Concerns about a swelling bubble in the artificial intelligence sector are intensifying among top technology executives, deepening anxiety within an industry whose valuations have surged at a pace many experts now say is unsustainable.
The warnings didn’t come from the usual corners of Wall Street this time. They came from the stage lights and crowded walkways of Lisbon’s Web Summit, where some of the most visible builders of artificial intelligence quietly admitted what many in the industry whisper behind closed doors. According to a CNBC report, they all agreed that something in this booming sector feels overheated.
Their warnings come as markets begin to confront the scale of capital pouring into the AI boom. Cash is chasing everything from semiconductor giants to small experimental startups, muddying the outlook for long-term revenue and raising serious questions about the reliability of current valuations.
Until recently, most alarms were being sounded in financial circles. Goldman Sachs’ David Solomon and Morgan Stanley’s Ted Pick have both flagged the possibility of a market correction as valuations across major tech firms touched historic highs.
The sharpest intervention came from Michael Burry, the famed “Big Short” investor, who accused major AI infrastructure and cloud operators—often called hyperscalers—of understating chip depreciation costs. He argued that profits at large players such as Oracle and Meta may be significantly overstated. Burry disclosed put options against Nvidia and Palantir, signaling his position that these companies are vulnerable if current assumptions collapse.
Admitting the concerns, this week, top executives who build and deploy AI, during conversations with CNBC at the Web Summit technology conference in Lisbon, acknowledged that the apprehension is no longer confined to Wall Street.
Jarek Kutylowski, chief executive of German AI firm DeepL, said valuations are running ahead of reality.
“I think the evaluations are pretty exaggerated here and there, and I think there is signs of a bubble on the horizon,” he told CNBC on Tuesday.
Picsart CEO Hovhannes Avoyan agreed. He said many young AI companies are being funded at extraordinary levels even though they have no meaningful revenue. Some are being valued on what he described as “noise and vibe revenue,” a phrase derived from “vibe coding,” which refers to using AI to create software without deep technical expertise.
The tension between optimism and caution runs straight through the industry. Executives acknowledge that AI demand is rising, and they remain confident that the technology will shape the future of work and economic productivity. But they also say current pricing carries substantial risk.
Lyft CEO David Risher put it bluntly. “Let’s be clear, we are absolutely in a financial bubble. There is no question, right? Because this is incredible, transformational technology. No one wants to be left behind.”
Risher said the transformative nature of the technology creates a gulf between long-term industrial value and the short-term financial enthusiasm that is driving valuations upward.
“The data centers and all the model creation, all of that is going to have a long, long life,” he said, arguing that real-world utility is not in doubt. “On the other hand, you know, the financial side, it’s a little risky right now.”
Kutylowski said demand for AI inside companies remains strong, with businesses increasingly aware of how automation can improve efficiency. He said companies understand AI can “do magical things,” but full adoption still lags the hype. Many firms, he noted, are “struggling in adopting” the technology. His view is that AI integration will advance, although the pace will not match some of the more exuberant predictions about 2026. DeepL’s main business is an AI translation tool, though it has now developed a broader “agent” to perform tasks for employees.
Cohere’s chief financial officer, Francois Chadwick, also said demand is real and rising, particularly in the enterprise sector, where companies are racing to understand how to incorporate AI into daily operations. His comments align with the sense that underlying interest remains robust even as valuations stretch.
Underpinning the debate is the scale of investment. A report published this week by venture capital firm Accel projected that new AI data center capacity could reach 117 gigawatts by 2030. The report estimated that this buildout could require about four trillion dollars in capital spending over the next five years. Accel calculated that about three point one trillion dollars in revenue will be needed to cover that capital outlay.
Even now, capital commitments remain aggressive. Companies such as Nvidia and OpenAI have announced billion-dollar infrastructure deals worldwide as they seek enough data center space to keep their models updated and commercially viable.
Accel partner Philippe Botteri said future revenue will be powered by three forces: larger and more powerful AI models that require substantial training capacity, the use of new AI services, and what he called the “agentic revolution in the enterprise,” a term often used to describe AI systems capable of carrying out tasks automatically.
The enormous spending has also produced a countercurrent of skepticism. Novo Capital managing partner Ben Harburg said the headline investment figures shouted by large tech companies may be inflated. He said there is growing awareness that expectations around energy use, chips, and data center needs may have been overstated.
“We hear these crazy headline numbers about how much energy is going to be needed, how many chips are going to be needed,” he told CNBC on Tuesday. He warned that enthusiasm around infrastructure is itself becoming a bubble.
Harburg said even Sam Altman “would privately admit” that the industry needs fewer chips, less capital, and less energy than earlier projections suggested.
The tension between excitement and caution is now embedded in the sector. On one side is the conviction that AI will reshape entire industries. On the other hand, there is the mounting concern that financial expectations have sprinted ahead of what the technology can achieve in the near term.
The sector’s leaders, who are often the ones urging acceleration, are now among the voices urging restraint.
