Is the Nvidia/OpenAI deal evidence of an AI bubble?
Chipmaker Nvidia announced this week it will be investing up to $100 billion (with a B) in OpenAI to help it build out its computing power. This follows OpenAI’s move earlier this month to buy $300B in computer power from Oracle.
Those are some pretty eye-popping sums. But that may not be a good thing.
AI companies need unprecedented sums of money to increase their computing power. The question is whether AI models will be able to improve enough to make those investments worth it. After GPT-5’s rocky rollout, some people believe the AI bubble is about ready to pop.
And it’s not a fringe idea.
The Wall Street Journal ran it on the front page Friday. And none other than Sam Altman himself recently compared AI to the dot-com bubble, telling reporters that he believes “investors as a whole are overexcited about AI.”
So, the questions are:
Are we in an AI bubble?
If so, who will be the winners and losers after it bursts?
Replies
It’s a fascinating moment for sure. While the sums are eye-popping, I think it’s less about a bubble and more about front-loading the infrastructure for AI’s next phase. The winners will likely be those who can turn compute power into real, repeatable value for users, applications, enterprise integrations, and efficiency gains. Those chasing hype without a sustainable product or use case might struggle if the market cools.
@ritik_ranjan A buildup of infrastructure is a key part of technological innovation — and also a precursor to an inevitable and necessary bubble pop. (That's according to Carlota Perez, if I'm understanding her correctly. https://www.foreignaffairs.com/reviews/capsule-review/2003-03-01/technological-revolutions-and-financial-capital-dynamics-bubbles)
@jeff_benson1 That’s a great point, bubbles often clear the noise and leave behind the real builders. Perez’s perspective fits perfectly here: infrastructure lays the groundwork, and the correction refines who truly adds value.
Feels like deja vu of dot-com. Nvidia will likely win regardless but many AI startups may vanish. The long-term survivors will be those who actually solve real problems profitably, not just burn GPU cycles.
@paul_mbetobong Where does timing come in for those survivors? Who's too soon?
CalPulse
I think that it's challenging to identify clear indicators for an AI bubble right now. The growing enthusiasm and exploration in the field suggest significant development potential.
If we are indeed in a bubble, it will be fascinating to see which companies can innovate and adapt to maintain their competitive edge. The winners are likely to be those that deliver successful products and tangible results quickly :)
Cal ID
There’s definitely some bubble vibes, massive investments and lots of hype. But I'm sure that not every AI company will survive.
Nvidia, Oracle, and other giants will likely benefit no matter what. The real winners will be groups solving actual problems with all that computing, the losers are the startups chasing hype without real value. Even if the bubble pops, the tech and infrastructure will stick around, just like after the dot-com era.
@sanskarix Not saying you're wrong about Nvidia and Oracle, but were there previous booms where seemingly surefire winners got replaced post-bust?
I attended a presentation less than two weeks ago where a few venture capitalists addressed this issue. In short, they're aware that we are in a massive AI bubble but they don't care: the dot-com bubble turned out okay for them and lead to the creation of some of the largest companies by market cap. As for picking winners, just look at the companies they are investing heavily in.
@william_z Yeah, when a literal bubble pops, you still have something left — just minus the air. We'll come out the other side with loads of infrastructure.
Bubble is such a binary word. Is it a bubble if it never pops?
My personal takes is that it feels like there is a gold rush overspending happening, but ultimately no one really knows how big the gold mine is. I think well funded companies (think Meta, Google) can easily justify the stance that it's safer to over spend then under spend.
@build4passion Right, it'll be years before we know whether companies are overspending, underspending, or in the Goldilocks zone.
We are in an AI bubble and it will definitely burst. Most probably mid-late 2026.
Until the AI bubble burst there will be no major advancement in the AI because LLM's are not the right path to AGI. We need to surpass the hype and start a serious approach. There will be no winners, but a complete transformation of the AI environment. Dot-Com vibes in the air.
@mirad_ai What makes you say mid-late 2026?