AI Valuations: The Euphoria Is Real, but Michael Burry May Still Have a Point

The AI boom has taken global markets by storm. NVIDIA’s latest quarterly results blew past expectations, chip demand remains red-hot, and every tech giant is racing to build bigger AI models. Valuations have skyrocketed, and the AI trade today looks similar to the early cloud boom - only much faster.

But in the middle of this excitement stands one loud contrarian voice: Michael Burry, the investor who famously predicted the 2008 financial crisis. He recently revealed a short position against NVIDIA - and while the stock continues to hit new highs, some of his concerns are worth understanding.

In other words: the present belongs to AI euphoria, but the future may still have a Michael Burry-shaped question mark.

The Euphoria Around AI Stocks

AI infrastructure has become the hottest theme in global technology. NVIDIA dominates the market with its GPUs and data-center chips, and the ecosystem around it - cloud providers, hyperscalers, AI labs - is aggressively spending to scale capacity.

This has created a perfect storm of high margins, explosive growth, and a massive jump in expectations. The market has rewarded NVIDIA with one of the fastest marches to trillion-dollar territory in history.

But when everything looks perfect, that’s exactly when contrarians begin digging deeper.

Burry’s Concern #1: Depreciation May Be Masking Real Costs

One of Burry’s first criticisms is about how profits are being reported. According to him, a large chunk of NVIDIA’s profitability today comes from the continued use of older chips - chips that should have been retired after 3–4 years.

These older chips are:

• less energy-efficient,
• not optimal for new workloads, but
• fully depreciated - so they carry zero expense on the income statement.

That means profits look higher simply because depreciation costs are artificially low.

As Burry put it: “Just because something is being used does not mean it is efficient.”

In short, extending chip life is inflating margins, not improving real efficiency.

Burry’s Concern #2: Too Many ‘Quid-Pro-Quo’ AI Deals

The second criticism is about the nature of AI demand itself. On paper, it looks like everyone is buying AI capacity. But Burry argues that a large part of this is a closed-circle of reciprocal deals.

A lot of the multibillion-dollar “AI spending announcements” we see are simply:

• NVIDIA selling to cloud giants
• Cloud giants building AI infrastructure
• AI labs (OpenAI etc.) training models on that cloud infrastructure
• Cloud giants reinvesting into NVIDIA

It starts to look less like broad-based demand and more like a “you buy from me, I buy from you” loop.

In Burry’s words, the web of deals “raises questions on genuineness” of true end-market demand.

Burry’s Concern #3: Stock-Based Compensation Is Underreported

This is his biggest criticism - and the most serious.

Tech companies often compensate employees through stock-based compensation (SBC). While SBC isn’t paid out in cash, it is still a real cost to shareholders because it dilutes equity.

Burry notes that since 2018:

• NVIDIA reported SBC of $20.5 billion, but
• The company spent over $113 billion on buybacks - far more than the SBC it disclosed.

His argument: If buybacks are being used to offset SBC dilution, then earnings are being overstated. Properly accounting for SBC could cut current earnings by as much as half.

That’s a huge red flag for valuation accuracy.

So Who Has the Upper Hand Today?

For now, the winner is clear: NVIDIA and the AI bulls. The stock continues to rally, demand remains strong, and analysts still expect 25%–40% upside despite its elevated valuation.

Yes, the stock did fall about 14% after Burry’s concerns went viral, but it quickly bounced back - showing that the market still trusts the AI story more than the doubts around it.

But it’s equally true that:

• hedge funds are tracking real demand signals more closely,
• concerns about SBC are getting louder, and
• reciprocal AI spending between a few large players is being questioned.

Back in 2008, Burry waited more than two years before his prediction proved right. Markets price the present. Burry often prices the future.

What Should Investors Watch?

This article is not advice - but here are neutral indicators worth tracking:

• How much of AI revenue is from true end-market demand
• Transparency on stock-based compensation
• Whether margins remain stable as new chips replace old ones
• The sustainability of AI capex cycles
• Behavior of cloud providers - are they ramping or pausing buys?

AI euphoria is real, but fundamentals eventually matter. Burry’s past shows that being early looks wrong - until it looks brilliant.

Key Takeaways

• NVIDIA continues to dominate the AI hardware space with strong financial performance.
• Michael Burry raises 3 key concerns: depreciation distortion, reciprocal AI deals, and understated SBC costs.
• The stock has rallied despite concerns, showing strong market confidence.
• Some hedge funds are turning cautious as valuations stretch.
• Whether Burry is early or wrong will only become clear over time.

Disclaimer: This article is for educational and informational purposes only and does not constitute investment advice. Mention of any company is not a buy or sell recommendation.