CNN

America may lose the AI race to China. Not because it lacks the technology, talent, or capital, but because Americans are increasingly voting against the infrastructure required to compete. The 2026 midterms exposed a remarkable trend: voters favored restricting data-center development in 40 of 41 local ballot contests. Opposition crossed party lines, driven by concerns about electricity prices, water consumption, environmental damage, and corporate subsidies. But how much of this opposition survives scrutiny? Data centers consume substantial electricity, but households don’t have to subsidize them. Dedicated generation, long-term energy contracts, and appropriate utility pricing can protect consumers. New generation and transmission investments can also strengthen the grid. Water consumption depends heavily on cooling technology. Closed-loop systems, air cooling, and appropriate site selection can dramatically reduce freshwater demand. Environmental impact is an engineering challenge, not an inherent argument against data centers. And the complaint that data centers create relatively few permanent jobs misses their economic purpose. We don’t measure the value of power plants, telecommunications networks, or ports by the number of people employed inside them, but by the industries they enable. Corporate subsidies are a legitimate policy debate. Governments can reject excessive incentives and require developers to finance infrastructure upgrades without obstructing an entire strategic industry. Meanwhile, China understands exactly what’s at stake. In June, OpenAI exposed a suspected Chinese influence operation called “Data Center Bandwagon”, using AI-generated content to amplify American fears about electricity costs and data-center development. In August, X identified approximately 200 accounts within a larger suspected Chinese bot network attempting to manipulate the same debate. And this week, reports emerged that U.S. intelligence suspects a Chinese intelligence operative cultivated a close relationship with Nvidia CEO Jensen Huang, allegedly seeking intelligence about American AI chips and influence over decisions about exporting advanced processors to China. The allegations remain disputed and unproven. But the competition extends far beyond hardware. Research into leaked documents from Chinese technology company GoLaxy revealed an alleged AI-powered propaganda system capable of profiling political figures, manufacturing convincing online personas, and tailoring narratives to influence public opinion. This is the other side of the AI race: not merely developing systems that answer questions, but influencing the information those systems consume and the historical narratives they reproduce. Whoever dominates the AI ecosystem could acquire extraordinary influence over how billions of people encounter information, interpret history, and understand political reality. AI isn’t merely about chatbots. It’s about military intelligence, cyberwarfare, autonomous weapons, scientific discovery, industrial productivity, and increasingly, the infrastructure through which societies acquire knowledge. And AI requires compute. Compute requires chips, electricity, and data centers. China is investing in those capabilities while Americans increasingly treat their own infrastructure as a political liability. The irony is extraordinary. Americans fear losing their economic and technological advantage to AI while opposing the infrastructure that would allow American companies to compete. Beijing doesn’t need Americans to support China. It benefits when Americans obstruct their own technological development. The AI race isn’t just about who builds the most powerful model. It’s about who possesses the infrastructure to run it, the capacity to shape its outputs, and ultimately the influence to define what the world believes.