Wall Street's seasoned analyst David Woo has issued a stark warning, suggesting a fundamental flaw in the logic behind AI valuations and predicting a catastrophic impact on U.S. equities if the prevailing narrative fractures.

In a recent podcast appearance, timed as Nvidia's earnings prompt fresh scrutiny of the AI boom, Woo argued the core risk isn't any single quarter's results, but the underlying assumptions propping up the entire AI valuation framework. He contends that the whole AI story, from Anthropic's staggering $2 trillion valuation target to the aggressive capital expenditure plans of tech titans, is increasingly steered by a "fear of missing out" rather than sound economic reasoning.

Woo confirmed he had been shorting AI-related assets but has temporarily stepped aside to await clearer signals. He stressed this is a tactical retreat, not a change in his fundamental outlook, warning that a collapse in the AI narrative would be "catastrophic" for U.S. stocks, given the substantial hidden exposure faced by nearly every investor holding the S&P 500.

The Winner-Takes-All Premise Under Scrutiny

Woo directed sharp criticism at Anthropic's estimated $2 trillion valuation target, asserting the figure rests on the shaky assumption of a "winner-takes-all" outcome. He dismissed the notion that AI will simply produce the next Google or Apple, stating his strong conviction that no matter how capable Anthropic is, it will never achieve the level of dominance Google holds in search. He characterized AI as a domain of commoditization and erosion rather than fortress-building. Without a natural moat, he argued, any attempt at protection would merely be a costly waste. He even suggested that Google's search business, once considered unassailable, is poised to become a major casualty—not because its search is poor, but because traditional search appears increasingly limited when compared to AI's capabilities.

Fear-Driven Spending and Questionable Accounting

Woo also challenged the nature of the tech sector's AI capital expenditure cycle, claiming it is propelled by the fear of being left behind rather than prudent financial calculation. He noted that fear operates without economic logic and has no ceiling.

Citing Microsoft as a prime example, Woo expressed strong doubts about its accounting methods. He pointed out that the company posted zero depreciation last quarter by extending the depreciation timeline for some AI data center projects from 15 to 25 years. This, he argued, defies reason given that chips account for roughly 60% of data center costs and have a practical lifespan of only two to three years, which should warrant rapid depreciation. He labeled the practice as "accounting trickery."

Earnings Quality Called Into Question

Beyond specific accounting treatments, Woo criticized the overall quality of tech earnings this season, asserting that impressive headlines conceal a heavy reliance on non-recurring items. His calculations suggest that stripping out investment gains from Amazon and Google's stakes in Anthropic would lower the sector's aggregate earnings growth by about 30%. Factoring out the effects of tariff relief would slash another 30%. He bluntly described the earnings figures as being dressed up for market consumption.

The Challenge of Betting Against the Hype

Despite his conviction that AI valuations are fundamentally unsound, Woo acknowledged the extreme difficulty of shorting the sector at present. He observed that the Trump administration is adept at activating the "hype machine" when necessary, and markets appear to buy into it. Meanwhile, companies like Nvidia are deeply woven into the political and commercial fabric of the AI ecosystem. Reports indicate Nvidia has provided $125 billion in guarantees to OpenAI, and SpaceX has announced an exclusive partnership with Nvidia for its next-generation data centers, backed by a $250 billion commitment. Woo asked rhetorically how any investor could contend with such forces.

Nevertheless, he reiterated that his current wait-and-see approach is tactical, not a surrender of his bearish thesis. If a crack appears in the AI narrative, the shockwaves would race across the entire U.S. stock market, given the massive hidden exposure to AI-related assets among S&P 500 holders. "It would be a disaster," he remarked.