X’s warning about Chinese influence has given the AI industry a ready-made villain. The harder problem is that Americans’ anger at data centers did not need to be invented online.

The story began in June, when OpenAI said it had disrupted two Chinese-linked covert networks that used ChatGPT to generate content touching U.S. arguments over AI energy use and technology tariffs. The campaigns failed to gain much attention, but they showed foreign actors were probing an increasingly sensitive domestic fault line.¹

On Thursday, X said its Safety team had identified a suspected Chinese bot farm of roughly 200,000 accounts. About 200, it said, posted material that could manipulate the “legitimate debate about American AI and energy policy,” including claims about strained power grids, rising bills and operators profiting at the public’s expense.David Sacks (DavidSacks)²

That finding has been embraced by advocates who see opposition to the AI buildout as a strategic vulnerability in the U.S.-China technology race. Yet the available evidence also points to a much smaller operation than the rhetoric suggests. Clemson researcher Darren Linvill found that at least 59 of the accounts had zero engagement; many posted only a few times in January before disappearing. “No real humans saw these posts about data centers,” he said, calling their effect negligible.³

The distinction matters. China may be trying to exploit anxieties around AI infrastructure, but researchers say it did not create them. Samuel Woolley of the University of Pittsburgh described the opposition as “absolutely” broad, bipartisan and grassroots, while warning that reports originating with companies that need more data centers deserve scrutiny.³

That local revolt is now visible far beyond X: residents have crowded public meetings, politicians in both parties have entertained moratoriums, and surveys have found substantial opposition to new facilities nearby. One poll cited by Axios found 61% of Americans opposed building data centers in their area. Silicon Valley consultant Jim Prosser put the political problem bluntly: if leaders as different as Greg Abbott and Kathy Hochul agree, “it’s probably not a Chinese psyop.”¹

X’s disclosure may establish an attempted influence operation. It does not settle the larger argument over who pays for AI’s vast appetite for electricity, land and water.⁴