The conversation around artificial intelligence has shifted beyond technical discussions and into economic and strategic territory. Investors, economists, and business leaders are examining whether the rapid growth of AI represents a sustainable transformation or another speculative boom. Billions of dollars are being committed to the sector, from training infrastructure to model deployments, which raises the question of whether we’re seeing solid foundations or inflated expectations. Michael Burry, the investor who predicted the 2008 financial crisis, has been unusually vocal in saying he believes a bubble is forming. His comments challenge the prevailing optimism that surrounds today’s AI narrative.
Hedge fund manager Michael Burry
The Speed of AI’s Advancement
Artificial intelligence has advanced rapidly in just a few years, and businesses are adopting it even faster. Companies are experimenting with automation, customer interaction, document analysis, planning, and dozens of internal processes. Investment fuels development and development fuels more investment, creating an accelerating cycle. At the same time, the cost of training and operating advanced models—and the infrastructure required to do it—is expanding just as quickly. Massive spending on data centers, chips, and energy consumption has become normal, and that level of capital intensity can be concerning from a historical perspective.
Supporters Say the Valuations Are Justified
Optimists argue that the spending makes sense because AI has the potential to fundamentally reshape how organizations work. They point to measurable productivity gains, especially in automation and decision support, and they highlight industries already reporting tangible improvements. From this point of view, AI isn’t a speculative promise; it’s a technology already producing business value. While many companies are still early in their adoption, usage is no longer theoretical, and supporters see this as evidence that valuations reflect future reality rather than wishful thinking.
Critics Say Revenue Isn’t Catching Up
Skeptics respond that revenue streams remain inconsistent and that many AI businesses are still figuring out how to make money. Some rely heavily on third-party infrastructure, meaning their costs scale with usage in ways that limit profitability. Others operate with significant uncertainty around future margins or pricing models. These concerns don’t necessarily imply the technology is overstated, but they do suggest that valuations may be running ahead of reliable business fundamentals.
Michael Burry’s Warning
Michael Burry’s core argument is that the AI market resembles periods when investors convinced themselves that exponential growth would continue indefinitely. He has compared today’s leading AI companies with Cisco during the dot-com era: a company with real technology but a valuation that dramatically outpaced reasonable expectations. According to Burry, investors are conflating technological progress with guaranteed financial outcomes, and markets are baking in assumptions that may not hold.
Cost Structure and Cash Burn
Burry also emphasizes the cost structure behind advanced AI. Training leading models requires enormous resources, and even inference at scale remains expensive. That level of capital commitment may not be sustainable if revenues take longer to materialize or if competition pushes prices downward. Burry’s position is not that AI itself will fail—it’s that the economics surrounding it may be weaker than the current enthusiasm implies.
The Optimistic Counter-Argument
Supporters of the AI boom note that unlike previous speculative waves, many of today’s AI companies do have real customers and real revenue. Enterprises are purchasing AI tools to solve specific operational problems, and some infrastructure companies are profitable. That distinguishes this moment from earlier cycles in which companies had no established business model at all. While valuations may still be ambitious, the underlying commercial activity is far more concrete than, for example, the dot-com era.
Even If It’s a Bubble, It Might Still Be Good
There is also a historical point worth considering. Many technology bubbles financed infrastructure that later became essential: railroads, telecom networks, broadband, and even elements of the early internet. The fact that investors lost money didn’t mean the technology failed; in fact, it accelerated progress. Some analysts believe we could see something similar with AI, where overinvestment speeds up innovation even if financial corrections eventually follow.
Both Things Can Be True
These perspectives aren’t mutually exclusive. AI may be transformative and overvalued at the same time. A bubble doesn’t require the technology to be fake or useless; it only requires prices to detach from fundamentals. It is entirely possible that valuations correct while AI adoption continues to expand. In other words, the existence of financial excess says little about the long-term importance of the technology itself.
What Could Burst the Bubble?
If valuations correct, it could be triggered by slower-than-expected enterprise adoption, intensified competition, regulatory friction, or macroeconomic shifts that make capital more expensive. Businesses might postpone investment if results are slower or harder to measure. Governments might introduce compliance burdens that raise costs. Competition might compress margins. Any of these could force expectations to adjust.
What If There Is No Bubble?
The opposite scenario is that adoption keeps expanding and infrastructure investment proves justified. In that case, current spending might even look cautious in hindsight. If AI automates large portions of cognitive work over the next decade, the economic consequences could be profound. Many analysts believe we’re still at the beginning of that transformation.
What Businesses Should Actually Do
For most organizations, the smartest approach is not to guess market timing but to focus on practical use cases that create measurable value. AI should be adopted where it reduces manual tasks, improves decision-making, or increases efficiency. Starting with small projects, validating results, and scaling what works is a strategy that holds whether valuations go up or down.
Preparing for Multiple Futures
Businesses that build AI capabilities now are better prepared whether or not the market corrects. If a downturn comes, they continue benefiting from automation and efficiency. If adoption accelerates, they’re already ahead of competitors who waited for certainty. The goal is to develop internal capability, not to predict how markets behave in the next quarter.
A Sensible Middle Ground
Michael Burry’s skepticism is a useful reminder to stay disciplined and avoid blind optimism. At the same time, dismissing AI entirely would ignore genuine technological progress. The balanced perspective recognizes that the market may be overheated and that AI is still an important long-term force.
The Practical Opportunity
At Zarego, we help companies adopt AI in ways that are grounded in real business needs rather than hype or speculation. We focus on identifying use cases that deliver value today and designing systems that scale tomorrow. If you’re looking for a partner who approaches AI with both ambition and realism, let’s talk.
