AI data center push collides with rising public resistance

Eclipse Ventures is celebrating a massive payday after its early bet on semiconductor startup Cerebras Systems delivered a 2.5 billion dollar return when the company went public this week. The firm poured 147 million dollars into Cerebras over the years starting with a modest 6.5 million dollar Series A in 2016, turning that stake into a windfall at the IPO price of 185 dollars per share. Founder Lior Susan framed the success as validation for a long-held view that the physical world, tied to 85 percent of global GDP, offers richer opportunities than pure software plays.

Susan told attendees at a recent StrictlyVC event in San Francisco that the old software moat has eroded, noting that anyone can now generate code with tools from companies like Anthropic and OpenAI. He pointed to surging shares in chipmakers such as TSMC and Micron as proof that investors and elite founders are shifting toward hardware and software hybrids. Yet this investor optimism collides with broader evidence that ordinary people see little upside from the same technologies.

Recent polling paints a stark picture of public sentiment. A Gallup survey found only 18 percent of Americans aged 14 to 29 feel hopeful about artificial intelligence. An Economist/YouGov poll released this week showed more than 70 percent of respondents believe AI is advancing too quickly, with majorities across both parties expressing concern. Negative views have climbed from 34 percent three years ago to just over 50 percent today, according to YouGov data.

The sources of that unease are concrete. Workers fear displacement in an economy already marked by stagnant wages and rising living costs. Households worry about higher electricity bills as data centers consume ever more power. Environmental advocates highlight the water and energy demands of training ever-larger models. These anxieties surfaced publicly when a Florida real estate executive's commencement speech praising AI as the next Industrial Revolution drew loud boos from graduates.

Industry leaders appear largely insulated from this discontent. Executives at frontier AI labs have expressed surprise in private conversations when confronted with the polling numbers, viewing the technology's spread as inevitable in the same way the internet once seemed. Superhuman Mail CEO Rahul Vohra, whose product relies on AI for email tasks, told Axios he does not see evidence of the backlash in his own operations.

The disconnect carries business consequences. State-level pushback and consumer skepticism are already slowing adoption in some sectors, even as venture returns like Eclipse's continue to climb. Susan's thesis about digitizing the physical world may deliver for limited partners, but it leaves unanswered questions about who bears the costs when that digitization accelerates job losses, energy strain, and wealth concentration among a narrow set of founders and investors. Public markets may reward the hardware winners, yet the broader population is signaling it wants a say in how fast and how far the technology proceeds.