Jim Cramer says the biggest danger to the AI trade isn't corporate spending—it's the story. The CNBC host warned that a souring public narrative could hurt AI stocks, arguing the industry is losing the fight over its own reputation even as capital keeps flowing into data centers and chips.
His comments landed the same day the 30-year Treasury yield hit its highest level since 2002, a backdrop that makes expensive growth stocks look shakier. Cramer sees no evidence that AI has actually cost jobs, but he says the perception that it has is already doing damage.
The mood has turned
Cramer pointed to higher electricity prices and job fears as the two forces driving public sentiment against AI. Neither is directly tied to the technology's economics, but both are now part of the political and cultural conversation around it.
The optics aren't helping. Cramer cited a White House AI lunch that included Nvidia's Jensen Huang and Elon Musk—a photo op that, in his view, reinforces the image of AI as a clubby elite project rather than a broad-based benefit.
There's real friction on the ground, too. Research group Data Center Watch counted 45 US projects worth $68 billion that were stalled by local opposition between April and June. That's not a rounding error. It's a sign that community resistance is becoming a material cost for the buildout.
Companies feeding the backlash
Cramer singled out Anthropic and OpenAI for openly flagging their own safety risks. It has been reported that Anthropic's IPO prospectus flags existential AI risks. Cramer's read: those disclosures are honest, but they also hand ammunition to critics.
His suggested fix is blunt. "The companies have to start telling better stories," he said. He did not address capital expenditure directly and still described AI as the strongest growth theme of the era. The problem, in his framing, isn't the spending—it's the narrative.
Spending keeps climbing
The money tells a different story than the polls. Goldman Sachs Research forecasts about $1 trillion of global AI investment in 2026. Hyperscalers such as Amazon, Microsoft, and Google anchor most of those estimates.
But credit markets are getting twitchy. Apollo Global Management chief economist Torsten Slok said credit markets now see more hyperscaler debt risk, tied to rising leverage and uncertain payback on AI spending. That's a shift from the free-money era, when debt for growth was an easy sell.
Investors continue to debate whether data center spending has run too high. Freedom Capital Markets' Paul Meeks, who leads technology research, told CNBC it is too early to judge. That's a fair position—but it also means the debate won't be settled by the next earnings call.
What October and November could bring
Cramer warned that October could turn tough if yields keep rising. Higher long-term rates pressure the discounted value of future earnings, and AI names carry a lot of future earnings in their valuations.
He also flagged the November midterms. A Democratic House could bring congressional probes targeting AI leaders, turning the industry's image problem into a legal and political one.
The next real test comes with third-quarter capex disclosures from the hyperscalers. Those filings may show whether investors are still judging AI by its spending—or by its story. Cramer's bet is that the story now matters just as much.
