Cathie Wood's AI Stock Picks and the Deregulation Bet

Cathie Wood has made the regulatory premise unusually explicit: her X bio says “government regulation is the risk,” and in recent interviews and profiles she has identified regulation as the primary threat to AI innovation while praising the Trump administration’s deregulatory turn as “astonishing.”[1] That matters more than the familiar performance debate around ARK funds. If the investment case is tied to deregulation, then the legal question is not whether AI is “early” or “overhyped.” It is which legal constraint is supposed to loosen, for which company, and under whose authority.

The sequence since early 2025 gives the thesis its shape. In January 2025, the administration rescinded President Biden’s AI executive order, removing the prior federal risk-management architecture as the starting point for national AI policy.[1] In December 2025, a new executive order created an AI Litigation Task Force and directed federal pressure against state AI laws viewed as obstructive.[2] In March 2026, the National AI Policy Framework proposed broad federal preemption of state AI rules, though its legislative path remained uncertain because of Democratic opposition and a narrow House Republican majority.[3] On June 2, 2026, Executive Order 14409, “Promoting Advanced AI Innovation and Security,” added a later federal statement in favor of AI innovation and security.[4]

That is a policy pattern, not a single deregulatory event. It also does not erase the difference between an autonomous-vehicle safety file, a defense procurement review, a medical-device oversight question, a crypto classification fight, and an export-control license analysis. Wood’s AI basket sits across all five.

The Portfolio Is Concentrated Enough for the Legal Thesis to Matter

As of July 16, 2026, Motley Fool’s review of ARK daily trade data put 30.4% of Wood’s portfolio across five AI-related names: Tesla at 9.73%, AMD at 8.1%, Tempus AI at about 5%, Coinbase at about 5%, and Palantir at about 4.2%.[5] Those percentages should not be treated as permanent; ARK’s daily trading makes that a poor compliance habit. They are useful because they show that the regulatory thesis is not decorative. It is attached to a meaningful portion of the portfolio.

The temptation is to treat this as one AI trade. For diligence purposes, that is too blunt. The same federal mood that benefits one holding may leave another exposed to state enforcement, sector-specific safety rules, procurement conditions, or export-control limits.

Tesla: Deregulation Still Has to Pass Through Safety Law

Tesla is the cleanest example of why “less regulation” is not a complete legal analysis. Wood’s robotaxi thesis is explicit and large: she has estimated that 86% of Tesla’s 2029 earnings will come from robotaxis.[1] That estimate makes autonomous-vehicle authorization, safety reporting, and state operating rules central to the investment case rather than incidental compliance issues.

A federal deregulatory posture may affect the tone and timing of autonomous-vehicle policy, especially around national standards. But robotaxi deployment still runs through the National Highway Traffic Safety Administration framework and state-by-state approval regimes.[1] That means a legal team assessing Tesla exposure should separate three questions that often get blurred in market commentary: whether the technology works, whether federal safety regulators tolerate the deployment model, and whether states allow commercial operation at scale.

The preemption push is relevant here, but it does not automatically settle the operating environment. The March 2026 National AI Policy Framework recommended broad federal preemption, yet the legislative path remained contested.[3] Until federal preemption becomes concrete and durable, state-level rules remain part of the risk map. That is especially true for autonomous vehicles, where road safety, insurance, licensing, and local operating restrictions do not collapse neatly into a single AI-governance category.

For Tesla, the legal signal to watch is not simply whether Washington is friendlier to AI. It is whether autonomous-vehicle supervision becomes more nationally standardized, whether NHTSA’s posture toward robotaxi deployment changes in a measurable way, and whether states resist, supplement, or accept the federal line.

Palantir: The Regulatory Temperature Is Procurement, Not Consumer AI

Palantir deserves a different lens. Its legal exposure is less about mass-market AI products and more about government procurement, defense contracting, data access, agency adoption, and public-sector accountability. In Q2 2025, Palantir reported $1.7 billion in U.S. government revenue, up 68% year over year.[1] That figure places public-sector AI policy near the center of the company’s growth narrative.

A deregulatory administration can help a company like Palantir in several ways without repealing a single consumer AI rule. Agencies may be encouraged to deploy AI tools faster. Procurement officers may receive stronger political signals to modernize. Defense and intelligence customers may frame AI capability as a national-security priority rather than a governance problem. Executive Order 14409’s innovation-and-security framing is relevant for that reason.[4]

But government procurement creates its own constraints. Public-sector AI vendors still face security requirements, contract-specific restrictions, audit rights, data-handling obligations, budget scrutiny, and political oversight. The compliance burden may shift from general AI rulemaking to procurement documentation and agency-level use controls. For counsel reviewing Palantir as a vendor, the practical questions are likely to be more granular than “Is AI deregulated?” They include what data the system touches, which agency controls the deployment, what representations appear in the contract, and how the agency documents model use in sensitive decisions.

Palantir is therefore one of the holdings most likely to benefit from a federal AI acceleration agenda, but also one of the least served by loose references to deregulation. Its risk lives inside public procurement and government use, where political support can speed adoption while increasing scrutiny from inspectors general, legislators, advocacy groups, and disappointed bidders.

Tempus AI: Healthcare Oversight Does Not Disappear Because the Product Uses AI

Tempus AI sits at the intersection of AI, genomics, and healthcare. That makes it a healthcare oversight story before it is a generic AI story. The company’s regulatory environment includes FDA oversight of AI/ML-enabled medical devices and state-level laws governing AI use in health insurance.[6]

The state layer is not theoretical. As of July 1, 2026, TechPolicy.Press, citing the NYU Center on Technology Policy, reported that 29 states had enacted 109 AI laws during the term, along with 28 data center laws; it also identified companion chatbot regulation as the most active category and noted that frontier model safety laws had passed in California, New York, and Illinois.[6] The same report noted that more than six states had enacted AI-in-health-insurance laws in 2026.[6]

Those numbers cut against the lazy version of the deregulatory thesis. Even if the federal government tries to preempt state AI rules, health-related AI remains exposed to sector-specific oversight, state insurance regulation, medical-device classification questions, and patient-impact concerns. A state rule on insurer use of AI may not look like a frontier-model law, but it can matter more to a healthcare AI company’s commercial workflow.

For Tempus AI, counsel should watch whether federal AI preemption proposals reach health insurance and clinical decision-support contexts, how FDA treatment of AI/ML-enabled tools evolves, and whether state insurance regulators continue to impose transparency, review, or human-oversight requirements. The key distinction is between easing general AI governance and easing healthcare oversight. The latter is harder to assume.

Coinbase: Classification Remains the Gate

Coinbase’s AI angle is less direct than Tesla’s or Tempus AI’s. The regulatory channel is crypto classification, enforcement posture, and the possibility that clearer crypto rules support AI-adjacent payments, agents, identity, or data-market infrastructure. ARK’s 2026 outlook identified potential upside from crypto-AI convergence regulation clarity.[7]

That is a narrower proposition than saying AI deregulation benefits Coinbase. The company’s immediate legal exposure still turns on how digital assets are classified, what the SEC and other regulators permit, and how exchanges, custody, staking, listings, and institutional products are treated. A friendlier AI policy environment may be supportive at the margin, but Coinbase’s core regulatory variable is still crypto law.

For legal professionals, Coinbase is useful because it shows where the Wood portfolio stretches the AI label. The relevant monitoring questions are not the same as for an AI model developer. They are whether classification becomes more predictable, whether enforcement risk declines, whether Congress or agencies create usable market-structure rules, and whether AI-related crypto use cases receive enough clarity to move from narrative to product design.

AMD: Export Controls Can Override a Domestic Innovation Story

AMD’s regulatory exposure is narrower but no less material. Motley Fool’s July 2026 analysis described Wood’s preference for AMD over Nvidia as partly a valuation call, with regulatory tailwinds connected to ASIC adoption; it also noted AMD’s exposure to CHIPS Act compliance and export controls on AI semiconductors to China.[5]

This is where the innovation-and-security pairing becomes more than rhetoric. A federal administration may promote domestic AI infrastructure and chip development while maintaining or tightening controls on advanced semiconductors that could support strategic competitors. Export controls do not have to contradict an AI growth agenda. They can be part of it.

For AMD, the monitoring file is therefore different from the files for Tesla or Palantir. Counsel should track changes to China-related controls, licensing requirements, product-performance thresholds, CHIPS Act conditions, and supply-chain commitments. A domestic deregulatory signal may support AI infrastructure investment, while trade controls continue to limit addressable markets or require product redesign.

State Preemption Is the Pressure Point, Not a Settled Fact

The administration’s state-preemption campaign is the part of Wood’s thesis that most directly maps to legal monitoring. The December 2025 executive order’s AI Litigation Task Force was aimed at challenging state AI laws viewed as barriers to innovation.[2] The March 2026 National AI Policy Framework then recommended broad federal preemption, while acknowledging a difficult legislative path.[3]

The state landscape gives that fight practical importance. By mid-2026, states were not waiting for a uniform federal regime; the reported 109 enacted AI laws and 28 data center laws show active state-level policymaking across multiple categories.[6] Some state rules may be vulnerable to federal challenge. Others may survive because they are framed as consumer protection, insurance regulation, employment law, healthcare oversight, procurement policy, or safety regulation rather than general AI governance.

Wood has reportedly invoked New York as a warning case, arguing through a video reposted by Ark Invest Tracker that a proposed bill could block companies from using AI or large language models and drive talent to Texas, Florida, and Nevada. The underlying interview source has not been independently confirmed, so the comment is best treated as an illustration of the geographic regulatory fear she is emphasizing, not as a verified account of a specific bill’s operative effect.

For a compliance team, the preemption question should be broken into narrower checks: whether the rule being challenged is truly an AI law, whether it sits inside a traditional state authority, whether federal agencies have issued conflicting guidance, and whether the relevant company operates nationally or can route operations around stricter jurisdictions.

What Would Strain the Deregulation Bet

Wood has rejected AI bubble fears, and CNBC reported in October 2025 that she continued to frame deregulation as an important support for innovation while flagging broader market-correction risk.[8] Forbes also tied ARKK’s 87% one-year rally to the period following the January 2025 rescission of Biden’s AI executive order and the December 2025 state-preemption order.[1] Those facts explain why the regulatory thesis has market salience. They do not prove that each holding receives the same legal benefit.

A stall or reversal would not hit the five names uniformly. Tesla would be most sensitive to autonomous-vehicle safety and state operating rules. Palantir would be more exposed to procurement policy, agency budgets, and public-sector AI guardrails. Tempus AI would face the healthcare layer even if general AI rules were narrowed. Coinbase would still need crypto classification clarity. AMD would remain tied to export-control and semiconductor-policy choices.

Fragmentation is the harder case. A federal government can sound deregulatory while states continue legislating, agencies continue supervising sector-specific uses, and export-control authorities continue restricting strategic technology flows. That mixed environment is not a contradiction. It is the normal condition of AI regulation in a federal system.

A Practical Monitoring Framework

Wood’s portfolio can still be useful to legal professionals who have no intention of copying the trade. It is a compact map of the regulatory intersections that now define the AI market: transportation safety, public procurement, healthcare oversight, crypto classification, and semiconductor controls.

• For Tesla, monitor NHTSA action, state autonomous-vehicle approvals, robotaxi operating conditions, and any concrete movement toward federal standardization.
• For Palantir, monitor federal AI procurement guidance, defense and intelligence contracting requirements, agency deployment policies, and audit or oversight mechanisms.
• For Tempus AI, monitor FDA treatment of AI/ML-enabled tools, state AI-in-health-insurance laws, clinical decision-support rules, and payer transparency requirements.
• For Coinbase, monitor crypto classification, SEC enforcement posture, market-structure legislation, and any formal treatment of crypto-AI convergence use cases.
• For AMD, monitor AI semiconductor export controls, China licensing rules, CHIPS Act compliance conditions, and domestic manufacturing policy.

That is the narrower and more durable lesson. Cathie Wood’s AI stock picks reflect a deregulatory bet, but the bet is not one legal instrument and not one compliance profile. It is five different encounters with law, each requiring its own watchlist.

References

1. Inside Cathie Wood's AI Stock-Fueled Comeback, Forbes, Oct. 22, 2025.
2. President Trump Signs Executive Order Preempting State AI Laws, Seyfarth Shaw, Dec. 2025.
3. White House Releases Long-Awaited AI Framework, Akin Gump, March 2026.
4. Promoting Advanced AI Innovation and Security, White House, June 2, 2026.
5. 30% of Cathie Wood's Portfolio Is Invested in These 5 AI Stocks, Motley Fool, July 16, 2026.
6. Where State AI Legislation Stands Half Way Into 2026, TechPolicy.Press, July 2026.
7. Cathie Wood's 2026 Outlook: The US Economy Is A Coiled Spring, ARK Invest.
8. Cathie Wood flags market correction risk but rejects AI bubble fears, CNBC, Oct. 28, 2025.