Wall Street veteran analyst David Woo: AI trading is driven by "fear of missing out," making short selling difficult now.

Wall Street veteran analyst David Woo warned that the valuation logic for AI has a fundamental flaw, and once the narrative breaks down, it will have a catastrophic impact on the US stock market.

As Nvidia's earnings report sparked a new round of scrutiny of the AI boom, David Woo stated bluntly in a podcast that the core risk of current AI deals is not the performance of a single quarter, but rather the underlying assumptions that underpin the entire AI valuation system. He believes that from Anthropic's valuation target of $2 trillion to the aggressive capital expenditures of tech giants, the entire AI narrative is increasingly driven by "fear of missing out" rather than economic logic.

Woo explicitly stated that he had been shorting AI-related assets and is currently temporarily withdrawing to observe and wait for the situation to become clearer. However, he emphasized that this is merely a tactical avoidance, not a change in judgment. He warned that if the AI narrative collapses, it will be "catastrophic" for the US stock market—almost every investor holding the S&P 500 will face huge hidden exposure.

The winner-takes-all assumption underpins exorbitant valuations; the logic is questionable.

Woo directly targeted Anthropic's valuation target of approximately $2 trillion, arguing that this figure is based on the assumption of a "winner-takes-all" outcome, which is fundamentally flawed.

"People's mindset is, oh, we're about to have the next Google or Apple," he said.

"But one thing I'm absolutely certain of is that no matter how good Anthropic is, it will never achieve the same dominance in search as Google does."

He further pointed out that the essence of AI is commodification and encroachment, not building moats. "If you can protect something with a natural moat, that's another matter. But I don't see any natural moats here. What I see is encroachment, commodification, and any attempt to protect it will only be a costly waste."

In his view, even Google Search, a business that was once rock-solid, will become a "big loser" under the impact of AI—not because Google's search is bad, but because the capabilities of traditional search are inferior to those of AI.

Capital expenditures driven by fear, accounting treatment raises questions

Woo questioned the nature of the tech giants' latest AI capital expenditures, arguing that the driving force behind them is "fear of missing out" rather than prudent economic calculations.

"The entire story of capital expenditure is driven by fear—the fear of falling behind—rather than economic considerations," he said.

"The only thing I know about fear is that it has no logic, no economic logic, and no upper limit."

He cited Microsoft as an example, expressing strong criticism of its accounting practices. He pointed out that Microsoft's depreciation figure was zero last quarter because the company decided to extend the depreciation period for some of its AI data center projects from 15 years to 25 years. Woo believes this is completely unreasonable—60% of data center costs come from chips, whose actual lifespan is at most two to three years, and depreciation should be accrued quickly. "I feel this is pure accounting trickery," he said.

The profit figures are significantly inflated; the actual growth is far less than advertised.

In addition to accounting treatment, Woo also criticized the overall earnings quality of technology companies this earnings season, arguing that the impressive figures were masked by a large number of one-off factors.

He calculated that once the investment gains from the revaluation of Amazon and Google's Anthropic shares are excluded, the overall profit growth rate of tech giants will drop by about 30%; if the benefits of tariff reductions are further excluded, the growth rate will be cut by another 30%. "The profit figures are basically dressed up for the market to see," he said bluntly.

Short-selling opportunities are blocked, but the risks should not be ignored.

Despite the fundamental flaws in judging AI valuations, Woo admits that shorting this sector is currently extremely difficult.

He pointed out that the Trump administration was adept at turning on the "hype machine" when needed, and the market seemed to believe it wholeheartedly. Meanwhile, companies like Nvidia are deeply embedded in the political and business networks of the AI ecosystem—for example, Nvidia reportedly provided a $125 billion guarantee to OpenAI, and SpaceX reportedly announced an exclusive partnership with Nvidia to build next-generation data centers, backed by a commitment of up to $250 billion. "How do you counter this?" he asked rhetorically.

However, he emphasized that this wait-and-see attitude is tactical, not an abandonment of short-selling. Once the AI narrative falters, the impact will quickly spread to the entire US stock market, given that almost all investors holding S&P 500 stocks have massive implicit exposure to AI-related assets. "That would be a disaster," he said.

Risk Warning and DisclaimerInvesting involves risk; please exercise caution. This article does not constitute personal investment advice and does not take into account the specific investment objectives, financial situation, or needs of individual users. Users should consider whether any opinions, views, or conclusions in this article are suitable for their specific circumstances. Any investment decisions made based on this information are at your own risk.