is to determine which companies capture value after competition, financing, dilution, regulation, and changing margins (ARK investment process).

Time horizon is the second pillar. ARK says investors overreact to near-term events while underestimating multi-year change, and it uses five-year company models to create distance from quarterly benchmark pressure. Wood argues that backward-looking indexes naturally emphasize incumbents and that conventional sector teams miss cross-industry convergence. This is a claimed organizational and horizon edge, not proof of alpha: it succeeds only when ARK forecasts adoption, market share, capital needs, and terminal economics better than the market (Wood, 2021; ARK thematic process, pp. 2–4).

The third pillar is transparency. ARK publishes research, valuation models, holdings, and many trade notifications in part to invite criticism and build investor trust. But transparency is incomplete by design: the trade files are unofficial and unreconciled and exclude IPOs, secondary offerings, and creation/redemption activity. Public research can also change without notice, while ARK may trade differently from a published view. Transparency therefore supplies testable inputs, not a complete contemporaneous audit trail (ARK trade-notification disclosure; ARK Form ADV, 2026, pp. 9–10).

The Edge - What Markets Misprice and Why

ARK's claimed edge has four components.

First, linear expectations confront exponential-looking cost curves. Analysts and incumbents can extrapolate current volumes or profits while learning effects lower price and unlock new demand. ARK tries to model the path before conventional earnings screens recognize it. The risk is that technical improvement may benefit customers rather than shareholders, or that competing producers commoditize the economics (ARK thematic process, pp. 2–4).

Second, organizational silos miss convergence. A sector analyst can understand today's industry yet underweight how software, batteries, automation, and biology alter one another. ARK organizes researchers around technologies and uses internal and external theme developers, academics, company dialogue, social media, and crowdsourced criticism to widen the evidence set (ARK investment process; ARK ESG policy, pp. 2–3).

Third, benchmarks reward established capitalization. ARKK's process starts from a distilled innovation universe rather than an index. That freedom can find emerging winners before they become benchmark weights, but it creates large tracking error and removes the benchmark as a portfolio-construction brake. A broad market can rise while ARK's opportunity set falls (Wood, 2021; ARK investment process).

Fourth, volatility can separate price from a five-year thesis. When short-term sentiment weakens but company milestones remain intact, a lower price raises model-implied return and may justify adding. The danger is circularity: the same internal thesis determines whether the price decline is opportunity or evidence. Without frozen milestones, “living research” can turn falsification into target revision (TIME, 2022; ARK FAQ).

Process: Idea Sourcing to Sell Discipline

1. Idea Sourcing

Top-down work asks where the world is going, identifies technological platforms, estimates cost declines and demand elasticity, and maps the value chain that may be transformed. This produces a focused research universe rather than a list derived from benchmark membership. Analysts then look for public companies with relevant exposure across market capitalizations and geographies. Big Ideas is best read as a research agenda; it is not an ARKK holdings document (ARK investment process; Big Ideas 2026, pp. 3–4).

The “open research ecosystem” is meant to counter institutional groupthink. Theme developers and analysts publish assumptions, engage domain experts, follow company and academic work, and absorb outside responses. Collaboration is organized across technologies because convergence is part of the thesis. The benefit is broad challenge; the risks are promotional feedback, selective attention, and reliance on sources with economic interests in adoption (ARK investment process; ARK ESG policy, pp. 2–3).

2. Research and Thesis Formation

Bottom-up work uses investment briefs and company models. ARK's disclosed portfolio tracker evaluates six areas: company, people, and culture; execution; moat or barriers; product leadership; valuation; and thesis risk. Warning signs include management departures, weak research and development, lost market share, worsening competition, regulatory or geopolitical exposure, slower adoption, and governance or legal problems. Scores are reviewed in weekly meetings and monitored continuously (ARK thematic process, pp. 3–4).

The scoring system makes the method more explicit, but the public record does not disclose the complete score-to-weight equation, the historical score archive, or how often Wood overrides the signal. ARK's FAQ says a proprietary score at 6/10 or below triggers full review; that is a diagnostic threshold, not an automatic sale (ARK FAQ). Wood's final authority means this is an informed central-decision system, not an independent risk committee veto.

3. Valuation and Entry

ARK builds five-year revenue and valuation models using unit volumes, cost decline, adoption or penetration, market share, share-count growth, margins, and future multiples. A disclosed methodology flag is an expected average stock return below 15% over five years [single-source official methodology]; ARK's ESG document describes that hurdle as roughly a doubling over five years. It is a score-review input, not a contractual fund-return target or a universal automatic entry/exit rule (ARK thematic process, pp. 3–4; ARK ESG policy, p. 6).

The 2024 Tesla model illustrates both ambition and model risk. Named analysts ran one million Monte Carlo simulations across 45 inputs and published bear, base, and bull outcomes. Their limitations section acknowledges subjective inputs, a positive-model bias, data problems, and ARK's incentive to present favorable results. Simulation can expose sensitivity; it cannot make uncertain adoption, autonomy, regulation, or competitive capture objective (ARK Tesla 2029 model, 2024).

Entry is therefore conditional on model-implied return and conviction rather than one public multiple. Lower prices can create an entry or increase a weight if operational evidence holds. No current public standard initial weight, fixed stop-loss, hard issuer-ownership ceiling, or complete factor-risk budget was recovered in the reviewed method and fund documents. Readers should not invent one from a particular historical holding (ARK thematic process, pp. 3–4; ARK ETF Trust prospectus, 2025, ARKK pp. 2–12).

4. Sizing and Portfolio Construction

ARK describes conviction weighting: expected reward relative to price, company scores, thesis risks, and competing opportunities inform position size. Analysts and portfolio staff contribute, but Wood approves final selections. For ARKK specifically, the 2025 prospectus describes an active, non-diversified ETF seeking long-term capital growth, normally investing at least 65% in domestic or foreign equities relevant to disruptive innovation and expecting roughly 40–50 issuers [single-source fund terms; ranges are not caps]. The fund can invest across company sizes and developed or emerging markets (ARK ETF Trust prospectus, 2025, ARKK pp. 2–12).

Those labels do not ensure economic diversification. On 2026-01-31 ARKK reported 47 holdings, $6.677 billion of net assets, and 20% turnover for the six-month period; the turnover figure should not be mechanically annualized. Its sector buckets spanned healthcare, information technology, consumer discretionary, financials, communications, and industrials, yet many holdings shared long-duration growth, financing, and risk-appetite exposure (SEC semiannual report, 2026).

Official returns show endpoint sensitivity. Through 2025-07-31, ARKK reported +65.31% for one year, -0.83% annualized for five years, and +14.45% annualized for ten years, versus +16.33%, +15.88%, and +13.66% for the S&P 500 comparator [fund returns, not Wood-personal; S&P 500 is not a tracked index]. The coexistence of a strong rebound, weak five-year result, and modest ten-year advantage is a regime and path lesson, not a one-number verdict (SEC annual report, 2025).

5. Ongoing Management and Sell Discipline

Weekly meetings test the thesis and position score. ARK may trim a liquid holding to fund an idea with greater modeled upside, reduce a weight when metrics no longer support it, sell when the thesis changes, or exit when a company no longer leads its innovation category. It may add during short-term negative sentiment if the operating case is unchanged. These are clear qualitative triggers, but public materials do not disclose a universal price stop or maximum portfolio drawdown (ARK investment process).

Five-year modeling is therefore compatible with active trading around a thesis. ARKK's fiscal 2025 portfolio turnover was 43% [single-period fund figure]; daily trade notices further reveal frequent rebalancing. The tension is real but not necessarily contradictory: the company outcome may be long horizon while relative weights respond daily to price and opportunity. Implementation skill and transaction costs are part of the result (ARK ETF Trust prospectus, 2025, pp. 2–3; ARK trade-notification disclosure).

Risk Management

Wood distinguishes temporary uncertainty from permanent impairment and often treats volatility as an opportunity to increase conviction. ARK's operational controls are thesis monitoring, six-part company scores, valuation updates, comparative opportunity ranking, weekly review, and trading around price. Wood has also told advisers that ARK should not be the core of a portfolio, a useful boundary for a volatile specialist allocation (Morningstar interview, 2025).

The prospectus supplies the harder risk vocabulary. ARKK is non-diversified and can concentrate loss in a limited number of issuers; disruptive technologies may fail commercially; early-stage companies may lack revenue or financing; micro- and small-cap securities may be difficult to sell; regulatory, political, adoption, obsolescence, issuer, market, and management risks can destroy value. Authorized Participants and market makers can withdraw in stress, widening spreads and price/NAV gaps. A large shareholder can impair the fund through a large transaction (ARK ETF Trust prospectus, 2025, pp. 5–12, 87, 92–95).

ETF mechanics need precision. Secondary-market shareholders trade with one another; only Authorized Participants create or redeem large units with the fund. In-kind redemptions can transfer securities instead of forcing one-for-one portfolio sales, reducing dilution and taxable gains. Cash baskets, transaction costs, reduced scale, AP retreat, and later liquidation of received securities still matter. Capacity depends on ARK's ownership and overlap across vehicles and the liquidity of underlying companies—not simply ARKK's exchange volume (ARK ETF Trust prospectus, 2025, pp. 87, 92–95, 117–119).

Morningstar's 2021 process review questioned aggregate risk controls, analyst depth, ownership stakes, capacity, and key-person dependence. Those staffing and policy observations are historical, not confirmed 2026 facts. Its April 2026 review says the team had stabilized and recent hires were constructive, while continuing to view the portfolio as one correlated technology/growth exposure and succession as unproven (Morningstar Australia, 2021; Morningstar research report, 2026, pp. 1–6).

Temperament and Psychology

The philosophy requires curiosity, tolerance for visible tracking error, willingness to look beyond current profits, and enough patience to let adoption compound. Wood uses the five-year horizon and public research as behavioral anchors. During corrections, she says ARK concentrates into its highest-conviction names rather than moving toward a benchmark. That can exploit short-term mispricing, but it also demands client patience precisely when correlations, redemptions, and liquidity can worsen (TIME, 2022).

The central psychological danger is conviction becoming immunity to evidence. A falling price raises model-implied return mechanically; only operating milestones can distinguish opportunity from a damaged thesis. A disciplined implementation should freeze base/bull/bear assumptions, adoption deadlines, dilution and financing limits, competitive-share evidence, and the condition that forces a thesis change. Seven-year investor-horizon guidance cannot excuse a matured five-year model that failed (ARK FAQ).

Evolution Over Her Career

ARK's 2014 founding combined disruptive-innovation research, cross-sector organization, active ETFs, and unusually public work. Wood later described openness as a response to trust lost after the 2008 crisis. The 2020–2021 risk-on and pandemic-adoption regime strengthened the argument that innovation deserved a strategic allocation; the prospectus's calendar series shows +152.51% in 2020, -23.36% in 2021, and -66.99% in 2022 [fund returns; one official series]. The swing made regime exposure and client timing inseparable from the philosophy's public reception (ARK ETF Trust prospectus, 2025, p. 12).

In December 2021 Wood published a 30–40% five-year compound-return forecast for ARK's broad innovation strategies [manager forecast; not particular-fund performance]; the article was revised to clarify that boundary. In 2023 she defended the process after the rate-driven repricing by emphasizing transparency, deflationary technology, and continuing operational progress. These are dated statements, not evidence that the forecast matured successfully (Wood, 2021, footnote 3; Wood, 2023).

Later research became more explicit about scenario distributions and model limitations, as the 2024 Tesla Monte Carlo work illustrates. By 2025 Wood described shifting AI emphasis from hardware toward software and embodied AI. In early 2026, ARK's own commentary reported ARKK down 12.22% in the first quarter versus roughly -4% for the S&P 500 and -6% for Nasdaq [manager-reported quarter and comparators], showing that a 2025 rebound did not eliminate short-horizon regime sensitivity (ARK Q1 2026 commentary).

What She Explicitly Rejects

Wood rejects benchmark-relative portfolio construction, sector-siloed research, a primary focus on current earnings, and treating all volatility as permanent risk. She rejects the idea that disruptive innovation belongs only in private markets. ARK's ESG policy does not use mechanical sector exclusion; it integrates environmental, social, governance, legal, and reputational considerations into the broader score. She does not reject valuation, fundamental company work, diversification among innovation platforms, or tactical trading (ARK ESG policy, pp. 2–6).

Regimes Where the Philosophy Thrives or Struggles

This matrix is a synthesis of ARK's models, prospectus risks, Wood's observed risk-off behavior, and full-cycle results—not a tested forecast for the next cycle (ARK thematic process, pp. 2–4; ARK ETF Trust prospectus, 2025, pp. 5–12; TIME, 2022; Morningstar research report, 2026, pp. 4–5).

The evidence is consistent with these exposures. ARK acknowledges that its extraordinary performance was partly associated with unusually favorable conditions and may not recur. Morningstar's current report places ARKK's October 2014–March 2026 annualized return at 12.3%, versus 12.9% for the style-relevant Russell Midcap Pure Growth Index and 12.7% for the S&P 500 [single independent methodology]. That comparison does not prove no skill, but it demands a style benchmark rather than an absolute-return story (Morningstar research report, 2026, pp. 4–5).

Tensions Between Stated Philosophy and Actual Behavior

1. Five-year horizon versus active turnover. Five-year company models coexist with weekly review, daily trading visibility, and 43% fiscal-year turnover. This can be rational reweighting around durable theses, but it makes execution and trading costs part of the edge (ARK thematic process, pp. 3–4; ARK ETF Trust prospectus, 2025, p. 3; ARK trade-notification disclosure).

2. A 15% stock hurdle versus realized fund outcomes. The model trigger is forward-looking and company-level; ARKK's negative five-year return through 2025 does not violate a guarantee that never existed. It does show why forecast calibration must be scored against realized portfolios (ARK thematic process, pp. 3–4; SEC annual report, 2025).

3. Thematic breadth versus economic concentration. Forty-seven reported holdings across several sectors still expressed one correlated long-duration technology/growth exposure in Morningstar's current assessment; name count does not ensure factor diversification (SEC semiannual report, 2026; Morningstar research report, 2026, pp. 1–6).

4. Open inputs versus centralized authority. Public feedback and collaborative research widen the evidence, but Wood's final approval and approximately 72% ownership make key-person and succession risk structural (ARK Form ADV, 2026, pp. 94–95).

5. Published models versus mutable living research. Transparency makes some assumptions testable, but changing targets and disclosed presentation incentives permit selective framing. A complete forecast ledger is more informative than a showcase of winners (ARK Form ADV, 2026, pp. 9–10; ARK Tesla 2029 model, 2024).

6. Volatility opportunity versus loss of principal. Wood's language emphasizes uncertainty; the prospectus treats non-diversification, issuer failure, liquidity, and market collapse as risks. Both can be true, but only capital at risk decides whether the investor survives the distinction (Morningstar interview, 2025; ARK ETF Trust prospectus, 2025, pp. 5–12).

7. Tax-efficient wrapper versus frictionless-story risk. In-kind redemptions improve tax and trading mechanics, yet turnover, cash baskets, spreads, investor sales, and underlying liquidity remain costs (ARK ETF Trust prospectus, 2025, pp. 87, 92–95, 117–119; ARK trade-notification disclosure).

8. Innovation insight versus price discipline. The broad direction of AI, robotics, genomics, or blockchain may be right while a specific company or entry price fails. Bessembinder's long-run evidence that net wealth creation is concentrated in a small minority of stocks supports both ARK's search for exceptional winners and the difficulty of identifying them prospectively (Bessembinder, 2018).

The 2021 Tesla 2025 forecast supplies a matured calibration case. ARK published pre-split bear/base/bull targets of $1,500/$3,000/$4,000. Tesla's 2022 three-for-one split makes those $500/$1,000/about $1,333; the stock ended 2025 at $449.72, below even the adjusted bear target [single market-data endpoint]. This is one miss, not a complete forecast scorecard, and open 2026/2027 targets should not be judged early (ARK Tesla 2025 model, 2021; Tesla split release, 2022; Yahoo Finance chart endpoint).

Criticism, Skill Versus Luck, and Transferability

Evidence consistent with skill includes early recognition of important technology platforms, explicit cost/adoption modeling, willingness to publish assumptions, a long-tenured cross-sector organization, and strong fund results at selected endpoints (Big Ideas 2026, pp. 3–4; ARK Form ADV, 2026, pp. 20–22; SEC annual report, 2025). Evidence consistent with factor exposure, luck, or non-repeatability includes extreme drawdowns, middling style-adjusted comparison, correlated holdings, a matured major forecast miss, and mutable targets (ARK ETF Trust prospectus, 2025, p. 12; Morningstar research report, 2026, pp. 1–6; Tesla forecast calibration).

Investor timing is a separate failure channel. Reuters reported Morningstar's estimate that ARK's US ETF family destroyed $14.3 billion of investor wealth over the ten years through 2023, including $7.1 billion attributed to ARKK [single independent estimate; dollar-weighted aggregate, not NAV return or Wood-personal P&L]. Late inflows after spectacular returns and exits after drawdowns can create poor investor experience even when a fund's time-weighted record is positive (Reuters, 2024).

Academic evidence is a prior, not an ARKK verdict. Ben-David and coauthors find that specialized ETFs on average suffered large risk-adjusted losses after launch, consistent with attention and extrapolation; ARKK requires its own holdings- and factor-level test (NBER, 2021). The evidence supports Wood's research agenda more strongly than persistent style-adjusted alpha. A fair future test needs frozen forecast vintages, company milestones, holdings-level contribution, liquidity costs, and full-cycle factor adjustment.

Transferable: cross-sector technology maps; explicit cost, adoption, dilution, and valuation assumptions; dated base/bull/bear cases; written disconfirmation milestones; comparative opportunity ranking; small innovation sleeves; style-consistent benchmarks; and a complete forecast ledger.

Not safely transferable: copying daily trades without thesis or portfolio context; concentrated early-stage ownership; treating a long horizon as permission to ignore failure; ARK's analyst network and execution; or assuming an individual can absorb ARKK-like drawdowns. ARK's Form CRS reports no disciplinary history, but this is adviser-filed and bounded; registration does not imply regulator approval or investment skill (ARK Form CRS, 2026).

Current and Legal Boundary

As of 2026-07-22, ARK's March 2026 ADV says neither the adviser nor its management persons has been subject to material legal or disciplinary action, and Wood's individual brochure answers “None” to its disciplinary-information item. Those are adviser-filed disclosures, not independent universal clearance; SEC registration does not mean the regulator approved ARK, verified its statements, or endorsed Wood's skill. Bounded lane searches found no matching public SEC enforcement action against Wood or ARK Investment Management, but cannot exclude sealed, private, unindexed, or later matters (ARK Form ADV, 2026, pp. 0–1, 70, 94–95).

The same ADV discloses a prospective conflict rather than a proven offense: affiliated ARK Capital Markets was seeking broker-dealer registration, and future fee-paying relationships with researched companies could create an incentive for favorable coverage. ARK says any such relationship will be disclosed. Underperformance, an aggressive forecast, an investee-company failure, or an exemptive regulatory order is not by itself illegality (ARK Form ADV, 2026, pp. 9–10).

Practical Takeaways

1. Separate “the technology wins” from “this company captures value at this price.”
2. Freeze adoption, dilution, margin, share, and terminal-multiple assumptions before the price moves.
3. Use a five-year model as an accountability schedule, not an ever-receding deadline.
4. Treat thematic diversification and factor diversification as different questions.
5. Size an innovation sleeve for the drawdown path, not the forecast's median outcome.
6. Compare results with a style-consistent benchmark and preserve every forecast vintage.
7. Treat ETF liquidity as the interaction of APs, baskets, underlying securities, ownership, and redemptions.
8. Attribute ARK, Wood, analysts, named funds, investors, and portfolio companies separately.

Evidence Standard and Ranking

This chapter ranks ARK investment campaigns, not Cathie Wood's personal-account trades. Wood has final investment authority, but analysts, portfolio staff, multiple ETFs, fund flows, and in-kind creation/redemption activity all affect the visible record. Regulatory schedules provide point-in-time shares, value, and sometimes cost or fund-period realized gains; they do not provide a complete tax-lot ledger. Consequently, an issuer's price change, a fund return, and an ARK campaign P&L are three different quantities (ARK Form ADV, 2026, pp. 4–5, 20–22).

Tesla is the single best documented campaign: ARK owned it within months of the firm's 2014 launch, added through 2019 weakness, identified it repeatedly as a leading contributor, and still held large positions in 2026. Bitcoin through GBTC is the strongest original-insight runner-up. The remaining ranks recognize genuine gains while preserving major givebacks, incomplete exits, and missing lifetime P&L (ARK EV research, 2014; SEC semiannual report, January 2019; SEC annual report, July 2019; SEC annual report, 2020; ETF.com, 2015; ARK ETF Trust portfolio schedule, April 2026).

1. Tesla, 2014–Open — Single Best Documented

Context, dates, thesis, and discovery. ARK published a Tesla thesis in September 2014, arguing that battery scale, low electric-vehicle penetration, and limited incumbent competition could support growth far beyond prevailing expectations. A 2015 case study documents the actual research network: James Wang, Brett Winton, Tasha Keeney, external theme developers, and Wood joined battery architecture, charging, direct distribution, storage, software, and autonomous mobility into one system. This was an organizational thesis approved by Wood, not a solo prediction (ARK EV research, 2014; ARK Tesla case study, 2015).

Size and structure. At January 31, 2015, ARKQ held 2,584 pre-split shares worth $526,102, or 6.2% of that fund, and ARKK held 1,355 worth $275,878, or 5.5% [single-source regulatory snapshot]. By January 31, 2019, ARKK held 383,558 shares worth $117.8 million and 8.7% of assets; by July 31 it held 792,438 worth $191.5 million and 11.1% [single-source regulatory snapshots]. The 107% net share-count increase is [derived from single-source regulatory endpoints]; subscriptions, redemptions, and other trading prevent calling it a sum of discretionary buys (SEC semiannual report, 2015; SEC semiannual report, January 2019; SEC annual report, July 2019).

Entry, path, and drawdown. The filing-implied Tesla price fell from about $307.02 at January 2019 to $241.60 at July 2019, a 21.3% endpoint decline [derived from single-source regulatory endpoints; not achieved return]; maximum intra-period drawdown was not recovered. ARK expanded its reported positions rather than abandoning the thesis. Tesla was then identified as a top contributor across ARKQ, ARKK, and ARKW in fiscal 2020 and again as ARKW's top contributor in fiscal 2021. At July 31, 2021, ARKK held 3,500,833 post-2020-split shares worth $2.406 billion, 10.7% of net assets [single-source regulatory snapshot] (SEC semiannual report, January 2019; SEC annual report, July 2019; SEC annual report, 2020; SEC annual report, 2021).

Exit and P&L. There is no full exit. On April 30, 2026, ARKK still held 1,653,653 shares worth $631.1 million, ARKQ held 543,841 worth $207.5 million, and ARKW held 376,625 worth $143.7 million [single-source regulatory snapshot]. No reviewed filing supplies lifetime purchase cost, realized gain, maximum position drawdown, or campaign return across funds; Tesla's contribution labels and ending values are not substitutes (ARK ETF Trust portfolio schedule, April 2026, pp. 2, 4, 8).

Later forecasting is important counterevidence. ARK's 2021 model produced pre-August-2022-split 2025 bear/base/bull values of $1,500/$3,000/$4,000, with roughly half the expected enterprise value tied to robotaxi operations [single-source manager forecast]. Tesla's later three-for-one split converts those to $500/$1,000/about $1,333; the stock closed 2025 at $449.72, below the adjusted bear case [single market-data endpoint]. That matured miss does not erase the original campaign, but it rejects a story of uniformly precise valuation (ARK Tesla 2025 model, 2021; Tesla split release, 2022; Yahoo Finance chart endpoint).

What it teaches. The repeatable part is research architecture: connect cost curves, product design, infrastructure, distribution, software, and adjacent markets, then add when the operating thesis survives a price decline. The non-repeatable part is ARK's research network and tolerance for concentration. Tesla also shows that getting the industry and company direction right can coexist with an overconfident terminal forecast (ARK Tesla case study, 2015; ARK Tesla 2025 model, 2021).

2. Bitcoin Through GBTC, Then ARKB, 2015–Open

Context, dates, thesis, and discovery. In September 2015, ARKW became the first ETF to obtain bitcoin exposure through publicly quoted Grayscale Bitcoin Trust shares. The entry followed Mt. Gox, Silk Road, and an approximately 51% twelve-month bitcoin decline [single-source contemporaneous report]. Wood and Brett Winton focused on network resilience, developer activity, adoption, and the possibility that a decentralized payment asset could reduce financial intermediation. The “first ETF wrapper” claim is independently contemporaneous; broader claims about being the first public asset manager remain issuer marketing (ETF.com, 2015).

Size and structure. By August 31, 2016, ARKW's GBTC lots had reported cost of $130,324 and value of $338,578, while ARKK reported cost of $88,879 and value of $196,062. Combined, that was $315,437 of unrealized appreciation on $219,203 of cost, or 143.9% [derived from a single-source regulatory endpoint; not campaign P&L]. At July 31, 2018, ARKQ, ARKK, and ARKW reported combined GBTC cost of $2.255 million and value of $5.751 million, a 155.0% unrealized gain [derived from a single-source regulatory endpoint; not campaign P&L] (SEC annual report, 2016; SEC annual report, 2018).

Entry, path, and drawdown. The original asset call was exceptional, but the fund vehicle path was not. At July 31, 2021, ARKW's GBTC position had $270.119 million of reported cost and $308.200 million of value; at July 2022, reported cost was $229.843 million against value of $88.877 million, a 61.3% unrealized loss; by July 2023, cost of $213.291 million stood against $105.658 million, a 50.5% loss [derived from single-source regulatory snapshots at each endpoint]. Changing lots and GBTC's premium/discount mean these endpoints cannot be spliced into one return (SEC annual report, 2021; SEC annual report, 2022; SEC annual report, 2023).

Exit and P&L. ARKW reportedly sold about two million remaining GBTC shares on December 28, 2023 and shifted roughly $92 million into bitcoin-futures ETF BITO while spot-ETF approval remained uncertain [single-source exit report]. It then migrated to affiliated spot ETF ARKB. In fiscal 2024, ARKW's ARKB purchases cost $101.660 million; sales proceeds were $7.833 million; realized gain on those disposals was $2.703 million; unrealized appreciation increased $57.429 million; and ending value was $153.960 million [single-source fund-period table]. Those ARKB figures are a new vehicle-period record, not the missing GBTC lifetime P&L (ETF.com, 2023; SEC annual report, 2024).

The thesis remains open: at April 30, 2026, ARKW held 3,775,474 ARKB shares worth $95.7 million and ARKF held 2,096,992 worth $53.2 million [single-source regulatory snapshot]. ARKB is affiliated and subadvised; it is not economically identical to old GBTC or interim BITO exposure (ARK ETF Trust portfolio schedule, April 2026, pp. 8, 10).

What it teaches. A contrarian technology insight and its investment wrapper must be underwritten separately. ARK recognized bitcoin early, but GBTC premium/discount, tax, regulatory, and transition risk made the vehicle outcome far less smooth than the asset narrative (ETF.com, 2015; SEC 2022; SEC 2023).

3. Square/Block, 2016–Open

Context, dates, thesis, and discovery. Square first appears in the reviewed August 2016 schedules: ARKK held 11,650 shares worth $142,013 and ARKW held 29,055 worth $354,180 [single-source regulatory snapshot]. ARK framed the company as a two-sided financial network: seller tools and payments could connect with Cash App, whose distribution, low acquisition cost, and peer-to-peer graph could support broader consumer finance and bitcoin services. No contemporaneous account of the initial discovery process or exact first fill was recovered (SEC annual report, 2016; ARK Square model, 2020).

Size and structure. The 2016 ARKK snapshot implies $12.19 per share. By July 31, 2020, ARKK held 3,777,708 shares worth about $490.5 million, implying $129.85—10.65 times the early issuer price [derived from single-source regulatory endpoints; not ARK return]. At July 2021, ARKK held 4,404,371 shares worth $1.089 billion and ARKW held 1,051,476 worth $260.0 million [single-source regulatory snapshot]. Square was an official contributor across several active funds in fiscal 2020 and again to ARKK and ARKW in 2021 (SEC annual report, 2020; SEC annual report, 2021).

Entry, path, drawdown, exit, and P&L. The July 2021 ARKK value implies $247.26 per share. By July 2024 the reported Block position implied $61.88, a 75.0% filing-to-filing decline [derived from single-source regulatory endpoints; not maximum drawdown or achieved return]. ARK continued trading and later reduced exposure. At April 30, 2026, ARKK held 1,411,163 shares worth $99.5 million, ARKW held 669,661 worth $47.2 million, and ARKF held 623,690 worth $44.0 million [single-source regulatory snapshot]. The campaign is open; no reviewed table provides its lifetime realized profit, absolute P&L, or internal rate of return (SEC annual report, 2021; SEC annual report, 2024; ARK ETF Trust portfolio schedule, April 2026, pp. 4, 8, 10).

What it teaches. ARK identified Cash App and ecosystem convergence early. The round trip shows the other half of a trade: good discovery does not preserve peak value, and an open position cannot be scored as though the 2021 mark were realized (ARK Square model, 2020; SEC 2021; SEC 2024).

4. CRISPR Therapeutics, 2018–Open

Context, dates, thesis, and discovery. CRISPR Therapeutics first appears in the reviewed July 2018 schedules: ARKG held 169,431 shares worth $8.09 million and ARKK held 297,684 worth $14.21 million [single-source regulatory snapshot]. ARK's genomics thesis emphasized CRISPR/Cas9 programmability, lower editing cost, and curative potential; late-2020 trial data increased its stated conviction that disease could be treated without unacceptable off-target editing. No contemporaneous account of how ARK originally discovered the 2018 position was recovered (SEC annual report, 2018; ARK genomics commentary, 2020).

Size, path, and drawdown. At July 31, 2021, ARKK held 4,876,394 shares worth $590.1 million and ARKG held 1,773,623 worth $214.6 million [single-source regulatory snapshot]. The ARKK affiliated-issuer table reports, for that fiscal year, $400.7 million of purchases, $389.4 million of sales proceeds, a $220.1 million net realized gain, a $9.1 million positive change in unrealized appreciation, and $590.1 million ending value [single-source fund-period table; not lifetime] (SEC annual report, 2021).

The scientific thesis achieved a concrete milestone when the FDA approved Casgevy in December 2023, the first FDA-approved therapy using CRISPR/Cas9. The stock path was much harsher: ARKK's filing-implied price was $121.02 in July 2021, versus about $52.33 in April 2026, a 56.8% endpoint decline [derived from single-source regulatory endpoints; not achieved return], while the position remained large (SEC annual report, 2021; FDA, 2023; ARK ETF Trust portfolio schedule, April 2026, pp. 4, 6).

Exit and P&L. The trade is open. At April 2026, ARKK held about 6.17 million shares worth $322.8 million and ARKG held about 2.10 million worth $110.1 million [single-source regulatory snapshot]. The 2021 realized gain is unusually strong evidence, but it cannot be called lifetime campaign P&L because later purchases, sales, open lots, and other vehicles remain; lifetime percentage return is also unknown (ARK ETF Trust portfolio schedule, April 2026, pp. 4, 6; SEC annual report, 2021).

What it teaches. Technology and security outcomes can diverge. ARK identified a platform that produced an approved medicine and realized a large fund-period gain, yet the security still suffered a severe endpoint decline and commercialization remained unfinished (SEC annual report, 2021; FDA, 2023; current schedule).

5. Nvidia, 2015–Exited and Re-entered

Context, dates, thesis, and discovery. Nvidia was already held across ARK's four original active ETFs by January 31, 2015. ARK viewed graphics processors as parallel-computing infrastructure for deep learning and autonomous systems, not merely gaming chips. ARKK then held 8,672 shares worth $166,546; ARKQ held 13,727 worth $263,627; and ARKW held 22,583 worth $433,707 [single-source regulatory snapshot]. No contemporaneous account of the original discovery process or exact first fill was recovered (SEC semiannual report, 2015; SEC annual report, 2018).

Size, entry, and path. ARKK's filing-implied price rose from $19.21 in January 2015 to $244.86 in July 2018, about 12.75 times, before Nvidia's later stock splits [derived from single-source regulatory endpoints; not campaign return]. Nvidia was a leading contributor in the 2016–18 reporting record, and ARK's holdings expanded materially. Neither those snapshots nor contribution labels reveal retained original lots or realized P&L (SEC semiannual report, 2015; SEC annual report, 2016; SEC annual report, 2018).

Drawdown, exit, and P&L. ARKK closed its Nvidia position in early January 2023 after reducing it during the semiconductor selloff, then missed most of that year's 161.6% rally [single-source independent retrospective]. This is the campaign's decisive weakness: ARK recognized the computing architecture early but exited the flagship just before generative AI made the thesis obvious. A later re-entry was open by April 2026, when ARKK held 385,802 shares worth $77.0 million and smaller positions appeared across ARKQ, ARKW, ARKF, and ARKX [single-source regulatory snapshot]. Lifetime realized P&L, lifetime percentage return, and exact maximum drawdown remain unavailable (Bloomberg Línea, 2023; ARK ETF Trust portfolio schedule, April 2026, pp. 2, 4, 8, 10, 12).

What it teaches. Theme discovery and sell discipline are separate skills. An investor can be early and directionally right yet surrender the most valuable phase by exiting before an approaching adoption inflection (SEC 2015; SEC 2018; Bloomberg Línea, 2023).

6. Palantir, Two Separate Rounds

Context, dates, thesis, and discovery boundary. At July 31, 2021, ARKK held 22,189,818 Palantir shares worth $481.7 million and ARKW held 4,576,908 worth $99.4 million [single-source regulatory snapshot]. No contemporaneous ARK thesis memo, discovery account, exact initial fill, or falsification rule was recovered from the reviewed corpus. Palantir was absent from the July 2022 schedules, so the first round had ended rather than continuously compounding (SEC annual report, 2021; SEC annual report, 2022).

Second entry, size, and path. By July 31, 2023, ARKK again held 5,000,660 shares worth $99.2 million. Palantir became ARKK's top positive contributor for fiscal 2025; at that endpoint ARKK held 2,305,393 shares worth $365.1 million [single-source regulatory snapshots and fund-period attribution]. The first exit and second entry must not be merged into one winning cost basis (SEC annual report, 2023; SEC annual report, 2025).

Drawdown, exit, and P&L. The first round was exited, but no reviewed table supplies its realized loss, entry price, maximum drawdown, or percentage return. The second round remained open on April 30, 2026: ARKK held $203.1 million, with additional positions in ARKW, ARKF, and ARKQ [single-source regulatory snapshot]. Its maximum drawdown and percentage return are unknown. “Top contributor” proves a favorable fund-period effect, not absolute or lifetime P&L (ARK ETF Trust portfolio schedule, April 2026, pp. 2, 4, 8, 10).

What it teaches. A sale need not permanently falsify a company thesis. Re-entry can produce a favorable fund-period result, but honest attribution requires scoring the losing or inconclusive first round separately from the successful second—and admitting when the public discovery record is missing (SEC 2021; SEC 2022; SEC 2023; SEC 2025).

Popular Claims That Fail the Full-Cycle Test

Roku fails the full-cycle ranking because ARKK's affiliated-issuer tables reported a $467.3 million realized loss in fiscal 2024 and a further $338.0 million in fiscal 2025—$805.3 million across those two fund periods alone [single-source fund-period figures at each endpoint; not lifetime] (SEC annual report, 2024; SEC annual report, 2025).

Zoom was a major fiscal-2022 detractor and was absent from the April 2026 schedules; lifetime P&L and the exact final exit were not recovered. Coinbase was present by July 2021, detracted in fiscal 2022, rebounded as a fiscal-2023 contributor, and remained open in April 2026. These are mistake or recovery cases, not stronger completed wins than the six ranked campaigns (SEC annual report, 2021; SEC annual report, 2022; SEC annual report, 2023; ARK ETF Trust portfolio schedule, April 2026).

Skill, Regime, and Legal Boundary

The campaign record supports genuine early thematic insight in electric vehicles, bitcoin, GPUs, digital payments, and gene editing. It is weaker evidence of repeatable sell discipline or style-adjusted alpha: Block and Roku surrendered large marks, Nvidia's flagship exit preceded the core AI rerating, and many positions shared long-duration growth exposure. Morningstar's April 2026 report places ARKK's October 2014–March 2026 annualized return at 12.3%, versus 12.9% for the Russell Midcap Pure Growth Index and 12.7% for the S&P 500 [single-source independent methodology]. A handful of great campaigns cannot by itself distinguish security selection from favorable factor regimes (SEC annual report, 2024; SEC annual report, 2025; Bloomberg Línea, 2023; Morningstar manager report, 2026, pp. 4–5).

As of 2026-07-22, ARK's adviser-filed Form ADV reports no material disciplinary action against the adviser or its management persons, and Wood's brochure reports none [single-source adviser disclosure]. Bounded research found no matching public enforcement action tied to these campaigns. That is not universal clearance, and investee-company disputes involving Tesla, Coinbase, Grayscale, or others are not Wood misconduct without evidence linking her or ARK (ARK Form ADV, 2026, pp. 70, 94–95).

What Remains Unknown

• Complete security-level tax lots, entries, exits, execution prices, and lifetime realized/unrealized P&L across every ARK vehicle.
• The portion of share-count changes caused by discretionary trading versus ETF creations, redemptions, and mandate differences.
• Wood-personal returns, which cannot be inferred from fund schedules, fund NAV, assets under management, or issuer returns.
• Maximum position-level drawdowns after accounting for additions, trims, splits, flows, and fees.
• A frozen complete forecast ledger and holdings-level factor attribution capable of separating security selection from growth, duration, and liquidity regimes.

As of: 2026-07-24

Research Framing and Attribution Boundary

Cathie Wood's mistake record has an unusual disclosure problem: ARK publishes research and its public ETFs file audited accounts, but neither source is a Wood-personal trading ledger. This chapter therefore separates six units that are often collapsed in commentary: Wood; ARK Investment Management; the ARK research team; ARKK; other ARK vehicles; and each fund investor. It also separates time-weighted fund return, dollar-weighted investor outcome, realized security loss, unrealized mark, assets under management, and opportunity cost.

The strongest evidence is the public-fund record. The weakest is Wood's pre-ARK personal attribution. No audited Wood-personal composite, complete tax-lot history, or continuous Jennison/Tupelo/AllianceBernstein record was recovered. Figures below are assigned to the exact fund and period that reported them; they are not Wood's personal P&L. The evidence still supports a clear diagnosis: ARK identified several genuine technological shifts, but valuation, duration, common-factor concentration, forecast calibration, and slow thesis invalidation made the path destructive for many late-arriving investors.

Loss and Error Ledger

The 2021–22 Regime Failure

ARKK's boom and collapse form the central capital-loss episode for the strategy and franchise. The fund gained 152.51% in calendar 2020, then lost 23.36% in 2021 and 66.99% in 2022. It rebounded 67.82% in 2023 and 8.36% in 2024, but the sequence matters: a 67.8% gain after a 67.0% loss still leaves the capital far below its starting point. These are official calendar NAV returns from one SEC table, not investor-dollar results [single official SEC return table] (SEC calendar-return table, 2025).

The audited fiscal-year accounting shows what drove the collapse. From July 31, 2021 to July 31, 2022, ARKK's net assets fell from $22.495 billion to $9.337 billion. The $13.159 billion decline consisted of a $12.901 billion decrease from operations, $123.8 million of net shareholder redemptions, and $134.0 million of distributions. Operations therefore explain about 98.0% of the net-asset decline [derived from one audited filing]. Gross creations and redemptions were both near $18 billion, but the net flow was small. This was overwhelmingly portfolio loss, not a forced net-redemption spiral (SEC annual report, 2022).

Recovery is endpoint-dependent. Through June 30, 2026, ARKK's official five-year annualized NAV return remained -9.05%, while the ten-year return was +16.16% and the since-inception return +13.63%. The S&P 500 returned +13.41%, +15.51%, and +13.97% over those respective periods [single issuer report]. Thus a competitive ten-year number coexisted with severe destruction for people who entered around the 2021 asset peak (ARK ETF Trust Q2 report, 2026). Morningstar's current review likewise treats ARKK as a correlated technology/growth exposure whose difficult exits and weak sell discipline are not captured by the number of portfolio names (Morningstar manager report, 2026).

Wood did make one unusually direct macro admission in June 2022: “We were wrong on one thing, and that was inflation being as sustained as it has been.” She attributed the error to supply chains and Russia's invasion, then argued that deflationary forces were building (CNBC, 2022). Her broader response was less a capitulation than a defense. In early 2022 she explained the prior year through pandemic acceleration, rotation, and a five-year horizon; ARK's 2023 review said markets had overlooked innovation amid macro fear (TIME, 2022; Wood, 2023). Her July 2024 letter acknowledged that macro conditions and some stock selections had challenged performance, but continued to frame the portfolio as deep value and defended concentration and the long horizon (Wood, 2024).

The distinction is important. Wood acknowledged a forecast error and disappointing stock selection, but the reviewed public record contains no equivalent of a formal postmortem that says a portfolio rule failed and was replaced. Persistence can be rational when prices diverge from improving fundamentals; it becomes dangerous when a five-year horizon postpones falsification.

Security-Level Failures

Teladoc: the clearest averaging-down and delayed-exit case

Teladoc is the strongest documented individual failure. In fiscal 2022, ARKK reported only an $11.2 million realized loss in Teladoc but a $1.148 billion adverse change in unrealized value. By fiscal 2024, the fund reported a $1.549 billion realized loss, with only a $27.9 million remnant at July 31; the company was absent from the July 2025 schedule [each period or endpoint is one audited filing]. These are bounded ARKK accounting periods, not a lifetime ARK-family loss (SEC annual report, 2022; SEC annual report, 2024).

ARK had defended the whole-person virtual-care thesis despite disappointing quarters, a lower EBITDA outlook, and rising customer-acquisition costs. The subsequent company record supplied harder disconfirmation: Teladoc's 2024 revenue fell 1%, and its $1.001 billion net loss included a $790 million BetterHelp goodwill impairment [single issuer report] (ARK, 2022; Teladoc, 2025). The failure was not that telemedicine lacked a future. It was that category growth, company execution, acquisition economics, and security value had been treated as one thesis.

Roku: adoption right, calibration wrong, campaign still open

Roku demonstrates why realized loss, open mark, and lifetime campaign result must remain separate. ARKK's audited affiliate tables show a $467.3 million fiscal-2024 realized loss and another $338.0 million in fiscal 2025. Yet the fund still held 5.917 million shares worth $557.1 million at July 31, 2025. Fiscal 2022 had also shown a $188.7 million realized gain alongside a $1.799 billion adverse change in unrealized value [each period or endpoint is one audited filing]. No public table supports one lifetime P&L number (SEC annual report, 2022; SEC annual report, 2024; SEC annual report, 2025).

ARK's 2022 model placed a $605 expected value on Roku in 2026, driven by connected-TV adoption and advertising economics [single ARK forecast]. Because 2026 was still incomplete at the research cutoff, this chapter does not score that target as a final miss. The bounded conclusion is that a credible secular migration did not protect the holding from large realized fund-period losses (ARK Roku model, 2022; SEC annual report, 2024; SEC annual report, 2025).

The failed-company cluster

Fiscal 2024 crystallized several more affiliated-security losses: Ginkgo Bioworks $510.3 million, Invitae $453.4 million, UiPath $270.9 million, and 2U $195.4 million. The audited table totals $3.499 billion of realized losses across its listed affiliated issuers [derived from one filing; not all ARK securities or funds] (SEC annual report, 2024).

The issuer outcomes show why these were not merely interest-rate marks. Invitae filed Chapter 11 on February 13, 2024. 2U announced a prepackaged Chapter 11 in July 2024. Ginkgo implemented a one-for-40 reverse split. In its Q2 2024 commentary, ARK itself acknowledged Ginkgo's revenue and earnings miss, lower guidance, and pricing-model change; it also acknowledged another UiPath guidance reduction, sales-execution issues, and a chief-executive transition (Invitae 8-K, 2024; 2U, 2024; Ginkgo 8-K, 2024; ARK Q2 commentary, 2024).

The shared failure was translation. A large theoretical market did not guarantee that a particular company could finance itself, acquire customers economically, maintain margins, or capture the platform's value. Averaging down increased exposure precisely when issuer-specific evidence deserved more weight than the size of the addressable market.

Errors of Omission and Forecast Calibration

Nvidia is the clearest omission. ARKK cut more than 750,000 shares in October 2022, held fewer than 39,000 by the end of November, and was out by mid-January 2023 before the defining generative-AI rally [single independent report]. A counterfactual dollar loss would be fiction without an assumed holding period, tax lots, and alternative trades. The fair diagnosis is narrower: ARK identified GPU acceleration early but shifted away from the company that captured an extraordinary share of the value. Wood's 2024 letter defended the move as a valuation and portfolio-concentration decision and emphasized software opportunities; it was not a formal admission of error (Bloomberg Línea, 2023; Wood, 2024).

The matured forecast record is more measurable. ARK's 2021 Tesla model projected a pre-split 2025 base case of $3,000 and bear case of $1,500. Tesla's later three-for-one split makes those $1,000 and $500. The stock's December 31, 2025 close was $449.72—10.1% below the adjusted bear case and 55.0% below the base [single market-data endpoint]. This is a forecast-calibration miss, not a finding that the long-running Tesla campaign lost money (ARK Tesla model, 2021; Tesla split release, 2022; Yahoo Finance chart endpoint).

Wood also forecast in 2020 that global electric-vehicle sales could reach 40 million units and roughly 45% share in 2025. The IEA reports just over 20 million and about 25% [single independent outcome source]. Adoption direction was right; units were roughly twice the outcome and share about 20 percentage points too high. The lesson is not to ignore technological S-curves. It is to freeze the base rate, forecast vintage, adoption path, company share, dilution, margins, and terminal value so a correct theme cannot conceal a poor magnitude estimate (Wood, 2020; IEA, 2026).

Pre-ARK Record: Evidence of Cyclicality, Not a Personal Scorecard

The pre-ARK gaps are themselves part of the mistake record. The recovered AllianceBernstein presentation identifies Wood as thematic chief investment officer but describes a portfolio oversight group supported by more than 200 researchers. Its Strategic Research Equity composite lost 16.48% in 2001, 23.39% in 2002, and 46.72% in 2008 after the presentation's maximum fee. It rebounded 40.36% in 2009 and 21.65% in 2010, yet its 2001–10 net annualized return was only 0.81%, below the S&P 500's 1.41% [single manager presentation; composite-specific verification not established] (AllianceBernstein, 2011).

ARK's current employment chronology places Wood at Tupelo from 1998 to 2001. A contemporaneous amended Tupelo Form 13F reported approximately $1.360 billion of listed long securities at March 31, 2000, with concentrated technology and internet exposure [single regulatory filing; incomplete long-only snapshot]. It was signed by co-founder Lulu Wang, omitted confidential holdings, and says nothing about return, cash, shorts, flows, or Wood's personal account. It cannot validate folklore about an “80%” loss. No Jennison mandate or Wood-attributable performance series was recovered. The defensible conclusion is that thematic cyclicality predates ARK, not that these records form one continuous Wood composite (ARK Form ADV, 2026; Tupelo Form 13F, 2000).

Governance Near-Death and Current Conflict Boundary

ARK's closest recovered business-continuity crisis was the 2020 Resolute dispute. Resolute had a minority stake and an option to acquire control. Its October notice of intent to exercise that option meant a change of control could assign and automatically terminate the funds' advisory agreements; replacement agreements would require board and shareholder approval. The December settlement extinguished the option, preserved Wood's majority control, and kept Resolute as a minority investor and distribution partner. This was a genuine governance near-death, not fund bankruptcy, portfolio insolvency, or a regulatory finding (SEC supplement, 2020; ARK/Resolute statement, 2020).

Key-person exposure remains. ARK's March 2026 Form ADV says Wood owns approximately 72% of the firm and retains final investment-selection authority [single adviser filing]. It also reports no material legal or disciplinary action involving ARK or its management persons; that is an adviser-filed, bounded statement, not universal regulator clearance (ARK Form ADV, 2026). Portfolio-company bankruptcies and litigation are not adviser misconduct.

The same ADV disclosed that affiliated ARK Capital Markets was seeking broker-dealer registration and might receive transaction-related compensation, restrict client opportunities, encounter material nonpublic information, or create incentives for favorable research. A May 2026 Eightco filing makes the conflict concrete but does not establish illegality: subject to regulatory requirements, the affiliate agreed to a five-year advisory arrangement with an annual fee tied to treasury assets, warrants, and capitalization bonuses; ARK Chief Futurist Brett Winton received separate cash and restricted-share compensation for a board-advisory role. No ARK-client holding overlap or completed broker-dealer registration was established in the reviewed record (Eightco 8-K, 2026; ARK Form ADV, 2026).

Behavioral Root Causes

These are evidence-based inferences, not diagnoses Wood has adopted:

1. Theme-to-security slippage. ARK could be right about telemedicine, streaming, synthetic biology, automation, or online education while the selected company failed to capture value (SEC annual report, 2024; Teladoc, 2025).
2. Exponential anchoring. Large top-down markets and steep cost curves created optimistic priors that company-specific evidence had to overcome (ARK Roku model, 2022; ARK Tesla model, 2021).
3. Macro attribution asymmetry. Inflation, rates, rotation, and benchmark concentration explained losses, while rebounds were more readily treated as thesis validation. Both explanations can contain truth, but the asymmetry delays disconfirmation (Wood, 2023; Wood, 2024).
4. Averaging down under model persistence. Falling prices raised model-implied upside. Without a separate execution and balance-sheet veto, the valuation model could reinforce rather than challenge conviction (TIME, 2022; Morningstar manager report, 2026).
5. Name diversification without factor diversification. Multiple technologies still shared long-duration growth, funding, liquidity, and risk-appetite exposure (Morningstar manager report, 2026).
6. A receding accountability horizon. Five-year models encouraged patience, but mutable assumptions and new target dates made it hard to distinguish a delayed thesis from a failed one (TIME, 2022; ARK Roku model, 2022).
7. Centralized override risk. Wood's final authority keeps the philosophy coherent, but it concentrates selection, sell, personnel, ownership, and succession risk (ARK Form ADV, 2026; Morningstar manager report, 2026).

Process Changes Made After

ARK did change its organization. A September 2022 restructuring created a Chief Futurist role, four research-director positions, five research-associate posts, and associate portfolio-manager responsibilities [single issuer announcement]. The current disclosed process uses weekly thesis monitoring and allows sales when the thesis changes, metrics or position size deteriorate, or another company becomes a better expression of the theme. Later risk hires and more hierarchy are also documented (ARK research reorganization, 2022; ARK investment process; Morningstar manager report, 2026).

The causal claim must stop there. No primary source tied the reorganization, score system, or risk hiring to Teladoc, Roku, Invitae, Ginkgo, UiPath, 2U, or the 2021–22 drawdown. No public fixed stop-loss, maximum-drawdown limit, complete factor budget, independent risk veto, CIO-override log, or named-security postmortem was recovered. Morningstar's April 2026 assessment says the additions are constructive but too new to prove better sizing, sell discipline, liquidity management, or succession. The changes are real; their efficacy and origin remain unresolved (Morningstar manager report, 2026).

What Cannot Be Known

• Public fund reports do not provide a continuous cross-fund, lot-level P&L for every security. Fiscal-period realized losses cannot be relabeled as lifetime campaign losses.
• AUM decline, investor wealth destruction, NAV return, and Wood's personal economics are different measures.
• Errors of omission such as Nvidia have no responsible dollar value without a frozen counterfactual.
• Jennison and Tupelo returns attributable to Wood remain unavailable; AllianceBernstein's record is a team composite.
• The internal score history, model-vintage ledger, risk overrides, and reasons for individual trades are incomplete.
• Adviser filings and bounded public searches cannot rule out private, sealed, unindexed, or later legal matters.

Takeaways for the Canon

This record shows why thematic foresight and security-level investment success must be tested separately. The failure channel was translation: selecting the company, paying the price, sizing the common risk, financing the path, updating the thesis, and helping investors survive the drawdown.

The most transferable repair is an accountability system ARK has not fully published: freeze every forecast vintage; separate industry adoption from company capture; predefine company, balance-sheet, and valuation disconfirmation milestones; keep an independent factor and liquidity veto; record every override; and compare dollar-weighted client outcomes with time-weighted fund results. A long horizon is useful only when it creates dated tests. Otherwise it can turn patience into an explanation that never expires.

How to Read This Corpus

Cathie Wood has a large public record, but it is not one undifferentiated voice. A Wood-bylined ARK letter is edited primary writing; institutional we describes ARK, not a Wood-personal account. A full producer transcript preserves an oral appearance but may contain transcription errors. A condensed Q&A or broadcaster-selected quotation is direct speech after editorial selection. ARK analyst reports, valuation models, FAQs, and Big Ideas are organizational work unless Wood is the named speaker or author.

The 30 excerpts below come from 29 underlying materials dated 2014–2026. Every excerpt is 21 words or fewer; the two excerpts from the same 2014 letter total 20 words, so combined quoted language from every underlying work remains below 25 words. Short fragments are kept in their local context and are not made to prove investment results. Forecasts and self-descriptions remain claims until outcomes or independent evidence test them.

Contrarianism, Change, and Innovation

1. Start with shared facts, not a private reality (2014). “we address the same concerns and come to a different conclusion.” Wood was contrasting ARK's reading of weak markets with the prevailing risk-off interpretation. Wood, A Different Interpretation of Recent Market Signals.

2. Name the macro possibility (2014). “We do see several reasons for a deflationary boom.” This is an early ARK-era forecast, not a timeless fact; later inflation makes its frozen date essential. Wood, A Different Interpretation of Recent Market Signals.

3. Treat technological position as existential (2020). “Staying on the right side of change will determine success and failure.” The full passage extends the warning from portfolios to companies, careers, and countries. Wood, Investors Beware.

4. Expect crises to accelerate adoption (2020). “we believe that disruptive innovation will take root and gain significant market share.” This was a contemporaneous pandemic forecast; the essay is evidence of Wood's thesis, not reliable retrospective medical history. Wood, The Coronavirus.

5. Reduce the worldview to one causal claim (2021). “Innovation solves problems and creates new opportunities.” In the full ICE transcript, Wood uses congestion and drones to illustrate how solving one problem can open another market. Inside the ICE House, episode 242.

6. Act before consensus certainty (2025). “Investors who wait for perfect clarity may find their performance lagging in the most transformative growth cycle in history.” This is a manager's argument for early action, not proof that any specific forecast will mature. Wood, The Recession Started Months Ago. Now What?.

7. Frame fear as an opportunity set (2022). “Fear of the future is palpable these days, but crisis historically has created opportunities.” The signed year-end message came after a severe two-year risk-off period, making it both philosophy and defense. Wood, Year-End Message.

8. Keep the asset thesis distinct from price (2018). “we remain convinced that bitcoin is the first of its kind in a new asset class”. Wood wrote after bitcoin had fallen sharply; the statement records conviction through a drawdown, not a verified return. Wood, $1,700? Even Bitcoin's Bear Case Is Bullish.

9. Express scarcity as a programmed rule (2026). “Bitcoin is mathematically metered to increase ~0.82% per year for the next two years”. The surrounding comparison is with gold supply; it does not make bitcoin demand or price mathematically certain. Wood, Cathie Wood's 2026 Outlook.

Research, Models, and People

10. Make openness part of the process (2023). “to create a transparent research ecosystem designed to surface and share information”. Wood gives this as one reason for founding ARK; transparency still does not make every assumption correct. Wood, What The Market Overlooked in 2022.

11. Reject ownership of expertise (2020). “I do not believe that domain expertise resides with any one firm.” The interview connects this belief to public research, social-media engagement, and outside challenge. Institutional Investor interview.

12. Treat team composition as an input (2020). “The diversity of the team, I do consider that part of our secret sauce.” Goldman presents the line in an official, producer-selected transcript excerpt about cross-disciplinary staffing. Talks at GS.

13. Locate exponential model risk early (2021). “If you make an incorrect assumption early on and you carry it on too long, you'll make an exponential mistake.” The SALT transcript ties outside criticism to battle-testing assumptions. SALT Talks #199.

14. Hire for bounded knowledge (2022). “They know what they know. And they know what they don't know. That's the most important thing.” TIME labels its Q&A condensed and edited for clarity. TIME interview.

15. Invite assumption-level challenge (2023). “Look at how wrong it could be. Put your own assumption in if you think ours is going to be mistaken.” Wood was explaining why ARK publishes models; an editable model is a falsification aid, not independent validation. Fortune interview.

16. State a formal risk vocabulary (2025). “We have a six-metric scoring system to help us manage the risks in the portfolio.” The edited Morningstar Q&A then distinguishes quantitative scores, valuation, and thesis risk without publishing the full weight formula. Morningstar interview.

17. Connect market structure to the research opening (2019). “The capital committed to innovation in the public markets has been disappearing because of the move to passive.” This is Wood's explanation for an opportunity, not proof that passive investing causes mispricing in every ARK holding. Livewire interview.

18. Expect consumers to adopt first (2025). “the consumer is going to move much faster than enterprises”. In the Federal Reserve conference transcript, Wood was discussing agentic commerce and AI adoption, not all innovation categories. Federal Reserve Payments Innovation Conference.

Valuation, Concentration, and Selling

19. Redefine value through duration (2018). “given the right investment time horizon, TSLA is a deep value stock today.” The open letter's claim depends on ARK's five-year assumptions; deep value here is prospective rather than balance-sheet value. Wood, Dear Elon.

20. Concentrate into stress (2018). “During periods of market volatility, investors should concentrate their portfolios toward their highest conviction names”. This is a direct prescription from Wood, but it presupposes that the conviction ranking is right. Wood, Addressing Concerns During Periods of Market Volatility.

21. Answer the no-valuation criticism (2022). “Despite ARK's reputation, price is absolutely a consideration in our process.” Goldman's publication-grade Q&A is direct evidence of the stated process, not proof of valuation accuracy. Goldman Sachs, Equity Bear Market: A Paradigm Shift?.

22. Defend a forward-looking method (2021). “We are looking forward, not backward, and many of the questions you've just asked are backward-looking.” The edited Morningstar debate preserves Rob Arnott's adversarial challenge rather than presenting the answer in isolation. Morningstar debate.

23. Own a controversial exit (2023). “We do not regret selling Nvidia in the flagship.” Wood distinguished ARKK from specialist funds; the sentence records her defense, not the later opportunity cost. Morningstar interview.

24. Describe drawdown behavior explicitly (2024). “If we're wrong, what we do during a downturn is we concentrate our holdings towards our highest conviction names.” The phrase if we're wrong does not itself resolve how ARK detects a broken thesis. Morningstar interview.

25. Reject the claim that ARK never sells (2026). “I sell all the time. I do. We take profits all the time.” The official panel transcript records Wood discussing reallocation toward multi-omics; Milken warns that transcripts can contain errors. Milken Global Conference transcript.

26. Address key-person risk (2026). “But we have a very firm succession plan here.” The Bloomberg recording is current direct voice, but the public statement does not disclose the full plan or remove Wood's centrality. Bloomberg ETF IQ.

Forecasting, Error, and Accountability

27. Admit fallibility before the miss (2021). “I am sensitized to the possibility that I could be wrong on inflation.” The signed letter still defended ARK's deflation view; this is calibration language, not yet an admission of realized error. Wood, Innovation Stocks Are Not in a Bubble.

28. Acknowledge the realized macro error (2022). “We were wrong on one thing, and that was inflation being as sustained as it has been.” CNBC selected the on-air remark rather than publishing a full transcript. CNBC, Squawk Box.

29. Challenge policy consensus tersely (2022). “Unanimous? Really?” Wood's open letter follows the question with conflicting commodity, inventory, currency, and labor evidence; the rhetoric does not settle the empirical dispute. Wood, Open Letter to the Fed.

30. Keep the falsifiable macro call visible (2024). “In our view, deflation should be the concern.” The letter restates Wood's cyclical and technological deflation thesis after the 2022 inflation miss. Wood, The Journey From Monetary Shock To An Innovation-Led Economic Boom.

Annotated Index of Primary Materials

Wood-Bylined Letters and Commentary

1. 2014 — A Different Interpretation of Recent Market Signals. Earliest recovered ARK-era statement of Wood's contrarian, deflationary-boom framework.
2. 2018 — $1,700? Even Bitcoin's Bear Case Is Bullish. Direct bitcoin conviction through a severe drawdown; unusually clear asset-class framing.
3. 2018 — Dear Elon. Signed open letter with a concrete Tesla horizon, valuation range, and public-market-access argument.
4. 2018 — Addressing Concerns During Periods of Market Volatility. Clearest written prescription to concentrate into higher-conviction names during corrections.
5. 2020 — The Coronavirus. Frozen crisis-era record of the accelerated-adoption thesis and several forecasts that require outcome checking.
6. 2020 — Investors Beware. Condenses creative destruction, incumbent value traps, and “right side of change.”
7. 2021 — Innovation Stocks Are Not in a Bubble. Direct defense of five-year valuation, concentration, and fallibility on inflation.
8. 2022 — Year-End Message. Signed reflection after the drawdown, with crisis/opportunity and commercialization arguments.
9. 2022 — Open Letter to the Fed. Falsifiable policy letter arguing that lagging indicators were producing excessive tightening.
10. 2023 — What The Market Overlooked in 2022. Post-drawdown explanation of ARK's transparent research ecosystem and deflation case.
11. 2024 — The Journey From Monetary Shock To An Innovation-Led Economic Boom. Restates the monetary-error and productivity-led recovery thesis.
12. 2025 — The Recession Started Months Ago. Now What?. Long argument for acting amid uncertainty as a rolling recession gives way to innovation-led growth.
13. 2026 — Cathie Wood's 2026 Outlook. Latest comprehensive outlook; Wood is named author and credits four collaborators.

Interviews and Debates

14. 2019 — Livewire, “The Five Platforms of Disruptive Innovation”. Host-published interview on passive market structure and the public-market research opportunity.
15. 2020 — Institutional Investor. Long profile/interview on crisis behavior, open research, Tesla, criticism, and communication.
16. 2020 — Talks at GS. Official producer-selected transcript excerpts on team design, mentors, and cross-disciplinary research.
17. 2021 — SALT Talks #199. Organizer-hosted conversation with a full transcript on open research and exponential-assumption error.
18. 2021 — Morningstar debate with Rob Arnott. Edited adversarial exchange on valuation, bubbles, profitability, and sell discipline.
19. 2022 — TIME. Condensed and edited Q&A on transparency, analyst temperament, faith, women in finance, and ARK's purpose.
20. 2022 — Goldman Sachs, Equity Bear Market: A Paradigm Shift?. Publication-grade Q&A after the drawdown on price, rates, profitability, and dot-com comparisons.
21. 2022 — CNBC, Squawk Box. Broadcaster-selected direct admission that sustained inflation exceeded Wood's forecast.
22. 2023 — Fortune. Full edited interview on model challenge, AI, responsibility, and shareholder losses.
23. 2023 — Morningstar on Rize, AI, Nvidia, and bitcoin. Direct defense of the Nvidia sale and cross-platform portfolio construction.
24. 2024 — Morningstar on Tesla, AI, and bitcoin. Direct Q&A on correction behavior, concentration, and the monetary shock.
25. 2025 — Morningstar on DeepSeek and portfolio risk. Current edited Q&A on AI cost curves, scoring, portfolio evolution, and investor fit.
26. 2026 — Bloomberg ETF IQ. Recorded current statement on liquidity and the existence—but not details—of a succession plan.

Speeches, Panels, and Podcasts

27. 2021 — Inside the ICE House, episode 242. Full host transcript on open research, innovation platforms, and problem-driven opportunity.
28. 2025 — Federal Reserve Payments Innovation Conference. Official panel transcript on AI agents, payments, stablecoins, productivity, and consumer adoption.
29. 2026 — Milken, “Manias, Panics, and Crashes”. Adversarial multi-speaker panel on bubbles, private/public valuation, SpaceX, and sell discipline.

No verified sole-authored Wood book, formal congressional testimony, Wood-signed SEC comment letter, or pre-ARK Jennison/Tupelo/AllianceBernstein writing archive was recovered. Faith Driven Investing credits Wood among eighteen contributors, but its public preview does not isolate a Wood chapter or page span. ARK's “In The Know” and FYI archives are substantial primary recordings, yet many pages offer summaries and chapter markers rather than full transcripts; those editorial labels are not quoted as Wood.

Provenance, Criticism, and Current Boundary

• One underlying work is one copyright budget. Mirrors, syndications, clips, transcript vendors, and the same interview in audio and text are not separate sources.
• Institutional voice stays institutional. Wood's bylined we can show what she communicated as ARK's CEO/CIO, not which analyst originated a model or what she held personally.
• Editing is visible. TIME and Morningstar Q&As are edited; CNBC supplies a selected broadcast quote; Milken warns its official transcript may contain errors. Those layers are named rather than silently upgraded to raw speech.
• Direct voice is not performance evidence. The record is rich in conviction and explanation but thin in stock-level postmortems. The 2022 inflation admission is real; the 2021 fallibility sentence is not a realized-error confession; the Nvidia defense must be read alongside the missed rally.
• Outcome criticism belongs beside rhetoric. ARK's published models invite challenge, but transparency does not eliminate selection bias, revision risk, factor exposure, or the possibility that concentration magnifies a broken thesis.
• Legal claims stay bounded. As of 2026-07-24, ARK's adviser-filed brochure reports no material disciplinary action against the adviser or its management persons, and Wood's brochure reports none. That is not universal regulator clearance, and investee-company disputes are not Wood misconduct without an evidentiary link (ARK Form ADV, 2026, pp. 70, 94–95).

What the Record Actually Says

Across twelve years, Wood's vocabulary is remarkably stable: disruptive innovation solves problems, cost declines open markets, crises accelerate adoption, and a five-year horizon can turn volatility into opportunity. Her claimed edge is organizational as well as thematic—cross-domain hiring, open research, public models, external challenge, and explicit scoring.

The tensions are equally stable. A future-first method can dismiss inconvenient base rates as backward-looking. Concentrating in a downturn can exploit indiscriminate selling or compound a model error. Publishing assumptions improves inspectability but does not freeze every forecast or establish that outsiders' challenges change the portfolio. ARK's published models and Wood's own words contain the correct safeguards—incorrect early assumptions become exponential mistakes, price matters, and inflation proved more persistent than expected—but they do not disclose a complete falsification rule or postmortem ledger.

This corpus therefore establishes what Wood chose to communicate, not whether ARK produced style-adjusted alpha or whether every forecast was personally hers. Its strongest transferable lesson is procedural: show assumptions, name the horizon, invite challenge, and preserve dated misses. Its strongest warning is also procedural: conviction becomes evidence only after the thesis, valuation, and outcome are tested independently.

Research current through 2026-07-24

Evidence Boundary: A Public Letter Writer, Not a Sole-Author Book Author

Cathie Wood has a substantial public-writing record, but it is mainly a record of named institutional authorship. A Wood byline or signature establishes responsibility for the published letter; it does not prove that she drafted every sentence or originated every chart, model, forecast, or portfolio decision described through ARK's institutional we. Explicit collaborator credits are therefore retained. Analyst-authored valuation models, Big Ideas, prospectuses, fund reports, videos, and corporate copy are not silently converted into Wood writings.

No verified sole-authored Wood book, academic paper, pre-ARK writing archive, formal congressional testimony, or Wood-filed regulator comment was recovered. The only verified book-form item is her contribution to an eighteen-author volume. The ranked shelf below consequently treats articles as articles: each has a central thesis, five paraphrased ideas, and the best sections to read rather than invented “chapters.”

Works By Wood and Explicitly Collaborative Writings

1. What We Do: A Letter to Investors From Cathie Wood (2024)

Central thesis. ARK's post-drawdown case is that public markets are underpricing long-duration innovation while over-rewarding a narrow group of mature companies; active, concentrated research can exploit that gap if its forecasts are right (Wood, 2024).

Key ideas:

• A five-year horizon is meant to separate technology adoption from short-term price and rate shocks.
• Benchmark concentration can create its own risk when a handful of incumbents dominate index returns.
• Research and stock-selection errors, not only macro conditions, contributed to ARK's weak results.
• Nvidia illustrates a genuine distinction between recognizing an enabling technology and owning its best expression at the right time.
• Adjusting profitability for research spending and stock compensation can illuminate reinvestment, but can also make an optimistic valuation look cleaner.

Best sections: “Executive Summary,” “Market Backdrop,” “What About Nvidia?”, the valuation comparison, the tech-bust retrospective, and the conclusion. This is a Wood-bylined first-person letter with ARK data and institutional claims, not a solo research paper.

2. Innovation Stocks Are Not in a Bubble (2021)

Central thesis. A sharp correction can increase the prospective return of innovation equities because ARK values them on expected five-year cash flows rather than current multiples (Wood, 2021).

Key ideas:

• Innovation is framed as problem-solving whose adoption often accelerates in crises.
• A five-year model is supposed to absorb near-term volatility rather than deny it.
• During corrections, ARK concentrates toward names it judges to have the strongest risk-adjusted upside.
• Technological cost declines are treated as a structural deflation force against cyclical inflation.
• Wood acknowledged that her inflation view could be wrong, yet the letter still made unusually aggressive strategy-level return claims.

Best sections: the five-year valuation defense, concentration explanation, inflation discussion, and closing forecast. A later issuer clarification says the forecast referred broadly to ARK strategies, not a guaranteed return for one fund; the dated claim should be evaluated, not generalized.

3. ARKK and Nasdaq 100: A Spurious Correlation (2022)

Central thesis. Similar recent price declines did not make ARKK a conventional technology-index proxy because ARK claimed a different research process, company set, and source of prospective growth (Wood and collaborators, 2022).

Key ideas:

• Price correlation over one regime does not establish identical business exposure or investment logic.
• First-principles research starts with technology cost curves and addressable markets, then moves to companies.
• Mature mega-cap technology and earlier-stage disruptive companies can carry different operating risks despite moving together.
• A large drawdown can create expected return only if adoption and company-level forecasts survive.
• Contrarian positioning is an input, not evidence that consensus is wrong.

Best sections: the correlation distinction, technology-cost discussion, first-principles research explanation, and valuation comparison. The page explicitly credits Sam Korus, Nick Grous, Dan White, and Julian Falcioni; it is collaborative ARK research under Wood's byline.

4. Cathie Wood's 2026 Outlook: The US Economy Is a Coiled Spring (2026)

Central thesis. A multi-year rolling recession, disinflation, improving policy conditions, and converging technologies could release an innovation-led expansion (Wood and collaborators, 2026).

Key ideas:

• Weakness across housing, manufacturing, and parts of consumption is presented as a recession spread over time rather than one synchronized contraction.
• Technological productivity is expected to reinforce cyclical disinflation.
• Artificial-intelligence investment is treated as both a capital-spending cycle and an economy-wide productivity catalyst.
• Entrepreneurship, deregulation, and easier financial conditions are cast as potential release valves.
• Bitcoin's programmed supply is contrasted with gold, but scarcity alone does not make demand or price certain.

Best sections: the rolling-recession chronology, inflation and policy discussion, AI-capex analysis, and innovation-platform outlook. The letter names Dan, Will, Katie, and Keith as helpers who crafted it; “Wood-authored” must not mean unassisted.

5. What the Market Overlooked in 2022 (2023)

Central thesis. The 2022 selloff obscured accelerating technological progress, and ARK's open research system was designed to keep that company-level evidence visible through the drawdown (Wood, 2023).

Key ideas:

• Transparency is presented as a way to surface information and invite challenges to assumptions.
• The five innovation platforms are expected to reinforce one another through convergence.
• Falling technology costs can create consumer surplus even while tightening financial conditions depress valuations.
• Commercial and scientific milestones are separated from short-term stock performance.
• A long horizon does not resolve whether ARK selected the right companies, sized them well, or updated models quickly enough.

Best sections: Wood's founding reflection, the open-research explanation, the five-platform review, and the conclusion. The work is strongest as a frozen post-drawdown defense, not as independent proof of the forecasts it catalogues.

6. Disruptive Innovation and Profitability (2022)

Central thesis. Current accounting losses can be rational when a company deliberately funds technology, distribution, and scale that may support larger future cash flows (Wood and collaborators, 2022).

Key ideas:

• Present profitability may understate a business that is investing heavily in a new market.
• Research and development can be economically productive even when expensed immediately.
• Gross margins, unit economics, and reinvestment choices should be separated from a simple profitable/unprofitable label.
• “Sacrifice now, harvest later” is plausible only when spending creates a defensible advantage.
• The framework can rationalize persistent losses unless milestones, dilution, financing needs, and competitive outcomes are tested independently.

Best sections: the accounting distinction, company examples, and reinvestment framework. Tasha Keeney, Thomas Hartmann-Boyce, Julian Falcioni, and Paul Wilson are named co-authors; its analysis is not Wood-personal work.

7. Investors Beware: Stay on the Right Side of Change (2020)

Central thesis. Portfolios built from incumbent-heavy benchmarks can become value traps when converging technologies destroy old profit pools faster than conventional valuation captures (Wood, 2020).

Key ideas:

• Creative destruction threatens companies, sectors, jobs, and capital invested in obsolete assets.
• Low historical multiples do not protect a business whose economics are deteriorating.
• Technology platforms interact, so disruption should be researched across sectors rather than within silos.
• Benchmark sensitivity can turn yesterday's winners into an unexamined default allocation.
• Avoiding incumbents is not enough; an investor must still identify, value, and retain the actual winners.

Best sections: the “creative destruction” setup, incumbent examples, convergence discussion, and final portfolio warning. This short letter is the cleanest statement of Wood's worldview and also the easiest to overextend beyond its evidence. Its forecast of roughly 40 million electric-vehicle sales and about 45% share in 2025 (Wood, 2020) materially overshot the IEA's later report of just over 20 million and 25% (IEA, 2026).

8. Broad-Based Benchmarks Seem to Be Short Disruptive Innovation (2022)

Central thesis. Market-cap-weighted benchmarks can be structurally underexposed to emerging platforms and overexposed to firms vulnerable to them (Wood, 2022).

Key ideas:

• Index weights describe past capitalization, not a forecast of future relevance.
• Active share is treated as necessary to express a genuinely different innovation view.
• Sector classifications can obscure technology convergence across traditional categories.
• Benchmark agnosticism removes one constraint while increasing tracking error and manager risk.
• Being different from an index does not establish superior security selection or portfolio construction.

Best sections: the benchmark-exposure comparison, active-share discussion, and innovation-platform mapping. The byline is Wood's, but the quantitative conclusions are ARK research and should remain attributed that way.

9. Dear Elon: An Open Letter Against Taking Tesla Private (2018)

Central thesis. Taking Tesla private at $420 would, in Wood's view, deny public investors access and crystallize a valuation far below ARK's five-year scenarios (Wood, 2018).

Key ideas:

• Public ownership preserves liquidity, price discovery, and participation by ordinary investors.
• A volatile share price need not invalidate a long-duration operating thesis.
• Autonomous mobility was central to the upper valuation cases, not a minor optional product.
• Five-year scenarios can reveal upside hidden by near-term auto-industry comparisons.
• Scenario range is not precision: assumptions, probabilities, capital needs, regulation, and execution determine whether a target matures.

Best sections: the direct request to Musk and Tesla's board, the public-access argument, and the scenario summary. The letter belongs to Wood; the page credits Brett Winton, Tasha Keeney, and Sam Korus with leading the valuation models.

10. ARK Extends an Open Letter to the Fed (2022)

Central thesis. The Federal Reserve was relying on lagging inflation and labor indicators while real-time prices and activity suggested that continued tightening could create deflation (Wood, 2022).

Key ideas:

• Commodity prices, freight rates, inventories, and the dollar can turn before headline inflation.
• Employment data can remain firm after more leading measures weaken.
• Synchronized tightening can amplify policy lags.
• A unanimous policy decision should still be stress-tested against conflicting market evidence.
• Selective leading indicators can also mislead; the letter is a dated forecast to score, not formal testimony or a regulator filing.

Best sections: the commodity and inventory evidence, labor-market lag argument, and concluding policy challenge.

11. A Different Interpretation of Recent Market Signals (2014)

Central thesis. The same weak-market facts that produced conventional pessimism could instead be consistent with lower costs, productivity gains, and a “deflationary boom” (Wood, 2014).

Key ideas:

• Contrarianism should begin from shared observations rather than a separate factual universe.
• Falling input prices can support real growth rather than always signal collapsing demand.
• Investor anxiety can coexist with a continuing bull market.
• Currency, rates, oil, and technology are read as one macro system.
• A novel interpretation still requires later outcome testing; the recurring deflation thesis became a persistent source of both insight and error risk.

Best sections: the opening alternative interpretation, the deflationary-boom case, and the closing wall-of-worry discussion. This is the earliest recovered ARK-era Wood byline, not proof that no earlier private writing exists.

Book-Form Contribution and the Missing Book

Tyndale credits Wood as one of eighteen contributors to Faith Driven Investing: Every Investment Has an Impact—What's Yours? (ebook 2022; hardcover 2023). The publisher describes her topic as finding a call to create and innovate in investing. Public catalog and preview material did not isolate a verified chapter title, page span, or full Wood-only text, so five detailed “key ideas” or “best chapters” cannot be supplied without inventing access. This is a contribution by Wood, not “Cathie Wood's book.”

Best Works About Wood, Ranked

1. Robby Greengold, Morningstar Managed Investment Report (2026). Best current independent diligence on Wood's centrality, personnel, succession, correlation, liquidity, exits, and full-period comparative returns. The linked vehicle is a European UCITS share class; qualitative analysis reaches ARK and U.S. ARKK, but its vehicle table is not the U.S. fund record.
2. Matt Phillips, “God, Money, YOLO: How Cathie Wood Found Her Flock” (2021). Best long biography of career, faith, founding, organization, public research, and retail following. Wood declined an interview, and the peak-era endpoint predates the 2022 collapse.
3. Amy Arnott, “15 Funds That Have Destroyed the Most Wealth Over the Past Decade” (2024). Essential flow-adjusted investor-outcome corrective. Dollar wealth destruction is sensitive to scale and timing; it is not Wood's personal P&L, a simple NAV return, or proof of misconduct.
4. David Gura, “A Superstar Investor With the Midas Touch or Just Lucky?” (2022). Best boom-to-bust audio/article treatment of skill, rate sensitivity, fandom, and the five-year defense. It is a May 2022 snapshot, and inverse-fund sponsor Matthew Tuttle had an opposing economic interest.
5. Evie Liu and Leslie P. Norton, “ARK's Cathie Wood Disrupted Investment Management” (2021). Strong peak-era synthesis of career, research architecture, active ETFs, staffing, faith, and audience design. Pair its pre-crash optimism with later diligence.
6. Antoine Gara, “How Cathie Wood Beat Wall Street by Betting Tesla Is Worth More Than $1 Trillion” (2020). Best early account of ARK's economics, ownership, team model, Tesla thesis, and Wood's earlier career. Forbes's firm-value and net-worth estimates are not audited.
7. Hannah Zhang, “Why the Michigan Retirement System Is Still Bullish on Cathie Wood” (2022). Rare allocator perspective on using ARK as a high-conviction manager. The pension's reported result is strategy-specific, then-current, and not a Wood-personal composite.
8. Amy Whyte, “Cathie Wood Still Thinks Tesla Is Going to $6,800. Why?” (2020). Best contemporaneous record of ARK's crash behavior, client communication, open models, and Tesla conviction. Targets require split normalization and outcome testing.
9. Belinda Luscombe, TIME's edited Q&A (2022). Best compact treatment of transparency, team temperament, faith, women in finance, and investor purpose, explicitly condensed and edited for clarity.
10. Michelle Celarier, “The Passion of Cathie Wood” (2022). Useful post-collapse account of brand, faith, forecasts, critics, and persistence. Its hard paywall limits independent passage checking, and several quoted critics were economically short ARKK or Tesla.

Criticism, Legal Status, and Reading Discipline

The strongest criticism is methodological rather than personal. Morningstar's current report argues that conviction became obstinacy in cases such as Teladoc, holdings remained highly correlated despite thematic labels, some exits were difficult, and new risk hires had not yet established effectiveness. That is analyst judgment, but it directly tests the letters' claims about long horizons, concentration, and research quality. It also prevents a few early thematic successes from proving repeatable stock selection.

ARK's March 2026 adviser brochure says Wood is founder, CEO, CIO, primary portfolio manager, and principal owner, and reports no material legal or disciplinary action against the adviser or its management persons; Wood's supplement lists none (ARK Form ADV, 2026, pp. 4, 70, 94–95). Those are adviser-filed statements, not universal regulator clearance. Routine product orders, public comments, impersonation cases, and investee-company disputes are not Wood misconduct without an evidentiary link.

The most useful reading method is paired and chronological: read the 2014 origin statement, the 2020 creative-destruction essay, and the 2021 valuation defense; then read the 2022 collaborative rebuttals, 2023 post-drawdown letter, and 2024 full defense beside Morningstar's 2024 investor-outcome analysis and 2026 manager report. Finish with the 2026 outlook as an open forecast. This preserves both the durability of Wood's framework and its central unresolved question: whether getting broad technological direction right can overcome company selection, valuation, sizing, liquidity, and exit error.

Missing and Excluded Materials

• No serious independent book-length biography or completed evidence-grade documentary was found. Short 2025 self-published biographies expose little sourcing or editorial process and do not outrank the institutional shelf.
• Big Ideas, valuation models, FAQs, prospectuses, and analyst papers remain ARK or named-team work unless a Wood-authored section is explicit.
• Recurring “Commentary From ARK's CIO” reports are institutional fund commentary, not independent essays or a personal trading ledger.
• The original, complete Jennison, Tupelo, and AllianceBernstein writing record remains unavailable.
• A byline establishes published responsibility; it does not convert institutional data into Wood-personal research or fund results into her personal performance.

Research current through 2026-07-24

Evidence Boundary

Cathie Wood has not published a personal checklist with fixed rules. The public method is Wood-led but organizational: ARK's analysts and external Theme Developers generate and challenge research; directors coordinate it; and Wood has final accountability for public-equity investment decisions. The firm's named vocabulary includes disruptive innovation, Wright's Law, convergence, the Open Research Ecosystem, five-year valuation models, the ARK Portfolio Tracker, benchmark agnosticism, and risk-on/risk-off regimes. The numbered labels below are reconstructions of how those elements fit together, not concepts Wood is claimed to have coined (ARK investment process; ARK Form ADV, 2026, pp. 4–5, 20–22, 94–95).

The distinction matters. An ARK model is not automatically Wood's personal analysis; ARKK is not every ARK strategy; a public methodology is not a complete control system; and a fund result is not Wood's personal return. No current fixed stop-loss, maximum portfolio-drawdown rule, full factor budget, complete score-to-weight equation, or personal forecast ledger was recovered. Current help-center weights are operating norms for active ETFs, not immutable prospectus limits or retail instructions (ARK position-weight guidance). The company-model horizon is five years, while current investor guidance asks for a full market cycle or at least seven years; neither is permission to postpone a failed milestone (ARK investor-horizon guidance).

ARK's March 2026 adviser brochure reports no material legal or disciplinary action against the adviser or its management persons. That is a bounded adviser-filed statement, not universal regulator clearance; routine product orders, public comments, impersonation cases, and investee-company disputes are not Wood misconduct without an evidentiary link (ARK Form ADV, 2026, pp. 70, 94–95).

Named Vocabulary and Reconstructed Frameworks

1. The disruption gate comes before the stock

ARK's official test asks whether a technology can produce sharp cost declines, cut across sectors and geographies, become a platform for further innovation, and unlock additional demand. This is a gate, not a buy signal. A promising technology can still fail commercially, arrive too early, transfer value to customers, or enrich a competitor rather than the selected company (ARK thematic process, pp. 1–3).

The operational sequence is therefore technology -> market -> value chain -> company -> security. Starting with the ticker reverses the method and invites a fashionable story to masquerade as a structural change.

2. Cost decline, elasticity, and value capture are three separate bridges

ARK uses Wright's Law and related learning curves to estimate how cumulative production may lower unit cost. It then models demand elasticity: how much a lower price can expand adoption. The third bridge is frequently omitted in simplified accounts—who retains the economics after competition, financing, regulation, dilution, and falling prices (ARK investment process).

A useful reconstruction writes three independent claims:

1. cumulative output should lower unit cost at a specified rate;
2. lower price should enlarge the addressable market and adoption path; and
3. the company should defend enough margin, share, or network value to reward shareholders.

Correctly forecasting the technology does not prove the third claim.

3. Convergence is a network of dependencies, not a theme count

Wood's recurring map contains artificial intelligence, robotics, energy storage, blockchain, and multiomics, with applications strengthened when platforms interact. The model's insight is that conventional sector research can miss cross-domain reinforcement. Its danger is double counting: the same AI, duration, funding, or risk-appetite assumption can appear in many differently labeled holdings (Wood, 2026; Big Ideas 2026).

The correct portfolio map therefore has two layers. The first shows technological dependencies and possible positive spillovers. The second shows shared financial risks. More thematic labels do not necessarily create more economic diversification.

4. Top-down maps must survive bottom-up underwriting

ARK first sizes a multi-year value-chain transformation and defines a potential company universe. It then builds company briefs, scores, and five-year models. This deliberately avoids starting from benchmark membership, but it also imposes a burden: every broad opportunity must be translated into company-specific adoption, share, margins, dilution, capital needs, and valuation (ARK thematic process, pp. 2–3). The architecture predates ARK: in a 2010 AllianceBernstein interview, Wood described combining secular themes with bottom-up fundamental research and judgment as “classical” investing (Wood interview, 2010).

The model is an evidence funnel. A trillion-dollar market is not a valuation. Each step should shrink uncertainty rather than simply repeat the same total-addressable-market claim in greater detail.

5. Five years is a modeling bridge, not a patience exemption

Five-year models force a connection from current operations to a future business rather than extrapolating the next quarter. Public methodology monitors unit volume, cost, penetration, share count, margins, and future multiples. A stock's expected average return dropping below 15% over five years appears as a score-warning input; it is not a guaranteed fund return, universal automatic sale, or promise that every forecast matures on schedule (ARK thematic process, p. 3).

Time can distinguish adoption from quarterly noise, but it cannot rescue a broken unit-economics or value-capture thesis. Each model needs intermediate milestones and a frozen vintage so that “long term” does not become an indefinitely moving deadline.

6. The Open Research Ecosystem is an error-correction claim

ARK organizes analysts around technologies rather than sectors and invites input from social media, outside specialists, academics, startups, and other Theme Developers. The structure can surface cross-domain evidence and public criticism. Theme Developers are not employees and receive no monetary compensation, while ARK's filing describes conflict and information controls (ARK thematic process, pp. 2, 4; ARK Form ADV, 2026, pp. 6–8).

Openness is not falsification by itself. Public attention can reward exciting forecasts, external participants can have economic interests, and a model can be revised without an accessible decision log. The system works only if contrary evidence changes assumptions, scores, weights, or exits.

7. Six scores turn a story into a monitored thesis

The current ARK Portfolio Tracker evaluates six areas: company, people, and culture; execution; moat or barriers; product leadership; valuation; and thesis risk. Warning signs include personnel departures, inadequate research spending, poor execution, lost market share, competition, wrong key indicators, regulatory or geopolitical exposure, adoption delays, legal threats, and governance problems. A score of 6/10 or below triggers full review for an active ETF; it is not an automatic sale (ARK thematic process, p. 3; ARK FAQ).

This is best understood as a dashboard, not an algorithm. A 2015 process document instead listed eight metrics, including thematic relevance and growth potential; the later six-part framework must not be projected backward as timeless (ARK process, 2015). The public does not have the score history, factor weights, override record, or exact mapping from score to position size.

8. Price is a relative-opportunity signal

ARK may add when negative sentiment lowers price without changing its operating thesis, or trim a liquid name to fund one with greater modeled upside. It may sell when the thesis changes, metrics no longer support the weight, or a company ceases to lead its innovation category. The decision is comparative: each holding competes for capital against the rest of the opportunity set (ARK investment process).

The danger is circularity. A decline mechanically raises model-implied return unless forecasts are marked down. Price becomes useful evidence only when operating milestones, dilution, financing, competition, and opportunity cost are re-underwritten independently. ARK's FAQ also distinguishes roughly 15% annual turnover in names from roughly 70% historical share turnover as weights change; those are descriptive figures for a representative strategy, not universal targets.

9. Concentrate when fear widens the claimed gap; broaden when supply returns

Wood describes concentrating into highest-conviction names during corrections. A 2026 ARK summary adds the other side: it says the portfolios consolidate in downturns and diversify in bull markets as initial public offerings and acquisitions restore price discovery. Current help-center guidance says active ETFs generally hold about 50% in their top ten, with a median position near 2%, a typical minimum near 1%, and a typical maximum purchase weight near 10%. At June 30, 2026, ARKK's official snapshot showed 50.9% in its top ten and 10.5% in Tesla [single-source]—one observation, not a rule (ARK Q4 2025 webinar summary; ARK position-weight guidance; ARK Q2 2026 report).

“Typical” is crucial. Market moves can lift a holding above its purchase weight, and the disclosed figures are neither hard fund limits nor suitable household allocations. Concentrating into a selloff is rational only if evidence quality rises faster than correlated risk.

10. Benchmark agnosticism trades one error for another

ARK rejects benchmark membership as the starting universe because capitalization records past success and sector boundaries can hide convergence. This permits genuine differentiation: ARK reported 43 ARKK holdings, roughly 91% active share, and only about 38% overlap with the S&P 500 as of March 31, 2026 (ARK, 2026).

Removing the benchmark can reduce incumbent and career-risk bias, but it also removes a diversification and valuation brake. The appropriate test is not whether a portfolio differs. It is whether the differences are compensated after concentration, fees, volatility, liquidity, and timing.

A Reconstructed Decision Checklist

This checklist operationalizes the public record. It is not a leaked ARK form or personal financial advice.

1. Define the portfolio role. State why the exposure exists, its horizon, and the loss the whole portfolio can absorb. Wood says ARK should not be the core of an adviser client's portfolio (Morningstar interview, 2025).
2. Pass the disruption gate. Identify the cost decline, cross-sector reach, platform effect, and source of incremental demand. Reject a theme supported only by a large market forecast.
3. Map the value chain. Identify suppliers, complements, competitors, regulators, financing needs, and which participant captures rather than merely creates value.
4. Separate the three bridges. Write the cost curve, elasticity/adoption, and shareholder value-capture assumptions independently.
5. Map convergence twice. Record technological reinforcement, then common duration, funding, valuation, liquidity, and sentiment exposure.
6. Build the company brief. Score people and culture, execution, moat, product leadership, valuation, and thesis risk. Attach observable warning signs to every category.
7. Create frozen cases. Model bear, base, and bull five-year paths for units, price, share, margins, dilution, capital needs, and terminal valuation. Preserve the dated original before revising it.
8. Calculate expected return honestly. Use current price and probability-weighted outcomes. Treat ARK's 15% methodology flag as its review input, not a borrowed guarantee or universal hurdle (ARK thematic process, p. 3).
9. Pre-register falsification. Specify milestones, dates, disconfirming evidence, acceptable variance, financing limits, and the event that converts volatility into thesis failure.
10. Set size outside the story. Begin with a portfolio loss budget and correlation map. Do not copy ARK's fund weights into a household account.
11. Inspect overlap. Look through ETFs and individual holdings for repeated companies and shared growth, AI, crypto, small-cap, rate, or liquidity exposure. A different sector label can conceal the same factor.
12. Underwrite implementation. Check spread, volume, taxes, currency, custody, and underlying liquidity. ETF exchange volume is not the same as liquidity in the fund's holdings.
13. Review on a schedule. Update each score and milestone at a fixed interval and after material events. Investigate a 6-or-lower score; do not turn it into an invented mechanical sale (ARK FAQ).
14. Add only after re-underwriting. A lower price is not enough. Recalculate the operating path, financing, dilution, competitive position, and portfolio-level loss.
15. Use the disclosed exit hierarchy. Trim for a better relative opportunity; reduce when metrics no longer support the size; exit on thesis change or loss of innovation leadership (ARK investment process).
16. Postmortem the vintage. Compare forecast with outcome and separate technology direction, company selection, valuation, sizing, execution, and investor timing. Never replace the old forecast in the audit trail.

Failure Modes of the Model

1. An exponential curve magnifies an early error. Small mistakes in learning rate, adoption start, market share, margin, dilution, or terminal multiple can become very large five-year valuation errors.
2. Technology success may not become shareholder value. Competition can pass cost declines to customers, an incumbent can adapt, or the chosen company can finance growth on punitive terms.
3. Living research can erase falsification. Model revisions are necessary, but rolling the date or changing inputs without scoring the previous vintage makes calibration impossible. ARK's public Tesla models moved from 2025 to 2026, 2027, and 2029 targets; that lineage should be read as separate forecasts, not one continuous proof (ARK Tesla model repository).
4. Concentration can turn conviction into obstinacy. A falling price can increase modeled upside at the same moment it is signaling a financing, competition, or execution problem. Morningstar's current report identifies Teladoc and other exits as examples of weak execution and difficult thesis change (Morningstar manager report, 2026).
5. Different themes can share one macro trade. Long-duration, unprofitable growth companies can move together when real rates, liquidity, or risk appetite changes. ARK itself says its strategies usually struggle in risk-off markets (ARK FAQ).
6. Benchmark rejection can become base-rate rejection. A future-first view can dismiss survivorship, competition, capital intensity, and historical failure as backward-looking even when they are relevant priors.
7. Transparency can become marketing feedback. Publishing bold targets attracts attention and useful criticism, but also clients, media reach, and confirmation from enthusiastic communities. The incentives are not neutral.
8. The ETF wrapper does not eliminate capacity risk. In-kind redemptions can transfer securities rather than force immediate sales, but underlying liquidity, cross-fund overlap, market makers, cash baskets, spreads, and large ownership positions still matter (ARK ETF Trust prospectus, 2025, pp. 87, 92–95).
9. Central judgment creates key-person risk. Wood's final authority makes the process coherent while concentrating selection, override, and succession risk. Morningstar's 2026 report says the team has stabilized but succession and the effectiveness of newer risk hires remain unproven.
10. Fund return and investor return can diverge sharply. Investors often arrive after visible success. Morningstar estimated that flow timing made ARKK shareholder outcomes much worse than time-weighted fund results; a thematic-ETF study separately finds specialized products tend to launch after strong underlying performance and attract attention-sensitive capital (Morningstar, 2023; Ben-David, Franzoni, and Kim).
11. Visible trades are not the process. ARK's trade files are end-of-day, unofficial, unreconciled, subject to later additions, and exclude offerings and creation/redemption activity. Copying them substitutes delayed output for the model, score, price, tax position, and portfolio context (ARK trade notifications).
12. The theme can be right and the investment wrong. Nvidia shows that recognizing AI did not guarantee holding its strongest expression at the decisive time; Wood's 2020 EV forecast shows that directionally correct adoption can coexist with a material volume and share miss (Wood, 2024; Wood, 2020; IEA, 2026).

Transferability to an Individual Investor

FINRA's practical counterweight is to look through every fund, check correlated exposures, rebalance, and understand underlying liquidity. Simply owning several funds does not remove concentration risk (FINRA). The most transferable Wood method is therefore a research discipline—not the headline forecast, ARK weight, or latest trade.

Critical Assessment

Wood's framework is strongest as a causal map. It asks where cost is falling, how demand responds, which platforms reinforce one another, and which company can capture the value. It also makes unusually much of the process inspectable: readers can see models, assumptions, holdings, direct letters, and qualitative sell rules. That openness creates real educational value.

Its weakest link is the conversion from a large technological truth to a well-priced, well-sized security. The same five-year horizon that creates independence can defer accountability; the same concentration that exploits panic can compound model error; and the same open research that invites challenge can amplify a compelling narrative. Observed outcomes show both sides: exceptional winners and rebounds, but also forecast misses, damaging exits, severe path dependence, and poor flow-timed investor experience.

The durable lesson is not to become a smaller ARK. It is to preserve the sequence and add harder brakes: technology before ticker, value capture before valuation, frozen assumptions before revision, falsification before averaging down, and whole-portfolio risk before conviction. That makes the framework usable without pretending that public transparency reproduces the institution behind it.

As of: 2026-07-24 Task: T0599 | Investor: 074-cathie-wood | Code: H-synthesis

Evidence Boundary

This synthesis combines the completed profile, philosophy, greatest-trades, own-words, key-writings, and mental-models chapters. A standalone mistakes-and-losses chapter is not yet complete, so adverse evidence is reconstructed from those six files, current fund and adviser records, forecast outcomes, and independent manager research. That gap prevents this chapter from claiming a complete loss ledger.

The public method is Wood-led but organizational. Wood has final investment authority, while ARK analysts and outside contributors build much of the research; ARKK is the best public flagship proxy, not Wood's personal account or every ARK strategy. ARKK NAV, market-price return, investor-dollar experience, firm assets, campaign P&L, and Wood's economics remain separate. ARK's March 2026 adviser brochure and Form CRS report no material disciplinary history for the adviser or its management persons; those are bounded adviser-filed statements, not universal clearance (ARK Form ADV, 2026, pp. 4–8, 20–22, 70, 94–95; ARK Form CRS). The brochure also describes prospective conflicts from affiliated ARK Capital Markets, then in the process of broker-dealer registration; no actual violation was found, and its later registration and issuer engagements remain unresolved.

Executive Brief

Cathie Wood's durable achievement is both an investment process and a product design. After a thematic career at Jennison, Tupelo, and AllianceBernstein, she founded ARK in 2014 around actively managed, transparent exchange-traded funds devoted to disruptive innovation. The method begins above the company: identify technologies with rapid cost decline, elastic demand, cross-sector reach, and platform effects; map the value chain; then underwrite which security can capture the economics. ARK's five recurring platforms—artificial intelligence, robotics, energy storage, public blockchains, and multiomics—are a research map, not proof that every associated stock is attractive (ARK Form ADV, 2026; ARK investment process; Big Ideas 2026).

Pre-ARK evidence adds a bounded checkpoint. An AllianceBernstein presentation identifies Wood as thematic CIO and the team as manager; through 2010, its ten-year composite returned 3.88% pure gross and 0.81% after maximum fees, versus 1.41% for the S&P 500 [single-source primary presentation]. Firm-level GIPS verification did not verify that composite specifically. It is evidence of an institutional strategy, not an audited Wood-personal record (AllianceBernstein, 2011).

Implementation joins top-down opportunity sizing with bottom-up five-year company models. Analysts score people and culture, execution, moat, product leadership, valuation, and thesis risk; Wood makes final decisions. Portfolios are benchmark-agnostic and concentrated, while additions, trims, and sales depend on relative modeled return and thesis evidence. That public architecture is unusually inspectable, but incomplete: no full score history, score-to-weight equation, fixed stop, maximum drawdown, factor budget, override log, or frozen forecast ledger is public (ARK thematic process; ARK FAQ).

The record supports early thematic insight, with Tesla the best documented campaign. It also shows that direction, security selection, valuation, and selling are separate skills: a 2020 electric-vehicle forecast overshot 2025 sales materially, while ARKK exited Nvidia before the decisive AI rerating (ARK Tesla model repository; Wood, 2020; IEA, 2026; Wood, 2024).

Its real contribution is disciplined question formation; its unresolved burden is converting technological maps into repeatable security selection after valuation, fees, and market frictions.

Performance is mixed and endpoint-sensitive. ARK's June 2026 report shows a competitive ten-year ARKK NAV result, a deeply negative five-year result, and slight since-inception lag to the S&P 500. Morningstar's different endpoint places ARKK behind a mid-cap growth comparator and criticizes correlated exposure, difficult exits, and key-person dependence. Flow timing made estimated investor outcomes worse than the fund's time-weighted return. These facts reject both blanket dismissal and proof by a few winners (ARK Q2 2026 report; Morningstar, 2026; Morningstar investor outcomes).

Wood belongs in the Canon as a thematic cartographer, institutional founder, and test of transparent high-conviction growth. Her strongest transferable lesson is causal sequencing: technology before ticker, value capture before valuation, and operating evidence before averaging down. Her greatest unresolved question is whether broad foresight can become repeatable, style-adjusted stock-selection alpha after fees, capacity, concentration, and investor timing. Individuals can copy the questions and frozen-accountability tools; they cannot reproduce ARK's analysts, access, execution, ETF plumbing, or risk tolerance.

The result is a powerful hypothesis engine whose value depends on disciplined calibration, implementation, and client use across complete market cycles.

Ten Transferable Lessons, Ranked

1. Put the technology-to-security chain in writing

Separate cost decline, demand elasticity, value-chain position, company execution, shareholder value capture, and security valuation. A correct platform forecast can still enrich customers, a competitor, or an incumbent rather than the chosen stock (ARK thematic process).

2. Freeze forecasts before revising them

Living research is necessary, but overwriting the old model destroys calibration. Preserve every vintage, probability, deadline, valuation contribution, and intermediate milestone. The Tesla model repository demonstrates version history; it is not a complete outcome ledger (ARK Tesla model repository).

3. Map convergence twice

First map how technologies reinforce one another. Then map shared duration, financing, valuation, liquidity, and sentiment exposures. Different theme labels can conceal one macro trade, and ARK itself says its strategies usually struggle in risk-off markets (ARK FAQ).

4. Separate a great theme from its best expression

Recognizing artificial intelligence did not guarantee owning Nvidia during its strongest rerating; identifying a platform never settles which security captures its value. Re-underwrite company share, margins, dilution, capital needs, competition, and price independently of the thematic narrative (Wood, 2024).

5. Size outside the story

Begin with a portfolio loss budget, overlap map, and realistic exit capacity. Position size should not be reverse-engineered from enthusiasm or copied from an ETF. ARK's current weights are descriptive vehicle norms, not hard limits or household allocations (ARK position-weight guidance; FINRA concentration guidance).

6. A long horizon needs short accountability intervals

Five-year modeling can distinguish adoption from quarterly noise; it cannot excuse missed operating milestones. ARK asks end investors for a full cycle or at least seven years, a different horizon from its five-year company models. Neither horizon turns patience into a sell rule (ARK investor-horizon guidance).

7. Transparency is an input, not an audit

Open models and daily trade notices invite scrutiny, but a visible output omits scores, tax lots, flows, intraday execution, and rationale. ARK's trade files are end-of-day, unofficial, unreconciled, and incomplete. Use them to form questions, not to mirror decisions (ARK trade-notification disclosure).

8. Treat selling as a separate skill

Theme discovery, averaging down, trimming, and exiting require different evidence. A lower price raises model-implied return mechanically unless assumptions fall too. Pre-register thesis-change, milestone, financing, leadership, and relative-opportunity exits before the position becomes an identity (ARK investment process).

9. Measure the investor, not only the fund

Time-weighted NAV can coexist with poor dollar-weighted experience when clients arrive after visible success. Specialized ETFs more broadly tend to launch after strong underlying performance and attract attention-sensitive capital. Product design, communication, rebalancing, and client timing are therefore part of realized investment risk (Morningstar investor outcomes; Ben-David, Franzoni, and Kim).

10. Copy the discipline, not the institution

An individual can build simple scenarios, frozen milestones, scorecards, and portfolio-overlap checks. They cannot recreate ARK's analyst network, company access, compliance, trading, creation/redemption machinery, or tolerance for severe drawdowns. A narrow ETF also is not diversified merely because it holds many tickers (Investor.gov).

Style Taxonomy

Regime Dependence

This matrix combines ARK's own risk-on/risk-off description with observed fund paths and independent critique; it is a causal interpretation, not a fitted timing rule (ARK FAQ; ARK Q2 2026 report; Morningstar, 2026).

ARK's issuer-reported Q2 2026 rebound—five active ETFs ahead of broad global indexes for one quarter and one behind—shows that opportunity and dispersion can return quickly [single-source issuer commentary]. It does not settle full-cycle alpha or validate open forecasts (Wood, Q2 2026 commentary).

Closest and Most-Opposite Investors Already in the Canon

Skill, Factor Exposure, and Luck

The defensible skill claim is strongest in research architecture, thematic direction, public communication, and institution building. ARK combines cost curves, domain specialists, open models, and active ETFs into a coherent operating system. Those public, testable hypotheses are an institutional achievement, not proof of persistent stock-selection alpha (ARK thematic process; ARK ETF Trust prospectus, 2025).

The evidence for repeatable security-selection alpha is weaker. Morningstar's April 2026 endpoint places ARKK at 12.3% annualized from October 2014 through March 20, 2026, versus 12.9% for Russell Midcap Pure Growth and 12.7% for the S&P 500 [single-source independent methodology]. ARK's June endpoint differs, and neither comparison alone isolates exposure to growth, duration, liquidity, crypto, small caps, or a few outliers. Portfolio outcomes also belong to Wood plus analysts, traders, fund flows, and vehicle mechanics—not a disclosed Wood account (Morningstar, 2026).

Independent factor models reinforce the uncertainty. A peer-reviewed 2018–21 daily study reports ARKK six-factor alpha of 0.05 with a 1.54 t-statistic, market beta of 1.24, positive size exposure, negative value and profitability exposures, and R-squared of 0.84 [single-source model]; alpha was not statistically significant, and the sample ended before the full 2022 collapse (Rompotis, 2022). A practitioner model ending February 2022 instead found a positive proprietary innovation loading and 5.4% annualized residual alpha [single-source proprietary model]. Its simulated factor and shorter endpoint cannot establish current alpha (Sparkline Capital, 2022). Their disagreement is the finding: attribution depends materially on factors, implementation, and endpoint.

Luck and regime clearly matter. The 2020 digitization boom, low rates, retail inflows, Tesla's rerating, and later crypto and AI cycles created unusually favorable paths for several holdings. The 2021–22 reversal exposed the other side: correlated duration, financing sensitivity, valuation compression, and investor timing. A thematic-ETF prior explains why vivid stories and past performance can attract capital near peaks, but it does not prove that ARKK's company research has no value (Ben-David, Franzoni, and Kim).

The correct verdict is therefore documented thematic and institutional skill; unresolved style-adjusted stock-selection alpha. A few large winners disprove the claim that the process is empty, while forecast misses, difficult exits, and flow-timed losses disprove the claim that broad foresight alone is sufficient. The missing ingredients for a stronger causal verdict are a complete frozen forecast ledger, score and override history, holdings-level factor attribution, consistent liquidity/capacity record, and audited pre-ARK or Wood-personal composite.

Unresolved Questions

1. What complete, frozen score and forecast ledger would show every original probability, revision, deadline, and outcome?
2. How much ARKK return remains after style-consistent growth, duration, liquidity, crypto, small-cap, and factor attribution?
3. What exact rules map the six-area Portfolio Tracker, modeled return, liquidity, and thesis risk into position weights?
4. Wood said in February 2026 that a succession plan exists; who is the named successor, and how would decision rights, ownership, and active-fund management transfer (Bloomberg ETF IQ)?
5. What are the current internal issuer-ownership, days-to-liquidate, factor, correlation, drawdown, and stress limits?
6. How should ARK's strongest open forecasts be scored without cherry-picking only matured successes or misses?
7. What complete tax-lot ledger separates realized and unrealized lifetime campaign P&L across ARK vehicles?
8. How much of ARK-family investor-dollar experience is attributable to product launch timing, marketing, client behavior, and portfolio returns?
9. Can a consistent Wood or strategy composite be recovered for Jennison, Tupelo, and AllianceBernstein?
10. What mistakes, omissions, and process changes will the still-incomplete standalone loss chapter establish beyond the adverse record synthesized here?
11. Have newer risk and research hires changed decision rights, exposure limits, or exit speed in a measurable way?
12. Will ARK publish postmortems that distinguish correct technology direction from company selection, valuation, sizing, and selling?
13. Did ARK Capital Markets complete broker-dealer registration, and what issuer engagements or trading restrictions followed?

Bottom Line

Wood's enduring contribution is a causal research sequence made public: cost decline, adoption, convergence, value chain, company, valuation, portfolio. Its enduring weakness is that every bridge can fail while a compelling long-horizon story remains intact. The safest adaptation is to preserve ARK's curiosity and add harder accountability—frozen forecasts, milestone dates, causal diversification, loss budgets, and independent sell rules. Copy the map-making discipline, not the headline target, fund weight, daily trade, or tolerance for drawdown.

Task B Source Map - Investment Philosophy

Research for T0593 was conducted through 2026-07-22. This is the exact 25-URL evidence set used in investment-philosophy.md; it ranks current regulatory and fund records, ARK's primary process and direct Wood material, independent manager research, then academic priors.

1. ARK Form ADV Parts 2A/2B, March 31, 2026 - Current primary evidence for Wood's roles, ownership, final authority, personnel, strategy assignments, conflicts, trade allocation, liquidity disclosures, and bounded disciplinary statements.
2. ARK ETF Trust prospectus, 2025 - Controlling ARKK mandate, 65% thematic floor, expected issuer range, fee and turnover figures, non-diversification, technology/issuer/liquidity risks, and ETF creation/redemption mechanics.
3. SEC ARKK semiannual report, January 31, 2026 - Primary point-in-time net assets, holdings count, sector weights, and six-month turnover; figures are fund-specific and not annualized or personal.
4. SEC ARKK annual report, July 31, 2025 - Primary one-, five-, and ten-year NAV returns and S&P 500 comparators at a specified endpoint.
5. ARK thematic investment process - Official framework for disruptive-innovation tests, six-part scoring, warning signs, weekly review, Wood's authority, and the dated 15% five-year expected-return flag.
6. ARK investment-process page - Current official chain from open top-down ideation and opportunity sizing through company modeling, conviction weighting, trimming, adding, and thesis-change sales.
7. ARK Form CRS, March 31, 2026 - Current adviser-filed summary of monitoring, fees, conflicts, standards, and reported disciplinary history; registration is not regulatory approval.
8. Big Ideas 2026 - Current organizational map of 13 converging innovation ideas; not evidence that every idea is Wood-authored or held by ARKK.
9. ARK ESG policy - Official scoring, engagement, governance and sector-exclusion boundaries, plus the five-year-return hurdle description.
10. ARK trade-notification disclosure - Primary limitation on daily transparency: files are unofficial and unreconciled and omit specified transaction categories.
11. ARK FAQ - Official score-review, risk-on/risk-off, and suggested investor-horizon statements; these are methodological guidance, not fund guarantees.
12. Wood, “Innovation Stocks Are Not in a Bubble,” 2021 - Direct benchmark, time-horizon, crisis-adoption, concentration and 30–40% strategy-forecast evidence with the issuer's later product-performance clarification.
13. Wood, “What the Market Overlooked in 2022,” 2023 - Direct post-drawdown defense of transparency, deflationary innovation and company evidence.
14. ARK Q1 2026 commentary - Manager-reported current-quarter ARKK and comparator results, used as a dated regime check rather than an audited personal record.
15. ARK Tesla 2029 model, 2024 - Analyst-authored one-million-simulation example whose limitation section discloses subjective inputs, data issues, positive-model bias, and presentation incentives.
16. ARK Tesla 2025 model, 2021 - Frozen matured forecast vintage used for one bounded calibration case, not a complete forecast scorecard.
17. Tesla split release, 2022 - Primary evidence for the three-for-one split needed to normalize ARK's 2025 targets.
18. Yahoo Finance chart endpoint, Tesla 2025 close - Machine-readable single market-data endpoint for the 2025-12-31 close; the forecast comparison remains one case.
19. TIME interview with Wood, 2022 - Direct evidence on correction behavior, concentration, analyst temperament, openness, and the five-year horizon.
20. Morningstar interview with Wood, 2025 - Current direct voice on volatility, scoring, changing AI exposure, regimes, and ARK's non-core allocation boundary.
21. Morningstar Australia process critique, 2021 - Historical independent criticism of analyst depth, aggregate risk, capacity, ownership, and key-person dependence; not treated as current staffing fact.
22. Morningstar manager report, April 2026 - Current independent assessment of team stabilization, succession, correlated growth exposure, liquidity, and style-consistent long-run comparison.
23. Reuters on Morningstar investor-loss estimates, 2024 - Independent dollar-weighted ARK-family and ARKK estimates, explicitly separated from NAV return and Wood-personal performance.
24. Ben-David et al., “Competition for Attention in the ETF Space” - Academic prior on specialized-ETF attention, launch timing and post-launch risk-adjusted returns; not an ARKK-specific test.
25. Bessembinder, “Do Stocks Outperform Treasury Bills?” - Long-run evidence on concentrated stock-market wealth creation, supporting both exceptional-winner search and prospective-selection difficulty.

Task B Evidence Limitations

• Wood leads and approves the process, but ARK's models, reports and decisions are organizational. Named-analyst research, ARK institutional publications, ETF portfolios and Wood's personal judgment are kept separate.
• ARKK, other active ETFs, index ETFs, SMAs, wrap accounts, private vehicles, non-discretionary model portfolios and portfolio companies have different mandates and records. Firm AUM, fund NAV, ETF flows, 13F holdings, investor dollar returns and Wood-personal performance are not interchangeable.
• ARK discloses six scores and qualitative review triggers but not the complete score-to-weight formula, current issuer-ownership ceiling, fixed stop-loss, maximum portfolio-drawdown rule, or full factor-risk budget.
• Five-year models depend on adoption, market share, cost, dilution, margins and terminal multiples. Public model revisions are not a frozen complete forecast ledger, and one matured miss cannot establish a universal forecasting rate.
• In-kind ETF redemptions can transfer securities without a forced portfolio sale. They do not remove cash-basket, AP, market-maker, underlying-liquidity, spread, capacity or investor-timing risk.
• The 2021 Morningstar staffing and ownership-policy observations are historical. The 2026 report notes team changes but does not eliminate Wood key-person or succession risk.
• Current adviser filings report no material disciplinary history, but those are adviser-filed and bounded. Public database searches cannot prove that no private, sealed, unindexed, or later matter exists.
• Performance figures are dated fund returns or independent estimates. No pre-ARK Wood composite, Wood-personal return series, complete holdings attribution, or full-cycle factor regression was recovered.

Task B Research Record

Three independent research lanes covered primary Wood/ARK philosophy and direct voice; operating process, fund mechanics, portfolio construction and regimes; and adversarial/current/legal, forecast-calibration, governance, performance and skill-versus-luck evidence. They performed 23, 27 and more than 60 discovery, document, database, passage and metric checks respectively, and each ended with exactly three deliberately dissimilar saturation searches followed by no further discovery. The final searches returned only already-covered process, filing, liquidity, performance, forecast, governance or false-positive legal material.

Exactly two further workstreams audit the frozen draft and its repaired exact-hash successor. The technical-lead audit reported P1=0, P2=2 and P3=1: two evidence-support gaps plus citation-locality cleanup. All findings were repaired by narrowing unsupported claims and repeating existing evidence at the relevant paragraphs and table rows, without expanding the source corpus. The fifth workstream is reserved for the final exact-hash regression. Reddit, Wikipedia, quote pages, unrelated entities, unverified social claims, ARKX-only documents, index-ETF mechanics misapplied to ARKK, and investee-company controversies misattributed to ARK were excluded.

Pre-audit endpoint QA returned 11 HTTP 200 responses, one HTTP 202 response, one access-controlled HTTP 401 response, eleven valid access-controlled HTTP 403 responses, and one known-live automation-resistant HTTP 406 response. Core ARK filings, process material, company-model disclosures, fund reports, Morningstar evidence and market data were independently inspected at the cited page, section, or endpoint level during the three research lanes.

Task A Source Map - Profile

Research for T0592 was conducted through 2026-07-23T02:07:03Z. This source map ranks the 25 core evidence sources used for the profile, with primary filings and fund reports first, then ARK primary materials, independent performance and critique sources, and finally false-positive legal/search-control sources. A stale-retry verification note below adds one current legal-boundary source without changing the original source hierarchy.

1. ARK Form ADV Parts 2A/2B, March 31, 2026 - Primary current source for Wood's legal name, birth year, education, employment chronology, ARK roles, approximately 72% ownership, supervision structure, AUM categories, conflicts, and disciplinary disclosure.
2. ARK Form CRS, March 31, 2026 - Adviser-filed current source for retail relationship summary, fee/conflict framing, and bounded legal/disciplinary disclosure.
3. SEC IAPD firm summary for ARK Investment Management LLC - Regulatory cross-check for ARK's adviser identity and registration status; registration is not an endorsement of skill.
4. ARK team page - Official biography for Wood's current role, Capital Group start, USC education, and public recognition claims.
5. ARK background FAQ, “What is Cathie Wood's background?” - Official plain-language biography that corroborates her education, Capital Group start, and ARK role.
6. ARK board-of-directors biography - Official chronology and AUM statements for Tupelo and AllianceBernstein, used with caveats because it is issuer biography.
7. SEC ARKK annual shareholder report, year ended July 31, 2025 - Primary fund report for ARKK one-, five-, and ten-year returns, net assets, holdings count, expense ratio, advisory fees, and turnover at a dated endpoint.
8. SEC ARKK semiannual shareholder report, period ended January 31, 2026 - Primary fund report for 2026 net assets, holdings count, expense ratio, sector weights, and six-month turnover.
9. ARK ETF Trust Q2 2026 quarterly report - Latest official performance table found during the run, used for ARKK one-, three-, five-, ten-year, and since-inception returns through 2026-06-30.
10. ARKK fund page - Official current product page for objective, active-management framing, risk caveats, ETF-market mechanics, and fund-level context.
11. ARK ETF Trust prospectus, 2025 - Controlling disclosure for active ETF mandate, expenses, non-diversification, thematic floor, issuer range, turnover, risk, and ETF creation/redemption mechanics.
12. ARK reports page - Official index of annual, semiannual, quarterly, prospectus, SAI, tax, and proxy reports; used to verify source recency and fund document hierarchy.
13. ARK investment-process page - Official current summary of top-down research, bottom-up modeling, open research, scoring, portfolio construction, and sell discipline.
14. ARK thematic investment process PDF - Official process detail for disruptive-innovation tests, score review, Wood's approval authority, and warning signs.
15. ARK FAQ - Official current source for time-horizon framing, scoring review, risk-on/risk-off language, and investor-fit caveats.
16. Big Ideas 2026 - ARK's current thematic map, useful for understanding the worldview behind Wood's public-market strategy but not proof of individual holdings or future returns.
17. Cathie Wood's 2026 Outlook - Current direct ARK/Wood market commentary, used as primary voice evidence rather than performance proof.
18. Institutional Investor, 2021 breakout-year profile - Secondary contemporaneous profile of ARK's 2020-2021 scale-up and public-market visibility.
19. InvestmentNews, ARK wipeout/AUM retrospective - Secondary evidence for the early-2021 active-ETF AUM peak range and later drawdown context.
20. ETF.com, ARK asset-flow retrospective - Secondary evidence for the ETF-family peak above $60 billion and later asset-base recovery attempts; used only with range caveats.
21. Reuters on Morningstar investor-loss estimates, 2024 - Independent report on Morningstar's dollar-weighted wealth-destruction estimates for ARK family and ARKK.
22. Markets Insider / Business Insider on Morningstar estimates, 2024 - Secondary corroboration of flow timing, ARK-family wealth destruction, ARKK-specific estimated loss, and 2022 drawdown context.
23. Morningstar manager report, April 2026 - Current independent assessment of ARKK's team, process, succession, concentration, liquidity, and style-consistent long-run comparison.
24. Morningstar Australia process critique, 2021 - Historical independent critique of analyst depth, risk controls, ownership and key-person dependence; preserved as historical, not automatically current, evidence.
25. Texas State Securities Board impersonation order, 2021 - Negative-control legal source: ARK was impersonated by a fraudster and was not the enforcement target.

Task A Evidence Limitations

• No audited Wood-personal composite, GIPS-style record, or continuous pre-ARK portfolio-return history was recovered. ARKK is a flagship public proxy, not a personal record.
• ARKK NAV returns, ARKK market-price returns, ARK-family ETF returns, firm AUM, ETF net assets, 13F holdings, client flows, and dollar-weighted investor outcomes are separate metrics and should not be merged.
• Peak AUM is a secondary-source range because sources variously define ARKK alone, active ARK ETFs, all ARK ETFs, or broader adviser assets.
• The primary birth evidence recovered is only Wood's 1955 birth year. Exact date and birthplace remain secondary-source biographical claims.
• Current ADV/CRS disciplinary statements are adviser-filed and bounded. Public web and database checks cannot rule out private, sealed, unindexed, or later legal matters.
• ARK biographies are useful for career chronology and assets managed but are issuer-controlled. Pre-ARK roles and AUM should be corroborated later from independent sources.
• Investee-company controversies and impersonation scams were excluded unless Wood, ARK Investment Management, or ARK ETF Trust was actually a party.

Task A Research Record

Five research lanes were used for T0592: biography/current status; vehicles, ownership and fund structure; track-record reconstruction; legal, regulatory and controversy checks; and source validation. The lanes prioritized current Form ADV/CRS filings, SEC fund reports, ARK primary documents, independent manager research, mainstream financial press, and adverse searches for enforcement or litigation. The work recovered enough primary evidence to write the required profile and to flag gaps without padding them.

Task A Stale Retry Verification - 2026-07-23

• Delaware Supreme Court Tesla/Musk compensation decision, 2025 - Current legal-boundary check: ARK appeared as an objector-appellant, not as an adviser-disciplinary or enforcement respondent; used to keep the profile's legal section narrow.

Task C Source Map - Greatest Trades

Research for T0594 was conducted through 2026-07-22. This is the exact 25-URL evidence set used in greatest-trades.md, ranked with regulatory holdings and fund-period accounting first, then contemporaneous research, corporate/regulatory evidence, and independent counterevidence.

1. SEC ARK ETF Trust semiannual report, January 31, 2015 - Earliest reviewed regulatory snapshot for Tesla and Nvidia across the original active ETFs; the semiannual schedule is not labeled as an audited annual report.
2. SEC ARK ETF Trust annual report, August 31, 2016 - Early Tesla, GBTC, Square and Nvidia holdings, costs, values, weights and contribution commentary.
3. SEC ARK ETF Trust annual report, July 31, 2018 - GBTC cost/value evidence, first reviewed CRISPR positions, Nvidia contribution, and risk disclosures.
4. SEC ARK ETF Trust semiannual report, January 31, 2019 - Pre-accumulation Tesla snapshot and reported weights across funds.
5. SEC ARK ETF Trust annual report, July 31, 2019 - Post-drawdown Tesla share-count and value snapshot used for endpoint comparison.
6. SEC ARK ETF Trust annual report, July 31, 2020 - Primary management discussion and schedules for Tesla, Square and the pandemic-year contribution record.
7. SEC ARK ETF Trust annual report, July 31, 2021 - Controlling peak-era holdings, contribution discussion, CRISPR fund-period realized gain, and first Palantir round.
8. SEC ARK ETF Trust annual report, July 31, 2022 - Risk-off holdings and contribution record, GBTC loss endpoint, and evidence that Palantir's first round had ended.
9. SEC ARK ETF Trust annual report, July 31, 2023 - GBTC cost/value loss endpoint, rebound contribution context, and Palantir re-entry.
10. SEC ARK ETF Trust annual report, July 31, 2024 - ARKB vehicle-period accounting, Block path, and first audited Roku realized-loss period.
11. SEC ARK ETF Trust annual report, July 31, 2025 - Palantir's top-contributor evidence, second Roku realized-loss period, and latest audited full-year schedules.
12. ARK ETF Trust portfolio schedule, April 30, 2026 - Current point-in-time holdings across active ETFs and affiliated-issuer period tables; not July 22 live holdings.
13. ARK, “The Electric Vehicle Market Will Grow Faster Than Expected,” 2014 - Earliest located dated Tesla thesis and adoption/cost framing.
14. ARK Tesla case study, 2015 - Primary organizational record of how Wood, analysts and external theme developers built the systems thesis.
15. ARK Tesla 2025 valuation model, 2021 - Frozen matured forecast vintage and robotaxi dependence; analyst-authored rather than Wood-only.
16. Tesla three-for-one stock-split release, 2022 - Primary corporate-action evidence required to normalize the 2025 targets.
17. Yahoo Finance Tesla chart endpoint, 2025 close - Single machine-readable market-data endpoint for the matured forecast comparison.
18. ETF.com on ARKW's first bitcoin exposure, 2015 - Contemporaneous independent entry, thesis, market and first-ETF-wrapper evidence.
19. ETF.com on the GBTC exit and BITO bridge, 2023 - Independent single-source date, approximate size and stated regulatory rationale for the vehicle transition.
20. ARK Square valuation research, 2020 - Primary Cash App, seller ecosystem, network-effect and financial-intermediation thesis.
21. ARK genomics commentary, December 2020 - Primary dated statement connecting CRISPR trial evidence to increased platform and company conviction.
22. FDA Casgevy approval, 2023 - Primary evidence for the first FDA-approved CRISPR/Cas9 therapy milestone.
23. Bloomberg Línea on ARKK's Nvidia exit, 2023 - Independent exit timing and subsequent rally; used with regulatory holdings rather than as lifetime P&L.
24. ARK Form ADV Parts 2A/2B, March 31, 2026 - Current Wood authority, vehicle/attribution and bounded adviser-filed legal disclosures.
25. Morningstar manager report, April 2026 - Independent full-cycle style comparison, factor/regime critique, liquidity and succession context.

Task C Evidence Limitations

• These are ARK ETF campaigns overseen by Wood, not Wood-personal trades. An analyst's model, fund holding, issuer return, fund contribution and Wood's individual decision are not interchangeable.
• “First reviewed” means the earliest recovered regulatory snapshot, not necessarily the exact first fill. Point-in-time share changes include trading and can also reflect fund creations, redemptions, mandate differences and corporate actions.
• Except for specified affiliated-issuer periods, complete tax lots, execution prices, realized gains, campaign IRRs, absolute P&L, and maximum drawdowns were not recovered. Fund AUM times return is not trade P&L.
• Filing-implied prices equal reported market value divided by shares. They measure the issuer at dated endpoints, not ARK's achieved return or intra-period maximum drawdown.
• GBTC, BITO and ARKB are distinct vehicles. GBTC premium/discount behavior and ARKB's affiliated/subadvised structure prevent one blended bitcoin-campaign return.
• Tesla's matured 2025 target miss is one forecast-calibration case. It neither erases the earlier trade nor scores open targets prematurely.
• Investee-company legal matters do not establish Wood or ARK misconduct. Current negative disciplinary statements are adviser-filed and bounded, not universal regulator clearance.

Task C Research Record

Three independent discovery workstreams covered Tesla and crypto; non-Tesla public equities and fund-period accounting; and adversarial ranking, current holdings, legal attribution, performance and skill-versus-regime evidence. They completed more than 45, more than 30, and well over 30 searches, filings, database, passage and arithmetic checks respectively. Each lane ended with exactly three deliberately dissimilar saturation searches and performed no later browsing. The frozen technical-lead audit returned P1=0, P2=4 and P3=2 for key-number labels, citation locality, Palantir's missing discovery record, excluded-trade chronology, semiannual-report wording and audit-state metadata; all six findings were repaired without expanding the 25-source corpus. The fifth-workstream CTO regression then returned P1=0, P2=6 and P3=0 for three missing discovery-process boundaries and three explicit return/drawdown unknowns; those findings were repaired, and the same workstream's full regression returned GO with P1=0, P2=0 and P3=0.

The lanes excluded Reddit, Wikipedia, generic guru lists, unsourced fund-size-times-return claims, search-result aggregators, investee controversies misattributed to ARK, and current holdings treated as realized success. Evidence that failed the full-cycle ranking—especially Roku, Zoom and Coinbase—was retained as counterevidence rather than silently discarded.

Post-repair endpoint QA returned six HTTP 200 responses and nineteen valid access-controlled HTTP 403 responses. All 25 sources had already been inspected at the filing, page, passage, or endpoint level during the three discovery lanes.

Task E Source Map - In Their Own Words

Research for T0596 was conducted through 2026-07-24. This is the exact 30-URL evidence set used in in-their-own-words.md: 29 underlying direct-voice materials plus one current adviser-filed attribution/legal boundary.

1. Wood, “A Different Interpretation of Recent Market Signals,” 2014 - Earliest recovered ARK-era Wood byline and the only source contributing two excerpts; their combined quoted text is 20 words.
2. Wood, “$1,700? Even Bitcoin's Bear Case Is Bullish,” 2018 - Named-authored bitcoin conviction through a severe drawdown and new-asset-class framing.
3. Wood, “Dear Elon,” 2018 - Signed open letter documenting Wood's five-year Tesla valuation and public-market-access argument.
4. Wood, “Addressing Concerns During Periods of Market Volatility,” 2018 - Direct written prescription to concentrate toward higher-conviction names during stress.
5. Wood, “The Coronavirus,” 2020 - Frozen crisis-era statement that disruption would gain share; used as dated voice rather than retrospective medical history.
6. Wood, “Investors Beware,” 2020 - Primary “right side of change” and creative-destruction framing.
7. Wood, “Innovation Stocks Are Not in a Bubble,” 2021 - Signed five-year valuation defense and pre-miss acknowledgment of possible inflation error.
8. Wood, year-end message, 2022 - Wood-bylined, signed section connecting the drawdown, fear, and crisis-created opportunity.
9. Wood, “What The Market Overlooked in 2022,” 2023 - Post-drawdown founder letter describing the transparent research ecosystem.
10. Wood, “The Journey From Monetary Shock To An Innovation-Led Economic Boom,” 2024 - Dated restatement of the deflation and productivity-led recovery thesis.
11. Wood, “The Recession Started Months Ago. Now What?” 2025 - Long-form argument for investing amid uncertainty rather than waiting for complete clarity.
12. Wood, 2026 outlook - Latest comprehensive bylined outlook, with named collaborators and bitcoin-supply framing.
13. Livewire interview, 2019 - Host-published direct quote on passive flows and the public-market innovation research gap.
14. Goldman Sachs, Talks at GS, 2020 - Official producer-selected excerpts on team diversity, mentors, and cross-disciplinary staffing.
15. SALT Talks #199, 2021 - Organizer-hosted interview and transcript on open research and exponential assumption error.
16. Morningstar debate with Rob Arnott, 2021 - Edited adversarial exchange on forward-looking valuation, profitability, and selling.
17. TIME interview, 2022 - Condensed and edited direct Q&A on analyst temperament, transparency, faith, and investor purpose.
18. Goldman Sachs, “Equity Bear Market: A Paradigm Shift?” 2022 - Official publication-grade interview on valuation, rates, profitability, and drawdown context.
19. CNBC Squawk Box, 2022 - Broadcaster-selected on-air admission that sustained inflation exceeded Wood's forecast.
20. Fortune interview, 2023 - Full edited interview on model challenge, AI, responsibility, and shareholder losses.
21. Morningstar interview on Rize, Nvidia, AI, and bitcoin, 2023 - Direct defense of ARKK's Nvidia exit and multi-platform construction.
22. Morningstar interview on Tesla, AI, and bitcoin, 2024 - Direct description of how ARK concentrates during a downturn.
23. Morningstar interview on DeepSeek and portfolio risk, 2025 - Current edited Q&A on cost curves, scoring, portfolio evolution, and investor fit.
24. Bloomberg ETF IQ interview, 2026 - Recorded current direct voice on liquidity and the existence of a succession plan.
25. Milken Global Conference transcript, 2026 - Official multi-speaker transcript on AI bubbles, valuation, SpaceX, and sell discipline; transcript-error warning preserved.
26. Wood, open letter to the Federal Reserve, 2022 - Primary falsifiable policy letter challenging a claimed inflation consensus.
27. Federal Reserve Payments Innovation Conference transcript, 2025 - Official panel transcript on AI agents, payments, stablecoins, and consumer-led adoption.
28. Institutional Investor interview, 2020 - Long profile/interview on crisis behavior, open research, criticism, and Tesla.
29. Inside the ICE House, episode 242, 2021 - Full host transcript on research openness, innovation platforms, and problem-driven opportunity.
30. ARK Form ADV Parts 2A/2B, March 31, 2026 - Current adviser-filed authority, attribution, and bounded disciplinary statements; not universal clearance.

Task E Evidence Limitations

• Wood's bylined ARK materials are edited primary writing, and institutional we remains ARK voice. Her CIO authority does not turn named-analyst research or every organizational forecast into Wood-personal authorship.
• Producer transcripts, edited Q&As, and selected broadcast quotes are labeled. Search snippets, automated transcripts without audio verification, derivative quote pages, and inaccessible recordings do not support exact excerpts.
• One underlying work has one copyright budget regardless of mirrors or formats. Every excerpt is 21 words or fewer, and combined quoted language from every underlying work is below 25 words.
• Direct conviction is not outcome evidence. The corpus preserves forecast dates, the realized inflation admission, Nvidia-sale defense, drawdown concentration, and the difference between an error possibility and a postmortem.
• No sole-authored Wood book, formal congressional testimony, Wood-signed SEC comment letter, complete pre-ARK writing archive, or complete official transcript archive for “In The Know” and FYI was recovered.
• Adviser-filed negative disciplinary statements are bounded. Investee-company disputes, an SEC filing, or an impersonation matter are not Wood misconduct without a direct evidentiary link.

Task E Research Record

Exactly four discovery workstreams covered ARK-hosted Wood-byline material; independent interviews, broadcasts, conferences and podcasts; the year-indexed primary-material bibliography and archive gaps; and adversarial provenance, forecast-error, current/legal, criticism and misattribution controls. They completed 58, 127, 110, and more than 30 substantive search, catalog, document, transcript, filing and passage checks respectively. Each lane ended with exactly three deliberately dissimilar saturation searches that found no new qualifying material, followed by no further discovery browsing.

The discovery lanes excluded quote aggregators, Wikipedia, Reddit, search snippets, analyst-authored models presented as Wood's words, institutional copy without a named speaker, duplicated interviews, automated transcript text without replay verification, and investee controversies misattributed to ARK. The fifth-workstream frozen audit returned P1=0, P2=1 and P3=0 for one ARK-model attribution leak; it was repaired, and same-workstream exact-hash regression returned GO with P1=0, P2=0 and P3=0.

Post-draft endpoint QA returned ten HTTP 200 responses, four HTTP 202 responses, fifteen valid access-controlled HTTP 403 responses, and one known-live automation-resistant HTTP 406 response. All 30 sources had already been inspected at the article, recording, transcript, filing, or passage level during the four discovery lanes.

Task F Source Map - Key Writings

Research for T0597 was conducted through 2026-07-24. This is the exact 24-URL evidence set used in key-writings.md: eleven Wood-bylined or explicitly collaborative writings, one multi-author book record, ten ranked works about Wood, one current adviser filing, and one independent forecast-outcome source.

1. Wood, “What We Do: A Letter to Investors From Cathie Wood,” 2024 - Richest post-drawdown defense of ARK's horizon, benchmark critique, valuation, Nvidia decision, and research and stock-selection errors; Wood byline with ARK data and institutional voice.
2. Wood, “Innovation Stocks Are Not in a Bubble,” 2021 - Compact statement of the five-year horizon, concentration in corrections, deflation view, inflation fallibility, and aggressive strategy-level return forecast.
3. Wood with named collaborators, “ARKK and Nasdaq 100: A Spurious Correlation,” 2022 - Post-drawdown distinction between price correlation and investment exposure; explicitly credits Sam Korus, Nick Grous, Dan White, and Julian Falcioni.
4. Wood with named collaborators, “Cathie Wood's 2026 Outlook,” 2026 - Current macro and innovation synthesis whose closing acknowledgments say Dan, Will, Katie, and Keith helped craft the letter.
5. Wood, “What the Market Overlooked in 2022,” 2023 - Post-drawdown letter on ARK's open research system, converging platforms, and operating evidence; not independent forecast validation.
6. Wood with named collaborators, “Disruptive Innovation and Profitability,” 2022 - Accounting and reinvestment defense co-authored with Tasha Keeney, Thomas Hartmann-Boyce, Julian Falcioni, and Paul Wilson.
7. Wood, “Investors Beware,” 2020 - Clearest short statement of creative destruction, incumbent value traps, convergence, and benchmark risk; contains the later-tested 2025 EV forecast.
8. Wood, “Broad-Based Benchmarks Seem to Be Short Disruptive Innovation,” 2022 - Wood-bylined institutional argument for benchmark agnosticism and active share.
9. Wood, “Dear Elon,” 2018 - Canonical signed open letter on public-market access and five-year valuation; the page credits Brett Winton, Tasha Keeney, and Sam Korus with leading the models.
10. Wood, “ARK Extends an Open Letter to the Fed,” 2022 - Dated, falsifiable policy argument using commodity, inventory, currency, and labor evidence; not a filed regulator comment.
11. Wood, “A Different Interpretation of Recent Market Signals,” 2014 - Earliest recovered ARK-era Wood byline and origin point for her contrarian deflationary-boom framework.
12. Tyndale, Faith Driven Investing - Publisher record for the eighteen-author volume and Wood's described contribution; public material does not isolate a verified chapter title, page span, or full Wood-only text.
13. Morningstar Managed Investment Report, 2026 - Best current independent diligence on Wood's centrality, personnel, succession, process, liquidity, correlated exposure, exits, and full-period comparative performance.
14. Phillips, “God, Money, YOLO: How Cathie Wood Found Her Flock,” 2021 - Best long biography of career, faith, founding, staffing, public research, and retail following; Wood declined an interview and the endpoint is pre-collapse.
15. Arnott, “15 Funds That Have Destroyed the Most Wealth Over the Past Decade,” 2024 - Flow-adjusted investor-outcome analysis whose dollar estimate is not Wood-personal P&L, simple NAV return, or evidence of misconduct.
16. Gura, “A Superstar Investor With the Midas Touch or Just Lucky?” 2022 - Boom-to-bust profile testing skill, rate sensitivity, fandom, and Wood's five-year defense; includes economically interested critics.
17. Liu and Norton, “ARK's Cathie Wood Disrupted Investment Management,” 2021 - Strong peak-era synthesis of career, ETF innovation, research architecture, staffing, faith, and audience design; requires later counterweights.
18. Gara, “How Cathie Wood Beat Wall Street by Betting Tesla Is Worth More Than $1 Trillion,” 2020 - Early account of ARK's economics, ownership, team model, Tesla thesis, and Wood's prior career; private-value estimates are not audited.
19. Zhang, “Why the Michigan Retirement System Is Still Bullish on Cathie Wood,” 2022 - Rare allocator perspective; the pension's reported result is dated, strategy-specific, and not a Wood-personal composite.
20. Whyte, “Cathie Wood Still Thinks Tesla Is Going to $6,800. Why?” 2020 - Contemporaneous process and crisis-behavior record whose targets require split normalization and outcome testing.
21. Luscombe, TIME Q&A, 2022 - Condensed and edited interview on transparency, team temperament, faith, women in finance, and investor purpose.
22. Celarier, “The Passion of Cathie Wood,” 2022 - Post-collapse account of brand, faith, forecasts, critics, and persistence; hard-paywall access and critics' economic interests are explicit limitations.
23. ARK Form ADV Parts 2A/2B, March 31, 2026 - Current adviser-filed authority for Wood's roles, ownership, career and bounded negative disciplinary statements; not universal regulator clearance.
24. IEA, Global EV Outlook 2026 - Independent primary outcome source showing just over 20 million EV sales and 25% share in 2025 versus the 2020 letter's roughly 40 million and 45% forecast.

Task F Evidence Limitations

• Wood is a prolific named letter writer, not a verified sole-author book or research-paper author. A byline establishes published responsibility, not unassisted drafting or personal origination of institutional data.
• Explicit collaborator credits are preserved. ARK analyst reports, valuation models, Big Ideas, FAQs, prospectuses, fund reports, videos, and corporate copy remain team or named-analyst work unless a Wood-authored section is identified.
• Faith Driven Investing is an eighteen-author volume. Public publisher and catalog material establishes a contribution by Wood but not a verified chapter title, page span, or sufficient text for a five-idea close reading.
• Direct writings communicate thesis and process; they do not prove performance. The EV miss, strategy-return clarification, open 2026 outlook, and current independent process criticism keep forecast and outcome separate.
• The source set mixes manager research, quantitative analysis, journalism, interviews, and an adviser filing because no serious independent biography or completed evidence-grade documentary cleared the source bar.
• Paywalls, dynamic reports, mutable ARK pages, and edited Q&As are labeled. No inaccessible passage was made to support a claim that could not be checked from available metadata or corroborating material.
• No pre-ARK Jennison, Tupelo, or AllianceBernstein writing archive, formal Wood testimony, Wood-filed regulator comment, complete personal forecast ledger, or Wood-personal audited composite was recovered.
• The adviser-filed disciplinary response is bounded. Public comments, routine product orders, impersonation cases, and investee-company disputes are not findings against Wood or ARK.

Task F Research Record

Exactly four discovery workstreams covered Wood's authored bibliography and authorship grades; the ARK-hosted corpus, version history and forecast outcomes; the best independent works about Wood; and edited/current/legal/adversarial evidence. They completed 51 web searches plus roughly 40 catalog, page and document checks; 34 searches plus more than 45 primary-page, PDF, byline, version and outcome checks; more than 80 search, catalog, document, passage and metadata checks; and 27 searches plus more than 35 page, filing, book and passage checks, respectively.

Each lane established guiding questions before searching and ended with exactly three deliberately dissimilar terminal saturation searches that found no new qualifying work, authorship fact, version-history fact, serious biography, evidence-grade documentary, enforcement matter, or pre-ARK/formal writing. No lane browsed after its final battery. The lanes excluded self-published low-source biographies, an unfinished AI/community documentary, quote lists, snippets, forums, analyst/team work attributed to Wood, routine SEC product mechanics presented as enforcement, third-party public comments presented as Wood filings, and investee controversies presented as Wood misconduct.

The fifth-workstream frozen audit confirmed the initial artifact hashes and returned P1=0, P2=2 and P3=0 for two citation-locality gaps. Both were repaired by placing the Wood forecast source and the Morningstar criticism source at the relevant claims. Same-workstream regression returned P1=0, P2=0 and P3=1 for wording that mistakenly placed the adviser filing on the ranked works-about shelf; the label was repaired to describe the broader source set. Post-draft endpoint QA returned nine HTTP 200 responses, one valid paywall HTTP 401 response, thirteen valid automation- or access-controlled HTTP 403 responses, and one known-live automation-resistant HTTP 406 response across the exact 24-source set.

Task G Source Map - Mental Models

Research for T0598 was conducted through 2026-07-24. This is the exact 25-URL evidence set used in mental-models.md, ranked from current regulatory and operating evidence through direct Wood material, forecast outcomes, historical process versions, independent criticism, academic context, and retail-risk controls.

1. ARK Form ADV Parts 2A/2B, March 31, 2026 - Current primary evidence for Wood's authority, team and Theme Developer structure, conflicts, controls, ownership, succession exposure, and bounded disciplinary statements.
2. ARK ETF Trust prospectus, 2025 - Controlling ARKK mandate and non-diversification, underlying-liquidity, market-maker, Authorized Participant, cash-basket and large-shareholder risk.
3. ARK thematic investment process - Current three-part disruption test, Wright's Law and elasticity chain, six-part tracker, warning signs, 15% valuation flag, meetings, open ecosystem and Wood authority.
4. ARK investment-process page - Current operational chain from top-down opportunity sizing to bottom-up five-year modeling, relative-opportunity adds and trims, and thesis-change sales.
5. ARK FAQ - Current 6/10 review trigger, benchmark and regime language, qualitative sell discipline, and name-versus-share-turnover distinction; not a mechanical trading rule.
6. ARK ETF Trust Q2 2026 report - Official point-in-time top-ten concentration and Tesla weight used as an observation rather than a fixed limit.
7. Wood, “Cathie Wood's 2026 Outlook,” 2026 - Current Wood-bylined institutional expression of the five-platform convergence map; named collaborators and open forecasts remain bounded.
8. Big Ideas 2026 - Current team-authored platform and convergence map; not Wood-personal authorship or an ARKK holdings list.
9. ARK active-ETF position-weight guidance - Current operating norms for top-ten, median, minimum and maximum-purchase weights; “typical” does not create a hard prospectus cap or retail allocation.
10. ARK investor-horizon guidance - Current full-cycle or seven-plus-year end-investor horizon, distinguished from the five-year company-model horizon.
11. ARK Q4 2025 webinar summary, 2026 - Current institutional account of score-based construction, multiple compression, downturn consolidation and bull-market broadening.
12. ARK, “ARKK Is Not a Tech ETF,” 2026 - Current issuer-reported holdings count, active share, benchmark overlap and rate sensitivity; useful structural data with an advocacy incentive.
13. ARK trade-notification disclosure - Primary boundary on trade copying: end-of-day files are unofficial, unreconciled and incomplete and omit specified transactions.
14. Morningstar interview with Wood, 2025 - Direct current voice on volatility, scoring, portfolio change, risk and ARK's non-core role in an adviser client's portfolio.
15. Morningstar Managed Investment Report, 2026 - Current independent test of score execution, correlated exposure, difficult exits, risk hires, key-person dependence and succession.
16. Wood, “What We Do,” 2024 - Direct acknowledgment of research and stock-selection errors, Nvidia decision-making, concentration and post-drawdown conviction.
17. Wood, “Investors Beware,” 2020 - Direct creative-destruction and value-trap framework plus the later-tested 2025 electric-vehicle forecast.
18. IEA, Global EV Outlook 2026 - Independent primary outcome for 2025 electric-vehicle sales and share, testing direction against calibration.
19. ARK investment-process document, 2015 - Historical eight-metric framework that prevents the current six-part tracker from being presented as timeless.
20. Wood AllianceBernstein interview, 2010 - Pre-ARK direct interview documenting continuity in thematic plus bottom-up “classical” investing; interviewer framing is kept separate.
21. ARK Tesla valuation-model repository - Analyst-authored open model vintages and assumptions; useful for public stress-testing and for demonstrating version-ledger risk.
22. Morningstar, “Why ARKK Shareholders Are Still Underwater,” 2023 - Independent time-weighted-versus-dollar-weighted investor-return analysis with explicit ETF-flow estimation limits.
23. Ben-David, Franzoni and Kim, “Competition for Attention in the ETF Space” - Academic prior on specialized-ETF launch timing, attention and post-launch performance; not an ARKK-specific causal estimate.
24. Investor.gov, Asset Allocation and Diversification - Regulator investor-education boundary that a narrowly focused ETF is not automatically diversified.
25. FINRA, Concentration Risk - Retail control for look-through overlap, correlated exposures, rebalancing and underlying liquidity.

Task G Evidence Limitations

• The public record contains named ARK concepts but no canonical Wood-authored checklist. Section labels and the operational checklist are explicit reconstructions.
• Wood has final accountability; ARK analysts and collaborators originate many models and publications. An organizational method or result is not silently converted into Wood-personal work or performance.
• The scoring process changed from eight public metrics in 2015 to six current tracker areas. The historical and current versions are dated rather than merged into a timeless rulebook.
• Five years is the company-model horizon; current guidance asks end investors for a full cycle or at least seven years. Neither horizon substitutes for dated intermediate falsification.
• Current help-center weights, issuer ranges and turnover figures are descriptive and vehicle-specific. They are not fixed household allocations, universal hard caps, or complete risk limits.
• No complete score history, score-to-weight equation, CIO-override log, frozen forecast ledger, fixed stop, maximum drawdown, full factor budget, or current issuer-ownership and days-to-liquidate ceiling was recovered.
• In-kind ETF mechanics can reduce forced selling by the fund but do not eliminate underlying liquidity, spread, capacity, market-maker, cash-basket, cross-vehicle or investor-timing risk.
• Performance is endpoint- and benchmark-sensitive. ARKK NAV, ARK-family investor-dollar estimates, firm AUM, public forecasts and Wood-personal economics remain separate.
• Current adviser filings report no material disciplinary history, but those are adviser-filed and bounded. Routine product orders, public comments, impersonation cases and investee disputes are not Wood enforcement events.

Task G Research Record

Exactly four discovery workstreams covered named heuristics and framework provenance; operational screens, scoring, sizing, selling, fund controls and governance; adversarial failure modes, forecast calibration, performance, liquidity, succession, conflicts and legal status; and individual-investor transferability, trade-copying limits and independent retail controls.

They completed 47 search queries plus 57 full-page, PDF and passage checks; 23 searches plus 65 page, PDF and passage checks; 35 distinct searches plus more than 60 page, filing and passage checks; and 27 searches plus more than 30 page, passage, document and local cross-checks, respectively. Each lane stated guiding questions before searching and ended with exactly three deliberately dissimilar terminal saturation searches that yielded no new qualifying rule, framework version, forecast ledger, risk limit, enforcement matter, replication study or individual checklist, followed by no later browsing. One lane correctly rejected two earlier proposed saturation batteries and resumed research when they surfaced Wood's 2010 process interview and ARK's 2015 eight-metric framework.

The lanes excluded Wikipedia, Reddit, quote lists, self-help books, search snippets, derivative summaries, unrelated ARK namesakes, product mechanics from other vehicles presented as ARKK rules, analyst work attributed to Wood, fund outcomes presented as personal results, routine SEC orders presented as enforcement, and portfolio-company disputes presented as Wood misconduct.

The fifth workstream confirmed the initial frozen chapter and merged-source hashes and returned P1=0, P2=1 and P3=1. Its two bounded findings were an exact point-in-time position-weight observation without the required single-source flag and a weight-norm statement whose already-mapped citation was not local; both were repaired without expanding the evidence set. Post-draft endpoint QA returned fourteen HTTP 200 responses, two HTTP 202 responses and nine access-controlled HTTP 403 responses across the exact 25-source set. The same fifth workstream is the final exact-hash regression gate.

Task H Source Map - Synthesis

Research for T0599 was conducted through 2026-07-24. This is the exact 25-URL evidence set used in synthesis.md, ranked from current regulatory and operating evidence through manager research, independent outcome tests, attribution studies, and retail controls.

1. ARK Form ADV Parts 2A/2B, March 31, 2026 - Current adviser-filed evidence for Wood's authority, organization, conflicts, ARK Capital Markets, succession exposure, and bounded disciplinary statements.
2. ARK ETF Trust prospectus, 2025 - Controlling ARKK mandate and risk evidence for non-diversification, underlying liquidity, Authorized Participants, market makers, baskets, and large shareholders.
3. ARK thematic investment process - Current primary statement of the disruption screen, cost and demand logic, six-area Portfolio Tracker, valuation flag, review process, and Wood's final authority.
4. ARK investment-process page - Current operating chain from top-down opportunity sizing through bottom-up five-year modeling, relative-opportunity transactions, and thesis-change sales.
5. ARK ETF Trust Q2 2026 report - Official performance, expense, holdings, and concentration evidence through June 30, 2026; issuer-reported and endpoint-sensitive.
6. Morningstar Managed Investment Report, 2026 - Strongest current independent diligence on performance, correlated exposure, difficult exits, risk staffing, key-person dependence, and succession.
7. AllianceBernstein, thematic-equities presentation, 2011 - Bounded pre-ARK team-composite evidence identifying Wood as thematic CIO; pure-gross and maximum-fee results are not an audited Wood-personal record.
8. ARK Form CRS - Current adviser-filed relationship, conflict, fee, and bounded disciplinary disclosure; not universal regulatory clearance.
9. ARK FAQ - Current issuer account of benchmark independence, qualitative sell discipline, review triggers, turnover, and risk-off weakness; descriptive rather than a complete ruleset.
10. Wood, “What We Do,” 2024 - Direct post-drawdown acknowledgement of research and stock-selection errors, the Nvidia decision, concentration, and continued conviction.
11. Wood, Q2 2026 market commentary - Current issuer-reported quarterly rebound and regime interpretation; not evidence of full-cycle alpha or resolved forecasts.
12. Big Ideas 2026 - Current team-authored five-platform and convergence map; not Wood-personal authorship or a holdings list.
13. ARK Tesla valuation-model repository - Primary version history for a major campaign model and frozen-forecast lesson; not a complete ARK forecast or trade ledger.
14. IEA, Global EV Outlook 2026 - Independent primary 2025 EV-sales outcome used to test ARK's dated 2020 forecast.
15. Wood, “Investors Beware,” 2020 - Direct statement of creative destruction, convergence, benchmark risk, and the later-tested 2025 EV forecast.
16. Rompotis, ARKK factor study, 2022 - Peer-reviewed six-factor result for 2018–21; its short sample and pre-collapse endpoint limit current inference.
17. Sparkline Capital, The Investing Revolution, 2022 - Practitioner innovation-factor and residual-alpha model; proprietary construction, simulation, and endpoint make it a bounded countermodel.
18. Morningstar, “Why ARKK Shareholders Are Still Underwater” - Independent flow-timing evidence separating investor-dollar experience from the fund's time-weighted return.
19. Ben-David, Franzoni, and Kim, NBER working paper - Broad thematic-ETF evidence on specialized narratives, flows, fees, and post-launch outcomes; a prior, not an ARKK causal test.
20. ARK trade-notification disclosure - Primary warning that end-of-day trade files are unofficial, unreconciled, incomplete, and unsuitable as a complete replication feed.
21. ARK active-ETF position-weight guidance - Descriptive portfolio-weight norms; neither a hard prospectus limit nor a household-allocation recommendation.
22. ARK investor-horizon guidance - Current full-cycle or seven-plus-year end-investor guidance, distinct from the five-year company-model horizon.
23. FINRA, concentration-risk guidance - Independent retail control for sizing correlated themes outside the manager's narrative.
24. Investor.gov, asset-allocation guidance - Independent retail guidance on horizon, risk tolerance, allocation, and diversification.
25. Bloomberg ETF IQ interview with Wood, February 2026 - Current Wood assertion that a succession plan exists; no named successor or transfer mechanics were recovered.

Task H Evidence Limitations

• A standalone mistakes-and-losses chapter is incomplete. This synthesis therefore uses adverse evidence from the other six completed chapters and current sources but cannot claim a complete loss ledger.
• Wood, ARK, ARKK, other ARK vehicles, research teams, fund investors, and Wood's personal economics are separate units. Public vehicle performance is not a Wood-personal composite.
• The AllianceBernstein figures are a team and institutional composite presented by the manager. Firm-level GIPS verification did not verify that composite specifically, and pure-gross and maximum-fee returns are not audited Wood-personal results.
• The factor evidence is model- and endpoint-dependent. One peer-reviewed six-factor study and one proprietary practitioner model disagree and neither covers a complete current cycle.
• Published weights, return ranges, review triggers, and horizons are descriptive and vehicle-specific unless a controlling document says otherwise. They are not universal hard risk limits or household allocations.
• Current legal and disciplinary statements are bounded to retrieved adviser filings and searches through the research date. ARK Capital Markets' later registration, engagements, and associated restrictions remain unresolved.
• No complete forecast ledger, score-to-weight history, CIO-override log, tax-lot campaign P&L, holdings-level factor attribution, liquidity-capacity record, or audited pre-ARK or Wood-personal composite was recovered.

Task H Research Record

Exactly four discovery workstreams covered the 500-word executive brief and ranked transferable lessons; style taxonomy, regime dependence, factor attribution and failure conditions; closest and most-opposite Canon investors plus current legal, governance and succession evidence; and final cross-chapter synthesis, source parity and unresolved questions.

They completed 24 searches plus at least 70 source and passage checks; 31 searches plus at least 23 documents and more than 45 source checks; 15 searches plus 27 external checks and 20 local-document checks; and 18 searches plus more than 30 page, source and local checks, respectively. Each lane stated guiding questions before searching and ended with exactly three deliberately dissimilar terminal saturation searches that yielded no new qualifying fact, comparison, enforcement matter, current operating rule, factor result, forecast ledger, or pre-ARK evidence, followed by no later discovery browsing.

The executive lane correctly reopened its first stopping pass when it recovered the AllianceBernstein team composite. The factor lane integrated two conflicting attribution models before its terminal battery. The root lane correctly reopened its first proposed final battery when it surfaced Wood's Q2 2026 commentary, then completed a replacement battery. The lanes excluded Wikipedia, Reddit, quote lists, search snippets, unsupported personal-performance claims, analyst work attributed to Wood, fund outcomes presented as personal results, routine product mechanics presented as enforcement, and investee-company disputes presented as Wood misconduct.

Post-draft endpoint QA returned sixteen HTTP 200 responses, one HTTP 202 response, and eight valid access-controlled HTTP 403 responses across the exact 25-source set. The fifth workstream confirmed the initial frozen hashes and returned P1=0, P2=1, and P3=2 for citation-locality gaps, one numeric single-source flag, and one uncited empirical-support paragraph. All findings were repaired without new evidence or URLs by narrowing claims, placing already-mapped citations locally, and labeling the issuer-reported statistic. The same fifth workstream is the final exact-hash regression gate.

Task D Source Map - Mistakes and Losses

Research for T0595 was conducted through 2026-07-24. This is the exact 32-URL evidence set used in mistakes-and-losses.md, ranked from audited and regulatory loss records through direct Wood/ARK explanations, issuer outcomes, forecast tests, independent criticism, governance, process, and current conflict evidence.

1. AllianceBernstein thematic-equities presentation, 2011 - Bounded pre-ARK team-composite record identifying Wood as thematic CIO; maximum-fee and pure-gross figures are not Wood-personal returns.
2. Tupelo Capital Management amended Form 13F, March 31, 2000 - Contemporaneous long-exposure evidence near the dot-com peak; not AUM, performance, a complete portfolio, or a Wood-personal record.
3. SEC ARKK calendar-return table, 2025 - Primary calendar NAV return sequence for 2020–24, kept separate from fiscal-period accounting and investor-dollar outcomes.
4. SEC ARK ETF Trust annual report, July 31, 2022 - Audited fiscal drawdown, net-asset decomposition, realized/unrealized accounting, fund returns, and early security-loss paths.
5. SEC ARK ETF Trust annual report, July 31, 2024 - Audited affiliated-security realized-loss ledger for Teladoc, Roku, Ginkgo, Invitae, UiPath, and 2U.
6. SEC ARK ETF Trust annual report, July 31, 2025 - Second bounded Roku realized-loss period, open-holding endpoint, and exit checks.
7. ARK ETF Trust Q2 2026 report - Current official five-, ten-year, and since-inception ARKK and S&P 500 performance through June 30, 2026.
8. Morningstar Managed Investment Report, April 2026 - Strongest current independent critique of correlated exposure, sizing, exits, liquidity, risk hires, key-person dependence, and succession.
9. Reuters on Morningstar investor-loss estimates, 2024 - Independent dollar-weighted ARK-family and ARKK estimates, explicitly distinct from NAV return and personal P&L.
10. CNBC on Wood's inflation admission, 2022 - Contemporaneous direct statement that ARK was wrong about the duration of inflation.
11. TIME interview with Wood, 2022 - Direct post-2021 explanation of rotation, concentration, correction behavior, conviction, and the five-year horizon.
12. Wood, “What the Market Overlooked in 2022,” 2023 - Direct macro and innovation-thesis defense after the drawdown; not a formal process postmortem.
13. Wood, investor letter, 2024 - Direct acknowledgement of macro and stock-selection challenges, Nvidia, concentration, and continued conviction.
14. ARK Q2 2024 commentary - Manager-reported Ginkgo, UiPath, and Block operational disappointments and continuing thesis responses.
15. ARK, “Disruptive Innovation and Profitability,” 2022 - Contemporaneous defense of Invitae and other unprofitable holdings before later outcomes.
16. Teladoc 2024 results - Issuer-primary revenue, net-loss, and BetterHelp-impairment evidence.
17. Invitae bankruptcy Form 8-K, 2024 - Primary Chapter 11 date and proceeding evidence.
18. 2U Chapter 11 announcement, 2024 - Issuer-primary restructuring and bankruptcy evidence.
19. Ginkgo reverse-split Form 8-K, 2024 - Primary one-for-40 reverse-split evidence after operational deterioration.
20. ARK Roku 2026 model, 2022 - Frozen expected-value and adoption thesis whose 2026 horizon was not complete at the research cutoff.
21. Bloomberg Línea on ARKK's Nvidia exit, 2023 - Independent exit timing used to define an unquantified omission rather than fabricated P&L.
22. ARK Tesla 2025 valuation model, 2021 - Frozen base/bear targets and robotaxi dependence for one matured calibration case.
23. Tesla three-for-one split release, 2022 - Primary corporate-action evidence needed to normalize ARK's targets.
24. Yahoo Finance Tesla chart endpoint, 2025 close - Machine-readable single market-data endpoint for the 2025-12-31 close.
25. Wood, “Investors Beware,” 2020 - Direct 2025 global-EV adoption forecast tested against the later outcome.
26. IEA, Global EV Outlook 2026 - Independent primary 2025 EV-sales and market-share outcome.
27. SEC ARK ETF Trust Resolute supplement, 2020 - Primary control-option, assignment, automatic-termination, and approval-risk disclosure.
28. ARK/Resolute resolution statement, 2020 - Primary resolution preserving Wood control and extinguishing the option for an undisclosed agreed value.
29. ARK investment-process page - Current weekly monitoring, portfolio construction, relative-opportunity transaction, and thesis-change sale description.
30. ARK research reorganization announcement, 2022 - Contemporaneous team and role expansion after the main drawdown.
31. ARK Form ADV Parts 2A/2B, March 31, 2026 - Current Wood authority, ownership, ARK Capital Markets conflict, AUM, and bounded adviser-filed disciplinary statements.
32. Eightco Form 8-K, May 2026 - Primary current fee, warrant, capitalization-bonus, and board-advisory compensation terms; no illegality or client-overlap inference is made.

Task D Evidence Limitations

• Wood, ARK, its research team, ARKK, other vehicles, fund investors, and Wood's personal economics are separate units. No audited Wood-personal composite was recovered.
• Calendar return, fiscal return, AUM change, realized loss, unrealized mark, investor wealth destruction, issuer return, and opportunity cost answer different questions.
• Affiliated-security figures are bounded fund-period accounting. They cannot be spliced into lifetime campaign P&L or extrapolated across ARK vehicles.
• AllianceBernstein's record is a manager-presented team composite; Tupelo's 13F is an incomplete long-exposure snapshot; Jennison performance attributable to Wood remains unavailable.
• Forecast tests use frozen, matured vintages. Roku's 2026 horizon was not complete at the cutoff, and no open target was scored prematurely.
• Public method documents do not disclose a complete score history, model-vintage ledger, CIO-override log, fixed stop-loss, maximum-drawdown limit, full factor budget, or independent risk veto.
• Current adviser-filed disciplinary statements and bounded public searches are not universal legal clearance. Investee-company bankruptcies and lawsuits are not ARK adviser misconduct.
• The ARK Capital Markets engagement creates a disclosed economic-conflict case but does not establish illegality, client overlap, completed broker-dealer registration, or investment harm.

Task D Research Record

Exactly four discovery workstreams covered fund-loss and investor-dollar accounting; company failures, realized-loss ledgers and forecast misses; Wood's direct responses, behavioral roots and process changes; and pre-ARK performance gaps, governance near-death, current conflicts and legal false positives.

They completed more than 60 searches plus extensive filing and arithmetic checks; 67 searches plus dozens of filing, model and issuer checks; 51 searches plus more than 70 source and passage checks; and 63 searches plus at least 59 additional source and database checks, respectively. Each lane stated guiding questions before searching, ended with exactly three deliberately dissimilar saturation searches that yielded no new qualifying fact, record, admission, process change, enforcement matter, or missing performance series, and performed no later discovery browsing.

The lanes excluded Wikipedia, Reddit, quote lists, unsourced personal-loss claims, AUM-times-return arithmetic, investor-dollar estimates presented as NAV return, fund-period realized losses presented as lifetime campaigns, open forecasts scored early, investee-company disputes presented as adviser misconduct, and unrelated or impersonating “Ark” entities.

Pre-audit endpoint QA returned nine HTTP 200 responses, one valid paywall HTTP 401 response, 21 valid automation- or access-controlled HTTP 403 responses, and one known-live automation-resistant HTTP 406 response across the exact 32-source set. The fifth-workstream frozen audit confirmed the initial hashes and returned P1=0, P2=5, and P3=3 for incomplete single-source labels, citation-locality and support gaps, an overbroad legal-search conclusion, the Wood–Tupelo connection, a mislabeled citation, an unbounded superlative, and one uncited empirical takeaway. All findings were repaired without new evidence or URLs. Same-workstream regression returned GO with P1=P2=P3=0, exact 32-to-32 URL parity, complete Task D coverage, correct quote and word limits, and preserved fund/person, calendar/fiscal, realized/unrealized, open-forecast, and current/legal boundaries.