Forget the Enron comparisons—Nvidia’s true challenge may be repeating Cisco’s boom-and-bust fate from the dot-com era. Michael Burry’s sharp warning isn’t about fraud, but about the dangers of overbuilding as AI investment soars. Investors should be watching not for a collapse, but for a drawn-out reckoning that could reprice the entire sector.
Nvidia (NASDAQ:NVDA) has become the epicenter of debate over the future of AI infrastructure – and the sky-high valuations that have come with it. This week, the company issued a detailed seven-page rebuttal directed at vocal skeptics, refuting suggestions of Enron-style accounting manipulations, secret debt, or revenue games with special purpose vehicles.
The back-and-forth is drawing headlines in part because the accusations come from Michael Burry, famed for calling the 2008 housing crash. Burry’s latest target isn’t accusing Nvidia of outright fraud, but of hiding risks beneath bullish accounting practices as the AI investment wave crests. His now widely-circulated Substack and social comments question Nvidia’s revenue recognition, stock-based compensation, and the assumed longevity of its AI chips—an echo of dot-com era over-optimism.
Nvidia’s corporate response is as forthright as it is rare: the company firmly asserts its financial statements are transparent and aboveboard. Yet the dust-up has sharpened the central question for investors—can today’s AI hype avoid the fate of yesteryear’s tech bubbles?
Nvidia Versus Cisco: A Chilling Historical Echo
Burry is explicit: he does not believe Nvidia is Enron. The better historical comparison, he argues, is Cisco Systems (NASDAQ:CSCO) circa 2000. Cisco wasn’t fraudulent—it was simply swept up as the key “arms dealer” of dot-com infrastructure, supplying the world’s internet backbone. Its market cap soared to $560 billion at its dot-com height, outpacing virtually every tech firm before or since.
In less than five years, Cisco stock climbed 3,800%, fueled by projections that internet traffic would double every 100 days. Much like today’s AI buildout, these “heroic assumptions” shaped an extraordinary capital expenditure cycle: telecoms spent billions, and capacity seemed limitless.
But when forecasts missed and demand failed to materialize at expected levels, the crash was brutal. Cisco’s value dropped 78% in the ensuing collapse. Orders dried up, inventories bloated, and it took years for the sector to find its footing even as the importance of the internet only grew.
The AI Capex Cycle: $3 Trillion in the Balance
Fast-forward to today, the AI hardware race is seeing a similar feeding frenzy. Five tech giants—Microsoft (NASDAQ:MSFT), Alphabet (NASDAQ:GOOG), Meta Platforms (NASDAQ:META), Amazon (NASDAQ:AMZN), and Oracle (NYSE:ORCL)—are projected to invest nearly $3 trillion over three years upgrading and building out AI-optimized data centers, powered by Nvidia’s industry-dominating GPUs.
- Nvidia’s market cap touched $4.3 trillion in 2025, up 1,000% in just three years.
- AI hyperscalers form a supply loop, with Nvidia sometimes even investing in customers (like CoreWeave (NASDAQ:CRWV)) that help sustain eye-popping demand for its products.
- Expensing and depreciation for hardware are key watchpoints—Burry estimates GPU depreciation could be understated by $176 billion between 2026 and 2028, inflating overall industry profits.
Burry’s deeper concern is that, just like with Cisco, the real consumer and business end-use cases for AI are still unproven at scale. Massive spending today may not deliver commensurate revenue tomorrow, creating the risk of an abrupt spending contraction once the “arms race” dies down.
Accounting Risks: The New Red Flags?
The specifics matter for investors. Burry highlights that stock-based compensation could be inflating Nvidia’s reported earnings by as much as 50%. Meanwhile, the way Nvidia and its customers handle depreciation may make current profit margins look rosier than they really are. These aren’t accusations of illegal activity, but indications of a cycle that historically ends with overcapacity and hard revaluations—a script all too familiar to observers of the 2001 crash.
The vendor-customer financial loop is another issue: Nvidia occasionally invests in or finances buyers, which can help keep sales moving but risks echoing the supplier credit trends that destabilized other expansions.
For Investors: What Happens Next?
With Nvidia’s shares already down 15% off their highs, the market is showing nerves. The lesson from Cisco is clear: when revenue growth is fueled by runaway capex, the cliff can be steep if the underpinning demand doesn’t justify the investment. Inventories build, pricing comes under pressure, margins compress, and long-term holders must live through years of sluggish recovery.
- Volatility is likely to remain elevated as market participants debate whether this cycle is different.
- Even dominant players can suffer drastic drawdowns—Cisco never regained its pre-dot-com peak for decades.
- Overweight exposure to hot sectors can quickly turn to value traps if growth fails to materialize.
None of this negates Nvidia’s foundational strength in AI hardware, nor does it guarantee a Cisco-style collapse. But the risks of overbuilding and the limits of current end-user demand suggest cautious position sizing, careful attention to earnings quality, and portfolio diversification are the smart play.
Key Lessons from the Dot-Com Echo
For serious investors, the lesson is timeless: Market bubbles rarely burst on fraud—rather, they unwind on economics. Nvidia’s current position as AI’s “arms dealer” echoes Cisco’s role in the dot-com revolution. The resolution, when it comes, will hinge on whether projected AI revenues arrive fast enough. As Michael Burry warns, “sometimes the new company is the same company on a pivot.”
Read more on onlytrustedinfo.com for the fastest and most authoritative coverage of the trends shaping tomorrow’s markets—and the insight to help you stay one step ahead of the headlines.
