When the Insiders Start Talking
Something shifted in late 2025. On a single Friday, three of the most prominent voices in artificial intelligence stepped forward and hedged, publicly, on where this whole thing is going. Goldman Sachs CEO David Solomon said he expects “a lot of capital that was deployed that doesn’t deliver returns.” Jeff Bezos, Amazon’s founder and executive chairman, called the current environment “kind of an industrial bubble.” And OpenAI’s own CEO Sam Altman warned that “people will overinvest and lose money” during this phase of the AI boom.
These aren’t outside critics or short-sellers. They are the architects of the boom itself. That makes their words worth sitting with, especially as investors tracking everything from msft stock to nvidia price are trying to figure out whether the rally has genuine legs or whether it’s running on borrowed confidence.
The Numbers Behind the Noise
The scale of spending is genuinely difficult to process. Amazon, Alphabet, Meta, and Microsoft alone spent nearly $300 billion on capital expenditures in 2025, a figure that GMO estimated at roughly 1.3% of U.S. GDP. AI-related capital expenditures surpassed the U.S. consumer as the primary driver of economic growth in the first half of 2025, accounting for 1.1% of GDP growth, according to Yale Insights. That is an extraordinary structural shift, and it happened fast.
Since ChatGPT launched in November 2022, AI-related stocks have accounted for 75% of S&P 500 returns, 80% of earnings growth, and 90% of capital spending growth, according to Michael Cembalest of JPMorgan Asset Management. Bloomberg Intelligence put the total market valuation added by AI-related stocks at around $17.5 trillion over that same period. For anyone watching microsoft share price climb through this cycle, those figures explain a lot of the momentum.
Nvidia made a $100 billion commitment to OpenAI in 2025. Cloud providers including Amazon Web Services, Microsoft, and Google are expected to collectively ramp their AI spending further in 2026. The spending trend, by any measure, appears strong. Whether the returns match it is a separate and much harder question.
OpenAI’s Uncomfortable Math
OpenAI sits at the center of this debate in a way that deserves direct attention. The company, backed by Microsoft, reported $3.7 billion in revenue against operating costs of $8 billion to $9 billion. It anticipates generating $13 billion this year, which would represent substantial growth. The projection from The Information, though, is harder to dismiss: OpenAI is expected to incur losses amounting to $129 billion by 2029.
That gap between revenue trajectory and cumulative losses is precisely what the bubble debate hinges on. Fidelity has noted that the monetization of AI still lags investment, even as the spending trend remains strong. That’s not a contradiction so much as a description of where the industry sits: enormous capital deployed, returns still catching up.
Is This a Bubble or Just a Boom?
The word “bubble” gets used loosely, and it’s worth being precise about what the disagreement actually is. A Yale Insights survey found that 60% of CEOs polled didn’t believe AI hype had led to overinvestment. The other 40% raised significant concerns. That split is not a ringing endorsement of the status quo.
Yahoo Finance’s framing that AI stocks like Nvidia and Microsoft are “nowhere near a bubble” represents one camp, and it’s not a fringe view. Nvidia’s second-quarter data showed the massive influx of capital into AI had significantly elevated stock values, and the earnings narrative around the company has been consistently strong. Nvidia’s November 19 earnings were flagged as a key test of whether the AI boom still had legs. The outcome of that kind of test shapes how investors read everything downstream, from perplexity ai valuations to coinbase’s exposure to AI-adjacent token speculation.
The circular quality of some deals deserves scrutiny. Nvidia committing $100 billion to OpenAI, while OpenAI’s primary backer Microsoft is itself one of the largest buyers of Nvidia chips, creates a financial loop that looks elegant until you ask where the external revenue justifying all of it actually comes from. The chainlink news cycle and broader crypto markets have shown repeatedly how quickly sentiment-driven valuations can detach from fundamentals, and some analysts see a parallel forming here.
What Comes Next
Crypto betting by TipsGG and other speculative platforms have long understood something traditional finance is still processing: markets can stay irrational longer than any single participant can stay solvent, but the reckoning does eventually arrive. Bitcoin’s own history of boom and correction cycles offers a useful, if imperfect, reference point for thinking about concentrated enthusiasm in a single technology narrative.
The AI spending cycle is real. The infrastructure being built is real. The question nobody can answer cleanly is whether the returns will arrive fast enough, and at sufficient scale, to justify the $300 billion in annual capex now being treated as routine. Fidelity’s assessment that spending trends appear strong is accurate. So is Solomon’s warning. Both things are true at once, and that tension is where the actual risk lives.