AI Labs Hire Consultants to Prove Nobody Actually Minds Data Centers

The leading AI labs have discovered a remarkable solution to the data center backlash problem: hire a consulting firm to prove the backlash doesn't exist. Milltown Partners, which explicitly counsels leading AI labs and tech startups, recently conducted polling showing that only a small fraction of data center opponents actually live near one. This finding was immediately shared with Axios, because nothing says "independent research" like a vendor's client paying for polling and then rushing the favorable results to press. The implication is clear: public opposition to massive electrical infrastructure is just NIMBYism from people who don't actually have to smell the cooling fans.

What makes this particularly exquisite is the underlying admission buried in the headline itself. Data centers have become "the face of AI backlash." Not a minor concern. Not a local zoning issue. The face. This means that across North America, people have started associating generative AI with sprawling industrial facilities, power consumption that rivals small cities, and the slow heating of their regional water tables. Rather than address why their core infrastructure is becoming a symbol of public unease, the industry's response is to weaponize a consultant's polling data. The message: your concerns don't count because you don't live in the blast radius.

The consultant's employer list is the real tell here. Milltown Partners works for the very companies that need data center buildout to accelerate. They are not a neutral arbiter—they are a hired gun with a financial interest in making data center opposition look irrational or geographically irrelevant. It's like hiring a tobacco industry consultant to poll whether cigarettes actually harm non-smokers, then celebrating when the non-smoking majority expresses concern but happens to live in smoke-free offices. The methodology isn't the point; the narrative is.

Notice the frame being sold: opposition is only legitimate if you personally experience the externality. By this logic, climate change is only a valid concern for people living on melting glaciers, water pollution only matters to people downstream, and AI labor displacement is fine as long as it happens to someone else's industry. This is the consultant's favorite move—redefine legitimacy so narrowly that only those bearing direct costs get a vote. Those bearing systemic costs (grid strain, environmental impact, energy market inflation) are dismissed as NIMBYs with no skin in the game.

What the polling doesn't measure is how many Americans would oppose data centers if they understood what they were for. The survey captures whether people living near data centers support or oppose them, but it says nothing about whether the general public would prefer their electricity bills and grid stability over OpenAI's ability to train GPT-7. Milltown Partners found the perfect metric: one that appears to vindicate the industry while actually proving nothing about whether people want this infrastructure if given full information and a real choice.

This is the state of North American tech consensus in 2026: when faced with public skepticism, hire a consultant, conduct a poll designed to create plausible deniability, and declare victory. The data center backlash isn't going away because a consulting firm found a clever way to redefine what counts as legitimate opposition. It's growing because people understand, at some level, that unlimited growth in computational power comes with real costs—costs that someone else will pay. The fact that those costs are geographically dispersed only makes the poll more necessary, not more convincing.

Nothing announces desperation quite like hiring researchers to prove your critics don't count.