X has dismantled a network of more than 200,000 accounts traced to China, with suspect profiles posting claims about data-center pricing and power-grid strain to shape the public conversation around AI infrastructure, the company said Friday, August 28, 2026.
The bot farm pushed coordinated content about electricity costs, AI training loads, and grid stress — claims designed to “manipulate” debate around the buildout of large-scale compute facilities, according to X’s safety team.
It is the largest single takedown tied to Chinese state-linked influence operations since the company began publishing suspension data in 2024.
Chinese ai: How the Network Operated
X’s investigation showed the accounts functioned as an amplification layer. A smaller set of core accounts seeded posts about AI electricity demand, energy tariffs, and grid reliability. The wider 200,000-account pool then quote-tweeted, liked, and replied to push those claims into trending topics tied to energy policy and data-center permitting.
Suspect accounts posted claims about energy pricing and power-grid strain to “manipulate” public debate around AI infrastructure, X said. The content surfaced most heavily during state-level utility commission hearings and federal rulemaking windows — moments when the marginal reader can swing outcomes. Here’s the thing: the timing is the tell. Energy permitting is one of the few areas where a synthetic 200,000-account swing can move real policy, because comment windows are short and the audience is small enough to be flooded.
Scale of the Takedown
The 200,000-account figure dwarfs prior Chinese-linked removals. X removed roughly 7,200 accounts in its April 2025 disclosure and about 28,400 in a December 2024 operation tied to spam-for-hire services.
The current bust is more than seven times larger than the December action. Worth noting: the network’s footprint crossed platforms. Researchers at the Stanford Internet Observatory flagged overlapping content on TikTok, YouTube, and Telegram comment sections, suggesting the same operator ran a multi-platform playbook rather than an X-only effort.
The Claims Being Pushed
Posts flagged by X hit three recurring themes:
- AI data centers are responsible for double-digit electricity
- Hyperscaler construction is destabilizing regional grid reliability.
- Domestic power-generation capacity cannot keep pace with AI compute expansion, justifying policy slowdown.
That last claim is the one researchers flagged as highest-leverage. If “we cannot build fast enough” becomes the consensus frame, it gives cover to permitting delays, grid interconnection freezes, and utility rate-cap fights — all of which slow U.S. AI capacity buildout relative to Chinese peers.
Why This Matters Now
The bust lands as U.S. utilities are approving record numbers of data-center interconnections, with roughly 50 GW of new load queued in PJM and ERCOT alone. The Federal Energy Regulatory Commission has also opened proceedings on co-location tariffs and behind-the-meter generation.
Any narrative pushback during these windows has outsize policy weight. X’s safety team said it shared indicators of compromise with peer platforms and with the FBI’s Foreign Influence Task Force. A spokesperson declined to comment on whether any U.S.-based amplifiers were identified.
What Comes Next
Expect two concrete follow-ups. First, X is likely to publish a full dataset of the suspended accounts, following its April 2025 transparency pattern.
Second, the FBI and the Office of the Director of National Intelligence have increasingly flagged Chinese AI-influence operations in their annual threat assessments — this bust gives them a fresh, citable case to point to.
For now, the platform’s enforcement team is treating the network as a single coordinated actor rather than a loose cluster, which raises the legal stakes for any human operators eventually identified.
The Chinese AI influence campaign shows that compute is now a contested information domain — and the platforms policing it are treating the energy debate as a frontline.
Related Articles
FAQs
What did X remove?
When was the takedown announced?
X disclosed the action on Friday, August 28, 2026.
What was the goal of the bot farm?
The accounts sought to manipulate public debate around AI infrastructure and energy policy, particularly during utility commission comment windows and federal rulemakings.
How does this compare to past Chinese-linked removals?
The 200,000-account bust is more than seven times larger than the 28,400-account removal X disclosed in December 2024.
Who is investigating beyond X?
X said it shared indicators of compromise with peer platforms and the FBI’s Foreign Influence Task Force. The FBI and ODNI have both flagged Chinese AI-influence operations in recent threat assessments.
Was this article helpful?
Your feedback directly improves future articles on this site.
