‘Vibe revenue’: AI companies admit they’re worried about a bubble

**Top Tech Executives Raise Concerns Over AI Bubble Amid Soaring Valuations**

*By Eakarat Buanoi | iStock | CNBC*
*Lisbon, Portugal*

Top technology executives have voiced concerns about a potential bubble forming in the artificial intelligence (AI) sector, highlighting growing unease within the industry as valuations continue to soar.

In recent weeks, markets have been grappling with the reality that excessive capital is flowing into the AI boom. This influx has cast doubts over future revenue and profit forecasts, leading many to question the sustainability of current high valuations.

### Warnings from Finance Leaders and Investors

So far, most cautionary signals about inflated valuations have come from investors and finance leaders. Goldman Sachs CEO David Solomon and Morgan Stanley President Ted Pick have both warned of possible market corrections as valuations of several major tech companies hit historic highs.

Adding to these concerns, renowned investor Michael Burry—known for predicting the 2008 financial crisis—accused major AI infrastructure and cloud providers, commonly known as “hyperscalers,” of understating chip depreciation expenses. Burry cautioned that companies like Oracle and Meta may be significantly overstating their profits. He recently revealed put options betting against Nvidia and Palantir.

### Voices from AI Industry CEOs

However, it’s not just financiers raising red flags. CEOs of companies actively developing AI technologies have also expressed skepticism about the market exuberance during interviews at the Web Summit tech conference in Lisbon.

“I think the valuations are pretty exaggerated here and there, and there are signs of a bubble on the horizon,” said Jarek Kutylowski, CEO of German AI firm DeepL.

Echoing similar sentiment, Picsart CEO Hovhannes Avoyan commented, “We see lots of AI companies raising tremendous valuations without any revenue. That’s a concern.” Avoyan referred to smaller startups being backed heavily despite minimal or “vibe revenue,” a term playfully linked to “vibe coding,” which involves using AI to write code without requiring deep technical skills.

### Growing AI Demand Despite Concerns

Despite worries about inflated valuations, industry leaders remain optimistic about AI’s long-term potential.

Lyft CEO David Risher acknowledged, “Let’s be clear, we are absolutely in a financial bubble. There is no question about it. This is incredible, transformational technology, and no one wants to be left behind.” He stressed the difference between the financial bubble and the underlying industrial prospects, adding, “Data centers and model creation will have a long, long life because it transforms how we live and work.”

Looking ahead to AI adoption in 2026, Kutylowski highlighted strong interest and demand: “Everyone understands that AI can do magical things for businesses and elevate efficiency. However, many organizations are still struggling to fully adopt AI. We will make progress, but it’s premature to say every enterprise has it all figured out.”

DeepL’s core offering is an AI translation tool, but the company recently launched a general-purpose AI “agent” designed to perform tasks on behalf of employees. Similarly, Francois Chadwick, CFO of enterprise AI company Cohere, told CNBC, “Demand is definitely there.”

### Massive Capex Outlook for AI Infrastructure

Investment in AI infrastructure is showing no signs of slowing down.

A new report from venture capital firm Accel estimates that new AI data center capacity will reach 117 gigawatts by 2030, representing roughly $4 trillion in capital expenditure over the next five years. To justify this investment, about $3.1 trillion in revenue would be needed.

So far this year, companies including Nvidia and OpenAI have announced multibillion-dollar deals aimed at expanding data center capacity worldwide to meet rising demand.

Philippe Botteri, a partner at Accel, expects that revenue growth will be driven by three major factors: more powerful AI models requiring extensive training capacity, the rising use of new AI services, and the so-called “agentic revolution” in enterprises. The term “agentic” typically describes AI tools capable of autonomously carrying out tasks for users.

### Divergent Views on Future Spending

Not everyone agrees that such massive spending is necessary.

Ben Harburg, managing partner at Novo Capital, suggested that the figures discussed by large tech firms may be inflated. “We hear these crazy headline numbers about how much energy and how many chips will be needed. But there’s probably more of a bubble brewing around infrastructure spending than on actual products,” Harburg said.

He added, “We’re starting to realize there’s been overexuberance around data centers. Even Sam [Altman] would probably admit privately that they need fewer chips, less capital, and less energy than initially anticipated.”

—

As enthusiasm for AI continues to grow, so too does the debate over where the technology’s promise ends and market speculation begins. Industry leaders and investors alike will be watching closely to see how this dynamic unfolds in the coming years.
https://www.cnbc.com/2025/11/14/vibe-revenue-ai-companies-admit-theyre-worried-about-a-bubble.html