Michael Burry Launches Substack and Declares the AI Boom a New Dot‑Com‑Style Bubble — With Nvidia as the New Cisco

Michael Burry, the contrarian investor famed for his billion‑dollar bet against the housing market in 2008, is once again sounding the alarm—this time through a paid Substack newsletter. His new platform delivers more than commentary. It offers a warning: today’s booming AI sector may be headed for a historic collapse. At the center of his thesis is Nvidia, the $2+ trillion chipmaker that Burry claims is playing the role Cisco did in the late 1990s—overvalued, overhyped, and vulnerable to a brutal comedown.

From Big Short to Big Blog

Burry’s pivot to newsletter writing signals a broader shift in his approach. Rather than managing external funds or surfacing through SEC filings, he now engages directly with retail and institutional audiences. His newsletter, infused with mythological references, frames him as a modern‑day Cassandra—offering insights others may choose to ignore until it’s too late.

He has been increasingly critical of market structures that reward momentum and penalize skepticism. With this new platform, he’s free to articulate what he sees as uncomfortable truths without needing consensus.

Nvidia: The Modern Tech Darling in Burry’s Crosshairs

At the core of Burry’s concern is the massive scale of AI infrastructure spending. He points to parallels between the current chip‑driven AI buildout and the bandwidth arms race of the late 1990s. Then, it was routers and fiber. Now, it’s GPUs and data centers.

Nvidia is, in Burry’s words, “this era’s Cisco.” While its fundamentals may look strong, he suggests that expectations baked into its valuation are unsustainable. Companies are buying high‑end chips with depreciation schedules of five to six years, but the real commercial value of the chips—driven by fast AI model evolution—may only last two or three. That mismatch, according to Burry, means reported earnings are overstated and will correct once depreciation assumptions hit reality.

He estimates that if depreciation schedules across hyperscalers and enterprise customers remain out of sync with product cycles, hundreds of billions in anticipated earnings could vanish over the next few years.

Warning or Opportunity? What Investors Should Take Away

Burry isn’t just writing. He’s betting. Public disclosures suggest he has taken bearish options positions against Nvidia and other AI‑heavy stocks. While he remains vague about the precise size and structure, the implication is clear: he sees this as one of the best asymmetric risk setups since the mortgage crisis.

His warning comes at a time when many investors have been chasing AI upside without reassessing downside risk. He challenges the bullish consensus and asks: what happens when demand slows, or when every company that needs an AI stack has already bought it?

From his view, current valuations appear to reflect perfection. In that kind of environment, any deviation—a pause in growth, a shift in cap‑ex, or a surprise in accounting treatment—could trigger a cascade of re‑ratings.

What’s Next for Burry—and for AI Markets?

The upcoming earnings seasons for Nvidia, AMD, and hyperscale cloud providers will be critical. If growth persists, Burry may appear early or wrong in the short term. But if the pace slows, margins compress, or chip refresh cycles shorten dramatically, his case will gain momentum.

The more revealing signals may not be in top‑line numbers but in strategic pivots—if companies scale back cap‑ex, adjust guidance, or hint at oversupply, it could spark a reevaluation across the AI infrastructure stack.

For now, Burry is using his voice and his trades to shape the conversation. And whether he proves prescient or premature, one thing is certain: in an era where AI is treated as inevitable and exponential, his thesis is a reminder that every boom carries the seed of its own undoing.