Is the AI bubble on the brink of a reckoning as dot-com fears return?

Investors are drawing parallels between the AI frenzy and the dot-com boom as major players reduce exposure and questions rise over whether the sector is heading for a long-awaited correction

After months of breathless optimism about the explosive rise of artificial intelligence (AI), the momentum finally appears to be slowing. A new generation of AI-led startups is erupting in both number and valuation, but many still lack a clear path to sustainable revenue or profit. Investors who once backed the sector with near-unquestioned confidence are now reassessing whether the technology’s commercial trajectory truly matches the extraordinary capital pouring into it. Even AI’s real-world use-case expansion, once projected to be nearly boundless, has turned out to be narrower and slower than early forecasts suggested.

 

Against this backdrop, investors and market watchers are asking the same question this autumn: after months of frenzied buying in AI-related stocks and startups, are the signs now pointing to a genuine AI bubble that is beginning to deflate? Recent market moves and growing scepticism around richly valued startups like Perplexity are sharpening those concerns.

 

 

Why is Oracle facing early pressure from AI apprehension? 

Among major tech players, Oracle has taken some of the earliest blows as fears of an AI boom losing steam ripple through markets. The computing-infrastructure giant’s share price has fallen nearly 30 per cent in the past month, reversing much of the gains it made after announcing major AI-infrastructure deals in September. Some analysts now suggest that the scale of Oracle’s AI commitments may be exposing it to heightened financial risk, reflected in the erosion of more than a quarter of its market value since early September.

 

Why are major investors reducing exposure to Nvidia? 

The unease is not limited to Oracle. Several prominent investors have pared or exited positions in Nvidia, the chipmaker powering much of the world’s AI computing backbone. Reports indicate that Peter Thiel’s fund fully exited its Nvidia stake in the third quarter, following similar moves by heavyweight investors such as SoftBank.

 

These exits come despite Nvidia briefly becoming the world’s most valuable company, crossing the five-trillion-dollar market-cap mark.

 

Adding to the scepticism, contrarian investor Michael Burry—whose prescient bet against US subprime mortgages inspired The Big Short—has issued warnings about how AI-focused companies may be overstating profitability. He has criticised the sector’s accounting practices, particularly the extension of asset-depreciation timelines, which he argues artificially boosts earnings.

 

Burry has also pointed to the enormous capital expenditure behind AI infrastructure, warning that overly optimistic assumptions about chip and server lifespans may be masking the sector’s true cost structure.

 

Why is Perplexity drawing scrutiny over its soaring valuation? 

In the startup world, Perplexity has become a symbol of AI-era valuation anxiety. The conversational-search startup raised multiple rounds in 2024–25, propelling it into reported valuations touching the high single-digit to low double-digit billions. More recently, estimates of 18–20 billion dollars have surfaced.

 

But at a recent San Francisco summit, some investors named Perplexity among the startups “most likely to fail.” Their concerns centre on the rapid fundraising, rapidly climbing valuation and an unclear path to a durable revenue model.

 

Are dot-com crash comparisons justified for the current AI boom? 

The growing concerns have rekindled comparisons to the late-1990s dot-com bubble. Analysts point to several parallels:

 

• Massive capital inflows chasing companies with unproven commercial models

 

• Sky-high valuations preceding sustainable revenue or profitability

 

• Market concentration, with a few dominant players—then internet giants, now AI chipmakers and cloud providers—powering most of the market’s upward movement

 

Still, critics argue that today’s AI ecosystem rests on more tangible foundations: established machine-learning models, robust data-centre infrastructure and real revenue streams for cloud providers. These features, they say, make the current cycle more durable than the dot-com boom.

 

What will determine whether the AI bubble bursts or stabilises? 

Whether markets are witnessing a temporary correction or the beginning of a broader collapse will depend on several indicators in the coming months:

 

• Corporate guidance on AI-related revenue

 

• Earnings and new orders at chipmakers like Nvidia

 

• Credit-spread movements and bond-market behaviour for companies funding large AI projects

 

These signals will shape whether the current scepticism marks a reset within a long AI bull market—or the start of a reckoning reminiscent of the dot-com crash.