Is Michael Burry Right to Short AI?

Michael Burry, the investor who famously predicted the 2008 housing crisis, has turned his attention to the AI sector, warning of a speculative bubble fueled by inflated valuations, questionable accounting practices, and a disconnect between hype and fundamentals. His bearish thesis hinges on the idea that AI companies are leveraging aggressive financial engineering to mask the rapid obsolescence of their hardware and the lack of sustainable demand. But is he right? To answer this, we must dissect the valuation realism versus speculative hype debate, using Tesla, Nvidia, Palantir, and AMD as case studies.

The Bear Case: Depreciation, Dilution, and Deception

Burry's core argument centers on accounting practices. He contends that major AI firms are artificially inflating profits by depreciating hardware over extended periods-often 5-6 years-despite the fact that GPUs and servers become obsolete in 2-3 years. This creates a "depreciation trap," where earnings appear robust today but will face massive write-downs in the future. According to a report by , Burry estimates this practice could lead to a $176 billion overstatement of profits across tech giants from 2026 to 2028.

He also highlights valuation disconnects. For example, Tesla trades at a P/E ratio of 260x estimated 2025 earnings, vastly outpacing legacy automakers. Burry argues this premium is unjustified, given Tesla's declining automotive margins, shrinking EV sales, and Elon Musk's stock-based compensation plan, which dilutes shareholders by 3.6% annually. Similarly, Nvidia's P/E of 56.93x, while slightly below peer averages, demands near-perfect execution to justify its valuation according to analysis.

Burry's analogy to past bubbles is stark. He likens the AI boom to the 2008 housing crisis, where circular funding and creative accounting masked underlying fragility. "The illusion of demand is being created through multi-year contracts and enterprise investments," he argues, but these may not withstand scrutiny as the sector matures according to Forbes.

The Bull Case: Infrastructure, Innovation, and Real Demand

Critics of Burry counter that AI is not a financial bubble but an infrastructure revolution akin to the energy sector's expansion. For instance, AMD's data center revenue surged to $4.3 billion in Q3 2025, driven by demand for its Instinct MI350 GPUs. The company's roadmap includes AI accelerators like the MI450 and MI500, with analysts projecting $22.9 billion in data center revenue for 2026.

Nvidia, too, defends its valuation by pointing to its dominance in AI infrastructure. Its Q3 2025 revenue hit $57 billion, with data center sales accounting for $51.2 billion according to financial analysis. CEO Jensen Huang argues that the "virtuous cycle of AI" ensures sustained demand, as older GPUs find secondary uses in inference and analytics according to Morningstar.

Palantir, despite its sky-high P/S ratio of 120x, has shown explosive growth, with U.S. commercial revenue surging 121% year-over-year according to AOL. Its ability to generate 47% free cash flow margins suggests its valuation, while speculative, is not entirely irrational according to PortfolioLab.

Contrasting AMD and Palantir: Realism vs. Hype

AMD and Palantir offer a useful contrast. AMD's valuation appears more grounded in tangible metrics. A discounted cash flow analysis suggests its intrinsic value is $391.22 per share, a 33.8% discount to its current price according to Yahoo Finance. While its P/E of 134.62x is high, its fabless business model and strong R&D spending (25% of revenue) position it to compete in the long term according to KoalaGains.

Palantir, by contrast, trades at a P/E of 580x, reflecting extreme optimism about its AI-driven analytics. However, its lack of transparency on depreciation policies for AI hardware-unlike AMD's industry-aligned practices-leaves room for skepticism. According to CNBC, Burry's short bets against Palantir highlight his belief that its valuation is built on fragile assumptions.

The Verdict: Bubble or Boom?

The AI sector is a mix of both. For companies like Tesla and Palantir, the valuation gap between hype and reality is significant. Tesla's automotive struggles and Musk's dilutive compensation plan make its 260x P/E unsustainable unless its AI ambitions (robotaxi, Optimus) deliver tangible revenue. Palantir's sky-high multiples rely on unproven scalability.

Nvidia and AMD, however, occupy a middle ground. Nvidia's dominance in AI infrastructure and AMD's strategic partnerships with Microsoft and Zoom suggest their growth is more defensible according to AMD's official blog. Yet even here, Burry's depreciation critique is valid: if hardware obsolescence accelerates, earnings could face downward pressure.

Ultimately, the AI sector is a hybrid of innovation and speculation. While Burry's bearish thesis has merit-particularly regarding accounting practices and valuation extremes-the sector's long-term potential cannot be ignored. Investors must weigh the risks of a correction against the rewards of a genuine technological revolution.