Something shifted this week. Not in the technology — the models keep improving, the benchmarks keep climbing, the agents keep getting sharper. What shifted is the mood. And if you’re watching the AI industry, that mood shift might matter more than any new model release.
In the span of a few days: The New York Times asked why the public doesn’t love the AI boom the way they loved dot-com. Bernie Sanders stood at Stanford calling this “the most dangerous moment in the modern history of this country.” Ads quietly appeared inside ChatGPT conversations. And a Bank of America survey showed that a full quarter of fund managers now see the AI bubble as the single largest risk to the market.
The money is flowing faster than ever. The people it’s supposed to serve are increasingly saying no thanks.
$650 Billion, Zero Patience
Big tech is projected to spend a record $650 billion on AI infrastructure in 2026. Amazon’s $200 billion AI announcement tanked its stock. Microsoft’s shares plummeted after stoking fears that ROI on AI investment may be even further off than expected.
A Bank of America survey of 162 fund managers found 35% believe corporations are overinvesting in capex — a record in 20 years of the survey’s history. Thirty percent said AI spending was the most likely source of a credit crisis. Only 20% approved of increasing capital spending at all.
The math is simple and brutal: these companies need AI to generate enormous revenue to justify the burn rate. Not eventually. Soon. And the revenue story isn’t there yet.
ChatGPT Gets Ads Nobody Asked For
OpenAI made it official. As of February 9th, brands like Expedia, Best Buy, Qualcomm, and Enterprise Mobility started appearing inside ChatGPT responses. Major holding companies — Dentsu, Omnicom, WPP — lined up with brands like Adobe, Ford, and Audible. OpenAI is reportedly charging advertisers at least $200,000 just to participate in early tests.
OpenAI’s ads lead Asad Awan framed it diplomatically: “We believe ads play an important role in continuing to support broad access to AI.” Translation: subscriptions aren’t cutting it.
They’re being conservative — Adthena found ads in only about 0.8% of responses across 500+ prompts. No ads in chats about mental health or politics. Advertisers don’t get chat logs or personal data.
But here’s the problem: ChatGPT’s entire pitch was that it was different from Google Search. The alternative to the ad-stuffed, SEO-gamed commercial wasteland. The moment a Best Buy ad appears in a conversation about which laptop to buy, that illusion shatters. The trust equation changes, and it’s changing at the exact moment public trust in AI companies is already eroding.
Bernie Sanders Isn’t the Only One Sounding the Alarm
At Stanford on Friday, Sanders didn’t mince words. After meeting with unnamed “senior leaders” at “the most prominent tech companies” in Silicon Valley, he warned that Congress and the American public have “not a clue” about what’s coming.
“The Congress and the American people are very unprepared for the tsunami that is coming.”
Congressman Ro Khanna — who represents Silicon Valley — described the current moment as a “new gilded age” run by tech billionaires who believe “they would have been heroic conquerors in a different era.” That’s not an outside critic. That’s their own representative saying his constituents have a conqueror complex.
The numbers back it up. A 2025 Pew survey found 64% of Americans believe AI will lead to fewer jobs over 20 years. Only 17% think AI will have a positive impact on the country. Even AI insiders have been, as The Atlantic reported this week, “spooked by advances coming from their tools.”
Why People Loved Dot-Com But Hate This
The NYT raised a question worth sitting with: why did people love the dot-com boom but remain deeply skeptical of AI?
In the late ’90s, the internet felt democratizing. Anyone could build a website. Amazon made shopping easier. Email connected people. The benefits were tangible, immediate, and personal.
The AI boom feels different. The benefits are real but uneven — mostly accruing to companies and developers. The costs feel personal: job displacement, surveillance creep, environmental impact of massive data centers, erosion of creative work. When Google CEO Sundar Pichai compares the AI boom to the industrial revolution “but ten times faster and ten times larger,” that’s not reassuring to people who remember what the industrial revolution did to workers before labor protections caught up.
Tech leaders keep trying to calm the waters. Jensen Huang says AI investments are “just the beginning.” Anthropic’s Dario Amodei argues that “humans plus AI together actually create more meaningful work.” Reasonable positions. But they land differently when your company is burning billions while 64% of the public thinks you’re coming for their jobs.
The Trust Gap Is the Real Threat
Here’s what the industry is getting wrong. The technology is genuinely impressive and probably transformative. But transformation without consent isn’t innovation — it’s disruption in the worst sense.
The internet had time. It rolled out over years, giving people space to adapt, find the parts they liked, develop new skills. The AI boom is trying to compress a decade of social change into 18 months while spending $650 billion on infrastructure the public didn’t ask for.
Google released Gemini 3.1 Pro this week — a “step forward in core reasoning” — and it barely registered. Not because it isn’t impressive, but because the public is drowning in AI fatigue. Another model, another capability, another thing to worry about.
When innovation outpaces your audience’s willingness to engage, you have a marketing problem. When it outpaces their trust, you have an existential one.
What Comes Next
The AI industry faces an inflection point that has nothing to do with model capabilities and everything to do with social license. The technology will keep improving. The spending will continue. But the question that matters most in 2026 isn’t “can AI do this?” — it’s “do people want it to?”
Right now, the answer is: not like this. Not this fast. Not without guardrails. Not without a clearer picture of who benefits and who pays.
The companies that slow down enough to bring people along — that build trust before building features, that share economic gains instead of hoarding them — will survive whatever comes next. The ones that keep insisting the public just doesn’t understand? They’re writing dot-com business plans in a world that’s already learned not to trust the hype.
The AI boom isn’t over. But the era of uncritical enthusiasm ended this week.