Overview

The Reuters article by Louis Juricic, published on 4 September 2026 at 09:36 pm, examines how the U.S.–China rivalry over artificial intelligence has escalated from economic competition to overt threats of military action against AI infrastructure.

Key Developments

  • Former Global Times editor‑in‑chief Hu Xijin used his X (formerly Twitter) account to warn that if the United States were to launch a military strike on Chinese data centres, China would retaliate with missile strikes on the U.S. mainland, not limited to the Asia‑Pacific region.
  • Jacob Stokes, deputy director of the Indo‑Pacific Security Program at the Centre for a New American Security (CNAS) and a former Obama‑administration national‑security staffer, released a report titled “Superpowers and AGI.” In the report he called for the United States to develop contingency plans for pre‑emptive military strikes on Chinese AI data centres, as well as state‑backed espionage and cyber‑operations designed to prevent Beijing from achieving artificial general intelligence (AGI) first.
  • Stokes argued that policymakers should “work backwards” from the unique nature of AI technology, drawing a parallel with the nuclear‑arms‑race era, and urged the Department of Defense and the National Security Agency to assess the intelligence thresholds required to justify extreme measures, including preventive strikes before China attains AGI.
  • Neither the Department of Defense, the National Security Agency, nor the Chinese government has issued an official response to Stokes’s proposals.
  • According to Epoch AI, the United States currently enjoys an estimated seven‑month lead over China in frontier AI model quality. Stokes cautioned that China could achieve a “surprise technological breakthrough,” especially through advances in robotics and embodied AI.
  • The report also suggested that reverse‑engineering or stealing Chinese AGI technology could be justified “legally and normatively” given China’s alleged history of intellectual‑property theft, but Stokes acknowledged that the U.S. intelligence community is not presently organized for long‑term technological espionage, implying a need for structural reform.
  • Arms analyst William Hartung of the Project on Government Oversight warned that strikes on Chinese data centres would risk igniting a shooting war between two nuclear‑armed powers, describing the idea as “the height of recklessness.”
  • On the diplomatic front, a U.S.–China summit between President Donald Trump and President Xi Jinping is scheduled for 24 September 2026 in Washington. Treasury Secretary Scott Bessent confirmed that AI governance will be on the agenda, as reported by CNBC on 2 September. Stokes described his outlook for the summit as muted, hoping for a “constructive” dialogue rather than a breakthrough in governance.
  • The article notes market reactions with Nvidia (NVDA) up 1.12 % and an unnamed index (VX) down 1.29 % following the coverage.
  • OpenAI President Greg Brockman announced the launch of GPT‑6 Astra on 3 September 2026, declaring “Welcome to the AGI era,” a development that intensifies the urgency of the scenarios outlined by Stokes.

Contextual Remarks

The piece underscores a shift from a purely economic competition over microchips to a scenario where AI infrastructure could become a target of kinetic military action, echoing concerns traditionally associated with nuclear deterrence. No official statements from the U.S. defense or intelligence establishments or from Beijing have been recorded at the time of writing.