Erik Gordon, a professor of entrepreneurship and technology markets at the University of Michigan’s Ross School of Business, is warning that the eventual collapse of today’s artificial intelligence investment surge could inflict far greater damage on investors than the dot-com bust of the early 2000s.
Gordon's reasoning is straightforward: the AI boom is bigger, more pervasive, and more capital-intensive — meaning the losses, when they come, could be staggering.
Gordon has been vocal in his assessment that the AI sector is in what he calls an “order-of-magnitude overvaluation bubble.” While some market observers compare the enthusiasm for AI to the dot-com era, others — like investor Kevin O’Leary — dismiss the analogy, citing real productivity gains from AI technologies. Gordon, however, sees troubling parallels, with key differences that point toward deeper financial pain when sentiment shifts.
To illustrate, Gordon draws a comparison between Pets.com — the online pet-supply retailer that became the archetype of dot-com excess — and CoreWeave, an AI infrastructure company backed by Nvidia. Pets.com, once valued at $410 million at its February 2000 peak, collapsed within a year, ultimately filing for bankruptcy and liquidating its assets. Even if every dollar of its peak valuation was lost, Gordon notes, the damage was tiny relative to what’s possible in today’s AI sector.
CoreWeave’s trajectory shows the new scale of risk. The company went public in March, riding intense AI optimism. But in just two days, its share price plunged 33%, erasing $24 billion in market value — roughly 60 times the peak capitalization of Pets.com. The drop followed disappointing earnings marked by widening losses and infrastructure bottlenecks. Despite the selloff, CoreWeave shares still trade at around $100, more than double their $40 IPO price, suggesting valuations remain lofty even after a steep correction.
“It takes a hype-driven tech stock to instantly destroy $20 billion in wealth,” Gordon says. He views this as a sign that when AI’s hype cycle breaks, the impact will reverberate across a much broader investor base than the dot-com crash did.
That broader reach is central to his concern. In 2022, Gordon told Business Insider that far more investors are exposed to AI today than were exposed to dot-com stocks 25 years ago. The sector’s reach into institutional portfolios, mutual funds, ETFs, and retirement accounts means the wealth destruction could be widespread, affecting not only direct equity holders but also millions of Americans through their retirement savings. The magnitude of capital flowing into AI — from venture funds to corporate balance sheets — amplifies the stakes.
CoreWeave’s drop is not an isolated case but part of a pattern Gordon expects to see as reality catches up with expectations. For wealth advisors and RIAs, the lesson is that client portfolios heavily exposed to speculative AI bets — directly or indirectly — carry asymmetric downside risk. These risks can emerge abruptly, leaving little time for tactical repositioning.
The dot-com bust offers a historical benchmark, though Gordon stresses the AI cycle could be worse. Between 2000 and 2002, the S&P 500 (including dividends) fell roughly 9% in 2000, 12% in 2001, and 22% in 2002. Many startups vanished, wiping out investor capital, and thousands of tech-sector jobs were lost. But the size and systemic importance of AI-related companies today — many of which are integrated into core market indexes and widely held funds — mean any collapse would hit far more portfolios, from large pensions to retail retirement accounts.
The difference isn’t just scale — it’s ubiquity. Tech titans now dominate index weightings and, in many cases, rely on AI as a primary growth narrative. This creates a concentration risk for investors who may believe they are broadly diversified but, in reality, are heavily exposed to a single theme. If AI valuations unwind, the knock-on effects could drag down broader benchmarks, impacting everything from passive ETF performance to 401(k) balances.
Gordon also injects a note of behavioral realism. In 2022, he quipped that after the AI bubble bursts, more households would be eating “bowls of spaghetti” — a metaphor for the belt-tightening that follows significant financial losses. It’s a reminder that beyond portfolio numbers, investor sentiment and spending patterns can shift quickly, reinforcing broader economic slowdowns.
While some, like Kevin O’Leary, counter that AI is already delivering measurable productivity gains — making it fundamentally different from the pre-revenue promise of many dot-com companies — Gordon points to the gap between technological potential and sustainable business models. Even companies with real products can be overvalued when expectations run far ahead of execution. CoreWeave’s infrastructure constraints, for example, highlight operational limits that can derail high-growth narratives.
For RIAs, the practical takeaway is not to abandon AI entirely, but to approach it with disciplined risk management. That means:
-
Stress-testing portfolios for concentrated exposure to AI-related equities, especially in index-heavy allocations.
-
Distinguishing between companies with defensible business models and those riding momentum without clear profitability pathways.
-
Preparing clients for volatility scenarios where even “blue-chip” tech names could experience double-digit drawdowns.
-
Considering alternative assets or sectors that can offset tech-heavy risk.
It also means tempering client expectations. Just as many investors in 1999 believed internet stocks would defy gravity, today’s enthusiasm for AI can lead to a suspension of valuation discipline. Advisors must act as a counterweight, helping clients separate long-term potential from near-term hype.
CoreWeave’s sharp drop is a case study in how quickly sentiment can shift when operational realities — such as earnings misses or scaling challenges — collide with sky-high expectations. In an environment where large market-cap companies can shed tens of billions in hours, advisors need to ensure clients are not overexposed to sudden valuation shocks.
As the AI narrative continues to dominate market headlines, Gordon’s warning serves as both a cautionary tale and a call for vigilance. The dot-com bust wiped out many speculative companies, but it also spared much of the broader market’s infrastructure from systemic damage. The AI boom, by contrast, is deeply embedded in today’s market structure. That integration means any sharp reversal could ripple through the entire equity ecosystem — and by extension, through the wealth plans of millions of investors.
For wealth advisors, the challenge is clear: balance participation in AI-driven growth with prudent safeguards against the potential for an equally dramatic collapse. The history of market cycles suggests that sectors with concentrated hype and heavy capital inflows are particularly vulnerable when growth narratives falter.
In the end, Gordon’s message is less about predicting the exact timing of an AI correction and more about urging preparation for its inevitability. Just as with the dot-com era, the market’s love affair with a transformative technology may not end abruptly — but when it does, the losses will be swift, deep, and widespread. Advisors who plan accordingly will be best positioned to help clients weather what could be the most consequential tech-sector downturn in decades.
