China’s Media Machine Is Selling Americans on Killing Their Own Data Centers. Trump Is Not Buying It.
Beijing’s English-language outlets and China-linked accounts are pushing one story to U.S. audiences: AI data centers are driving up power bills, straining the grid, and enriching tech companies at ordinary people’s expense. At home, the Chinese state subsidizes power for its own AI data centers and races to build more.
That split is the tell.

A large U.S. data center complex. Facilities like these have become the focus of both local opposition and Chinese state media coverage. (Fair use / news illustration.) - Credit qz.com
The AI data center race is getting way more complicated
President Donald Trump has treated the backlash as a strategic problem. On Truth Social, he wrote that communities rejecting data centers risk ending up “backward and poor,” and that “China could not be happier with this anti-data center movement.” He later called the campaign against the facilities part of a broader effort to slow American AI and said the United States is leading China and intends to stay ahead.


Interior Secretary Doug Burgum has described a “concentrated information propaganda war” aimed at places trying to permit data centers. House Speaker Mike Johnson has called the panic a Chinese psyop. Senator Tom Cotton asked the Justice Department to investigate. In September, Ways and Means Chairman Jason Smith and China Select Committee Chairman John Moolenaar asked Treasury and the IRS whether tax-exempt groups tied to a Shanghai-based network are using charitable status to slow U.S. AI infrastructure.
Chinese state media’s message
CGTN America, the English arm of Chinese state television and a registered foreign agent, has run multiple segments blaming U.S. data centers for higher electricity prices.
Relevant CGTN videos:
AI data center boom drives U.S. electricity costs higher (November 2025)
Data centers take over affluent US cities (July 2026)
China Daily and Global Times have run parallel pieces. The New York Times and threat-intelligence firm Alethea have separately documented Chinese, Russian, and Iranian state media treating the U.S. data-center fight as a domestic fracture point.
Covert accounts and the Trump response
In June 2026, OpenAI banned ChatGPT accounts it assessed as likely originating in China that generated English posts and cartoons blaming data centers for household bills. In late August, X identified roughly 200,000 inauthentic Chinese accounts in influence operations, about 200 of them posting into the American AI and energy debate.
Trump’s “let Data Reign” post drew immediate coverage:
Modern designs undercut the propaganda line
The Chinese state-media narrative treats every data center as an old-style facility that gulps power from the grid and evaporates millions of gallons of water. That description is increasingly outdated.
Leading operators are deploying closed-loop liquid cooling and on-site power co-generation. Meta has publicly detailed closed-loop systems that recirculate a water-glycol mixture in a sealed circuit. Heat is rejected through dry coolers rather than evaporative towers. The company states that a typical new AI-optimized facility using this approach consumes less water annually than a couple of full-service restaurants. Nvidia’s DSX reference design for its Rubin-generation infrastructure goes further: 100 percent liquid cooling of every chip and networking component, with no fans in the system. In favorable climates the design can operate with dry coolers alone, reducing on-site water consumption from the roughly 2.6 million gallons per megawatt per year typical of conventional cooling-tower systems to near zero. Nvidia has said a 50-megawatt facility can save more than $4 million annually in cooling-related energy and water costs under the new architecture.
On the power side, developers are increasingly using “bring your own power” strategies. On-site generation—primarily natural gas turbines or fuel cells—can be paired with combined heat and power (CHP). Waste heat from the generators drives absorption chillers that provide cooling without drawing large amounts of additional electricity from the grid. A Johnson Controls reference design for gigawatt-scale AI factories using this approach projects a power usage effectiveness (PUE) as low as 1.23 with zero on-site water use. Some projects also capture waste heat for nearby district heating.
Nuclear co-location deals are advancing as well. Microsoft has a long-term agreement tied to the restart of the Crane Clean Energy Center (formerly Three Mile Island Unit 1). Amazon holds a multi-decade power purchase agreement with the Susquehanna nuclear plant in Pennsylvania. These arrangements do not eliminate the need for careful siting and grid planning, but they directly address the claim that every new facility simply loads the existing residential grid.
Older evaporative systems and pure grid-dependent designs still exist and can create local impacts. Newer projects are being engineered to minimize exactly the externalities Chinese outlets highlight. The propaganda rarely mentions the design shift.
Local opposition vs. foreign amplification
Americans do not need a foreign script to worry about power bills, water, or noise. Polls have shown strong local opposition. The documented activity is amplification of an existing grievance, not the invention of it. Much of the local opposition has been fueled by the Chinese propaganda.

Bitcoin Policy Institute reports have tied the Party for Socialism and Liberation (overlapping with the Neville Roy Singham network) to multiple campaigns that delayed or blocked projects. Singham’s network denies government direction; no charges have been filed in the public record. European philanthropic funding of some environmental groups is a separate track. Reference Shasta Unfiltered's earlier story on Neville Roy Singham: Click Here
Administration policy
A July 2025 executive order directed federal agencies to authorize data centers on suitable federal land. DOE named sites including Idaho National Laboratory, Oak Ridge, Paducah, and Savannah River. Hyperscalers signed a Ratepayer Protection Pledge to bring their own power and cover grid upgrades.
For North State readers, California already has high electricity prices and has seen local bans (including Monterey Park). The strategic question is whether the United States slows computing infrastructure while China subsidizes it—and while American operators adopt closed-loop cooling and on-site generation that reduce the very impacts being advertised abroad.
It should be noted that the Circle Six efficiency study pointed out that Redding or Shasta County could be an ideal location for a data center and could bring increased revenue into the area through high-paying tech jobs where current budgets are being strained by increased expenses and a stagnant economy.
Shasta Unfiltered will keep following the money, the moratoriums, the power contracts, and the actual engineering of the facilities. The propaganda line is simple. The record, including the designs now being deployed, is not.
Sources
OpenAI, “PRC-linked influence operations are targeting AI debates in the US,” June 10, 2026: https://openai.com/index/prc-linked-influence-operations-ai-debates/
OpenAI case study, “Data Center Bandwagon” Campaign: https://openai.com/index/disrupting-malicious-uses-of-ai-data-center-bandwagon/
Bitcoin Policy Institute, “Foreign Influence in the Campaign against American AI, Part I,” May 2026: https://www.btcpolicy.org/articles/foreign-influence-in-the-campaign-against-american-ai
Bitcoin Policy Institute, Part II (Singham Ground Game), June 2026: https://www.btcpolicy.org/articles/china-aligned-ground-game-stalled-or-blocked-23-6-billion-in-american-ai-infrastructure
House Ways and Means / Select Committee on China letter, Sept. 23, 2026: https://admin-waysandmeans.house.gov/2026/09/23/chairmen-smith-moolenaar-demand-investigation-into-chinese-communist-party-linked-network-bankrolling-sabotage-of-american-ai-infrastructure/
CGTN America, “Data centers drive up local electric bills across US,” Oct. 14, 2025: https://www.youtube.com/watch?v=wysLxM1n0Ew
New York Times, “China, Russia and Others Seek to Inflame Debate Over A.I. Data Centers,” July 9, 2026: https://www.nytimes.com/2026/07/09/business/china-russia-ai-data-centers.html
Alethea, research note on state actors amplifying data-center opposition: https://info.alethea.com/research-state-actors-ai-slop-data-center-revolt
Meta, “Closed-Loop Cooling Explained,” August 2026: https://about.fb.com/news/2026/08/closed-loop-cooling-explained-the-plumbing-behind-metas-ai/
Nvidia, liquid-cooling / DSX design announcements, June 2026: https://blogs.nvidia.com/blog/liquid-cooling-ai-factories/
Johnson Controls, AI Factory CHP / absorption cooling reference design, 2026: https://www.johnsoncontrols.com/building-insights/feature-story/a-global-blueprint-for-how-gigawatt-scale-ai-factories-can-reuse-waste-heat-with-absorption-chillers
ABC News, “Trump goes all in on controversial data centers”: https://www.youtube.com/watch?v=Jl8H5X1Foxk
Truth Social archive of Trump’s Aug. 31, 2026 post: https://trumpstruth.org/statuses/41411







