Has talk of an AI bubble become… its own bubble?
The conversation around artificial intelligence has dominated the financial and business landscape all year. But this week, a few familiar names from past market manias joined the chorus — reigniting debate over whether the boom in AI investment is a sustainable growth story or another speculative cycle in disguise.
A warning from “The Big Short.”
Michael Burry — the hedge fund manager best known for calling the 2008 housing crash — is once again sounding alarms. His latest target: the massive capital commitments Big Tech is making toward AI infrastructure. Burry points to a potential accounting blind spot: the short lifespan of semiconductor chips. He believes companies like Meta, Oracle, and Microsoft may be underestimating depreciation costs by as much as $176 billion, distorting the true profitability of their AI investments.
Burry’s comments resonate with investors who remember how fast technology lifecycles can compress — particularly when competition accelerates innovation. If the pace of AI advancement renders today’s chips obsolete faster than expected, hyperscalers could find themselves writing down assets long before those data centers produce meaningful returns.
Echoes of the dot-com era.
JPMorgan analysts are drawing their own cautionary parallels. They liken today’s AI infrastructure buildout to the late-1990s fiber-optic boom, when telecoms poured billions into network capacity that took years to monetize. The fear now is that hyperscalers are racing to construct data centers and purchase GPUs before fully understanding the revenue potential of AI applications.
In other words, the market may be building the highways before there are enough cars to drive on them. For advisors and asset allocators, that raises a familiar question: are current valuations based on long-term utility, or on short-term enthusiasm?
Enter SoftBank — again.
Adding more intrigue, SoftBank made headlines by disclosing it sold its entire $5.8 billion stake in Nvidia. The move initially surprised the market, given Nvidia’s central role in the AI rally. But the Japanese conglomerate, known for its aggressive bets during the WeWork era, isn’t walking away from AI. Instead, it appears to be freeing up capital to reinvest in its growing exposure to OpenAI and other generative AI ventures.
For SoftBank, the shift underscores a belief that the next wave of value creation will come from the application layer — not just the hardware powering it. If that’s true, we may be witnessing the first signs of rotation within the AI ecosystem: away from chipmakers and toward the platforms and software companies using those chips to deliver real-world productivity gains.
Is this the start of the unwind — or just noise?
Despite the skepticism, few on Wall Street expect an AI pullback anytime soon. The scale of investment already committed is enormous, and the technology’s potential remains too compelling for most investors to ignore. Industry estimates suggest global AI spending could reach into the trillions over the next decade, spanning infrastructure, cloud, and enterprise adoption. That level of momentum doesn’t reverse quickly.
Even with rising questions about near-term profitability, Big Tech leaders are pressing forward. Microsoft’s CEO Satya Nadella and Alphabet’s Sundar Pichai continue to frame AI as the defining growth engine of the next decade — not an optional bet. Meta’s Mark Zuckerberg has also defended the company’s escalating AI expenditures, suggesting that AI integration across social platforms and advertising will ultimately boost both engagement and monetization.
For advisors evaluating client exposure to AI, this dynamic presents a double-edged sword. On one hand, the sector remains the primary driver of market performance — with Nvidia, Microsoft, and Amazon accounting for a disproportionate share of index gains. On the other, the same concentration risk that fueled the rally could magnify volatility if sentiment shifts.
Why this isn’t a simple bubble call.
Unlike the dot-com era, today’s AI investments are being made by companies with durable cash flows, deep R&D capacity, and global scale. Their balance sheets can absorb cycles of experimentation and failure in a way 1990s startups never could. That financial resilience may limit downside — even if valuations eventually compress.
Still, Burry’s caution isn’t misplaced. Investors have a tendency to underestimate how long transformational technologies take to translate into sustainable profits. The internet revolution created trillions in value, but it took more than a decade for the winners to emerge. For advisors guiding clients through this AI cycle, the key is distinguishing between strategic exposure and speculative excess.
The patience problem.
The next phase of AI’s evolution will test investor patience. Large language models, while impressive, are costly to train and operate. Profitability hinges on scaling user adoption, reducing compute costs, and developing differentiated applications. OpenAI, for instance, still faces a long road toward consistent profitability, even as ChatGPT reshapes digital productivity.
But there’s precedent for staying the course. Cloud computing went through its own “bubble” phase in the early 2010s, when companies like Amazon and Microsoft poured billions into infrastructure long before cloud margins improved. Today, those investments underpin some of the most profitable business segments in tech. If AI follows a similar trajectory, this current phase may look more like early buildout than overextension.
Wall Street’s balancing act.
Institutional investors aren’t blind to the risks, but they’re also aware of what’s at stake. The broader equity rally — from the S&P 500 to the Nasdaq — has been powered by AI optimism. Pulling back exposure prematurely could mean missing further upside if productivity gains materialize faster than expected.
At the same time, the sheer scale of AI-related capex — now topping $200 billion annually across the top hyperscalers — leaves little margin for execution errors. Advisors managing equity allocations may want to think in terms of diversification within the AI value chain: balancing chipmakers and cloud providers with software enablers, enterprise adopters, and cybersecurity firms positioned to benefit from the same trend.
A tale of two narratives.
Ultimately, the AI debate has become a reflection of investor temperament as much as market fundamentals. On one side are the believers — those convinced that AI represents the most transformative technological leap since the internet. On the other are the skeptics — veterans who’ve seen cycles of exuberance before and are wary of assuming exponential growth curves will persist indefinitely.
Michael Burry’s recent clash with Palantir CEO Alex Karp on social media captured this divide perfectly. Burry warned that AI hype risks outpacing its tangible impact, while Karp countered that AI is already delivering operational and defense breakthroughs that make skepticism shortsighted. Both are right in their own ways: the potential is real, but so are the growing pains.
For advisors, clarity matters more than conviction.
Rather than choosing sides, wealth managers can help clients focus on the framework — not the frenzy. That means differentiating between AI as a broad secular growth theme and AI as a concentrated trade. It also means watching fundamentals closely: earnings visibility, capital discipline, and evidence of monetization.
The temptation to time the AI “bubble” may be strong, but a more practical approach involves scaling exposure relative to valuation risk and cash flow momentum. As AI spending ripples across sectors — from semiconductors and data infrastructure to healthcare, finance, and industrial automation — opportunities will broaden. Advisors who maintain flexibility can participate in the upside without overcommitting to one segment of the hype cycle.
The bottom line.
There’s no clear resolution to the AI debate yet — and that’s exactly why it remains the market’s most important story. The technology’s potential to reshape productivity, profitability, and even economic growth is too significant to dismiss. But so are the warning signs flashing from history: overinvestment, overconfidence, and underappreciated execution risk.
Whether AI proves to be a revolution or a rerun of past manias, the truth will unfold over time — not in the next quarter. For now, advisors should expect volatility, maintain diversification, and treat every headline — bullish or bearish — as another data point in a long-term transformation that’s still just beginning.
