There are elements of irrationality in a moment like this
Sundar Pichai, speaking from the heights of a company whose valuation has doubled in seven months, offered a rare act of self-implication: the artificial intelligence boom, he suggested, carries within it the seeds of its own correction, much as the internet did before it transformed the world and then briefly broke it. His warning is not a prophecy of collapse but a reminder that transformative technologies and rational markets do not always travel together, and that the distance between the two is where fortunes—and reputations—are made and lost.
- OpenAI plans to spend 1.3 trillion euros over eight years against just 12.2 billion in expected annual revenue—a gap that echoes the reckless arithmetic of the dot-com era.
- Nvidia's valuation has reached 4.7 trillion euros, a figure that would have seemed fictional months ago, signaling a sector moving faster than any underlying reality can justify.
- Pichai draws an explicit parallel to the 2000 internet crash, acknowledging 'elements of irrationality' even as his own company's stock rides the same wave of enthusiasm he is cautioning against.
- Google argues its vertically integrated stack—custom chips, proprietary data, in-house models—gives it structural resilience that more narrowly focused competitors may lack when the correction arrives.
- The energy demands of AI infrastructure are already pushing Alphabet's climate commitments off schedule, exposing a deepening tension between technological ambition and planetary constraint.
Sundar Pichai sat down with the BBC and said something Silicon Valley rarely hears from its own: the AI boom is showing signs of irrationality, and when the correction comes, Google will not be exempt. The timing is striking. Alphabet's market capitalization has doubled in seven months to 3.3 trillion euros, yet Pichai chose that moment to invoke the ghost of the dot-com collapse—when internet companies soared on promise alone before crashing in 2000. 'There was clearly excessive investment,' he said of that era, while insisting the internet's transformative power was never in doubt. He hopes AI will follow the same arc, but he refuses to pretend the current moment is entirely sane.
The numbers give his concern weight. OpenAI has committed to 1.3 trillion euros in infrastructure spending over eight years against expected revenue of just 12.2 billion this year. Nvidia, the chip manufacturer underpinning the entire buildout, has reached a valuation of 4.7 trillion euros. The sector is operating on assumptions of unlimited progress and unlimited capital—neither of which history tends to honor indefinitely.
Pichai did make a case for Google's resilience. The company controls what he called a 'complete stack'—its own chips, its own models, decades of research, and the data wealth of YouTube—giving it a structural depth that more specialized competitors lack. If the market turns, that integration could absorb the shock more effectively than single-bet rivals.
He also raised two quieter concerns. He warned users not to treat AI as a reliable source of truth, urging a more discerning relationship with tools that are genuinely useful but not infallible. And he acknowledged that AI's voracious energy demands are already delaying Alphabet's climate goals, warning that constraining energy in a digital economy 'will have consequences.' What emerges is a portrait of a leader unwilling to pretend the moment is without risk—not predicting a crash, but refusing to look away from the math.
Sundar Pichai sat down with the BBC recently and said something that Silicon Valley doesn't often hear from its own leadership: the artificial intelligence boom is starting to look irrational, and when the correction comes—if it comes—Google will not be spared.
The timing of his warning is sharp. Alphabet's stock has doubled in value over seven months, pushing the company's market capitalization to 3.3 trillion euros, buoyed almost entirely by investor enthusiasm for its AI models and the custom chips designed to run them. Yet Pichai, who runs both Alphabet and Google, used his platform to pump the brakes. He acknowledged that the sector is experiencing an 'extraordinary moment,' but he also drew a direct line to the dot-com bubble of the late 1990s—when internet companies were valued at astronomical multiples, only to collapse in 2000, leaving behind a trail of bankruptcies and mass layoffs. 'There was clearly excessive investment,' Pichai said of that era, 'but none of us would question whether the internet was transformational. I hope it's the same with AI. So I think it's rational and, at the same time, there are elements of irrationality in a moment like this.'
The numbers backing his concern are staggering. OpenAI alone has committed to spending 1.3 trillion euros on infrastructure over the next eight years—against expected revenue of just 12.2 billion euros this year. That gap between spending and earnings is the kind of math that typically precedes a reckoning. Nvidia, the chip manufacturer that has become the backbone of the AI infrastructure buildout, recently hit a valuation of 4.7 trillion euros, a figure that would have seemed impossible just months ago. The entire sector is moving at a pace that assumes unlimited technical progress and unlimited capital availability, neither of which is guaranteed.
Yet Pichai also made a case for why Google might weather the storm better than others. The company, he argued, controls what he called a 'complete stack'—from the processors it designs internally to the vast trove of data from YouTube, from its own AI models to decades of advanced research. This vertical integration, in his view, gives Google structural advantages that more specialized competitors lack. If the market corrects sharply, Google's diversified assets and in-house capabilities could absorb the shock more effectively than companies betting everything on a single technology or service.
Beyond the market mechanics, Pichai raised two other concerns that reveal the deeper tensions in the AI moment. First, he cautioned users against blind trust in AI systems. These tools are genuinely useful for creative and exploratory work, he said, but they are not reliable sources of truth. People need to learn what AI is good for and what it isn't—a message that cuts against the hype that has driven so much of the investment in the first place. Second, he acknowledged that the energy demands of expanding AI infrastructure are already delaying Alphabet's climate goals. The company still aims for carbon neutrality by 2030, but Pichai warned that restricting energy availability in an increasingly digital economy 'will have consequences.' It's a bind: the technology that promises so much requires resources that the planet may not easily spare.
What emerges from Pichai's remarks is a portrait of a leader trying to be honest about the moment he's in. He's not calling for a halt to AI development. He's not predicting a crash. But he is saying, plainly, that the current pace of investment and expectation cannot continue indefinitely, and that when reality reasserts itself, no company—not even Google—will be immune to the fallout.
Notable Quotes
There was clearly excessive investment in the internet era, but none of us would question whether the internet was transformational. I hope it's the same with AI.— Sundar Pichai, CEO of Alphabet and Google
People have to learn to use these tools for what they're good for and not blindly trust everything they say.— Sundar Pichai
The Hearth Conversation Another angle on the story
When Pichai says 'irrational,' what does he actually mean? Is he saying the technology isn't real, or that the money being thrown at it is?
He's separating the two. The technology is real and transformative—he's clear on that. What's irrational is the assumption that it will scale infinitely without hitting technical or economic limits, and the willingness to spend trillions before there's any proof of return.
But Google itself has benefited enormously from this bubble. Why would he warn against it?
Because he's seen what happens when bubbles pop. The dot-com crash didn't erase the value of the internet, but it destroyed companies and wiped out investors. If Pichai can position Google as the company that saw the danger coming, that's actually a form of protection.
He mentions Google's 'complete stack' as a defense. What does that really mean?
It means Google owns the whole chain—the chips, the data, the models, the research. If the market contracts, a company that only makes chips or only runs a service gets hit hard. Google can absorb losses in one area by drawing on strength in another.
What about the energy problem he mentions? That seems like the real constraint.
Exactly. You can't argue your way around physics. AI systems consume enormous amounts of electricity. If you can't get the power, you can't scale the technology, no matter how much money you have. That's the hard limit nobody's talking about loudly enough.
Does he think a crash is coming?
He's not predicting one. He's saying the conditions for one exist, and if it happens, it won't be gentle. He's essentially saying: this is unsustainable, but I don't know when it breaks.