Blab AI
US National Security Official Calls for Bombing Chinese Data Centers to Stop AGI Development
When a former White House national‑security adviser warned that the United States might have to bomb Chinese data‑center facilities to halt the rise of artificial general intelligence (AGI), the headline itself felt like a headline from a spy novel. The warning, buried in the Center for a New American Security’s (CNAS) Superpowers and AGI report, was released on 27 August 2026 and has already sparked debate among policymakers, academics, and industry observers.
The report, written by Jacob Stokes—a former adviser in the Obama administration—frames AGI as an intelligence that can match or surpass human reasoning across all domains. Stokes argues that such a capability could pose an existential threat, and he urges the United States to prepare for a scenario in which China becomes the first nation to achieve AGI. He states that the U.S. must be ready to take “any necessary action” to prevent that outcome.
Stokes lays out a four‑step escalation plan. The first two steps are familiar tools of international competition: diplomacy and espionage. The third step shifts to the digital battlefield, calling for offensive cyber operations aimed at sabotaging China’s leading AGI systems. The report notes that cyber attacks “might not provoke large‑scale retaliation” and therefore could be a preferred option.
If cyber measures fail, the plan escalates to kinetic attacks—missile or bomb strikes on Chinese data‑center sites. Stokes cites the United States’ 2023 missile strike on Iranian Amazon Web Services (AWS) data centers as a precedent, arguing that such strikes can effectively shut down AI infrastructure. The report warns, however, that bombing Chinese data centers would constitute a declaration of war and would almost certainly trigger retaliatory action. Stokes acknowledges that China could respond with its own kinetic, cyber, or economic measures, potentially spiraling into a broader conflict.
A core challenge highlighted in the report is verifying whether China has indeed achieved AGI. Stokes points out that China’s AI ecosystem is dispersed across universities, research labs, cloud providers, and private companies, making it hard to assess progress. He questions whether the United States can be trusted to use AGI responsibly, and whether China could justifiably take pre‑emptive action.
Reactions to the report have been mixed. Some experts argue that bombing data centers is unrealistic, given the distributed nature of China’s AI infrastructure and the high risk of escalation. Others see the report as a reflection of a broader debate about how to manage the potential risks of AGI.
No U.S. government agency has announced a policy change in response to the report. The White House, the National Security Council, and the Department of Defense have not issued statements addressing the recommendations. CNAS has stated that the report is intended to stimulate discussion rather than prescribe immediate action.
As of September 2026, the United States remains engaged in a competitive race with China over AI development. Current measures include export controls on advanced semiconductors, increased funding for domestic AI research, and diplomatic efforts to shape international norms. The debate over whether kinetic action against data centers is a viable or prudent option continues.
Key unresolved questions include how the United States can verify China’s AGI progress, what the legal and strategic ramifications of bombing a foreign nation’s data‑center facilities would be, and how such an action might affect the broader U.S.–China relationship and global AI governance.
The Superpowers and AGI report has added a new dimension to the ongoing conversation about AI risk, highlighting the tension between national security interests and the potential dangers of advanced artificial intelligence.
The report, written by Jacob Stokes—a former adviser in the Obama administration—frames AGI as an intelligence that can match or surpass human reasoning across all domains. Stokes argues that such a capability could pose an existential threat, and he urges the United States to prepare for a scenario in which China becomes the first nation to achieve AGI. He states that the U.S. must be ready to take “any necessary action” to prevent that outcome.
Stokes lays out a four‑step escalation plan. The first two steps are familiar tools of international competition: diplomacy and espionage. The third step shifts to the digital battlefield, calling for offensive cyber operations aimed at sabotaging China’s leading AGI systems. The report notes that cyber attacks “might not provoke large‑scale retaliation” and therefore could be a preferred option.
If cyber measures fail, the plan escalates to kinetic attacks—missile or bomb strikes on Chinese data‑center sites. Stokes cites the United States’ 2023 missile strike on Iranian Amazon Web Services (AWS) data centers as a precedent, arguing that such strikes can effectively shut down AI infrastructure. The report warns, however, that bombing Chinese data centers would constitute a declaration of war and would almost certainly trigger retaliatory action. Stokes acknowledges that China could respond with its own kinetic, cyber, or economic measures, potentially spiraling into a broader conflict.
A core challenge highlighted in the report is verifying whether China has indeed achieved AGI. Stokes points out that China’s AI ecosystem is dispersed across universities, research labs, cloud providers, and private companies, making it hard to assess progress. He questions whether the United States can be trusted to use AGI responsibly, and whether China could justifiably take pre‑emptive action.
Reactions to the report have been mixed. Some experts argue that bombing data centers is unrealistic, given the distributed nature of China’s AI infrastructure and the high risk of escalation. Others see the report as a reflection of a broader debate about how to manage the potential risks of AGI.
No U.S. government agency has announced a policy change in response to the report. The White House, the National Security Council, and the Department of Defense have not issued statements addressing the recommendations. CNAS has stated that the report is intended to stimulate discussion rather than prescribe immediate action.
As of September 2026, the United States remains engaged in a competitive race with China over AI development. Current measures include export controls on advanced semiconductors, increased funding for domestic AI research, and diplomatic efforts to shape international norms. The debate over whether kinetic action against data centers is a viable or prudent option continues.
Key unresolved questions include how the United States can verify China’s AGI progress, what the legal and strategic ramifications of bombing a foreign nation’s data‑center facilities would be, and how such an action might affect the broader U.S.–China relationship and global AI governance.
The Superpowers and AGI report has added a new dimension to the ongoing conversation about AI risk, highlighting the tension between national security interests and the potential dangers of advanced artificial intelligence.
