King Charles hosts the frontier labs while Washington won't
Monday's pacing fallout went two directions at once: the one Western figure with no electoral stake in AI is convening the labs, while Beijing upgraded its rebuttal to an accusation — and quietly already regulates for Amodei's worst case.
Britain's King Charles will host leaders of Nvidia, Google DeepMind, OpenAI and Anthropic, along with the UK's AI minister, this week to discuss whether the technology needs a shared set of principles. Buckingham Palace framed the agenda as how AI can benefit society and contribute "to the flourishing of both people and the planet," the monarch's newest use of a convening power he built on environmental briefings. The timing is the story: it lands two days after Dario Amodei's essay, which his own researcher called an extinction risk and which Trump dismissed on Sunday with "whoever wins AI, wins." With Congress declining the job — Speaker Mike Johnson's position is that safety is the companies' problem, not the legislature's — the only head of state actively hosting the frontier is one with zero authority to bind it. The UK's own posture is light-touch regulation aligned with Washington, and its AI Security Institute sees models pre-deployment only under voluntary agreements. A royal invitation with no penalty behind it is the honest shape of what "global consensus" currently means: reputational, voluntary, and chaired by nobody who can enforce it. What the labs accept in a palace may define what they refuse in a treaty — the September 24 Trump-Xi meeting is the venue with actual teeth.
Beijing's rebuttal to the pacing push hardened Monday from dismissal into a specific counter-accusation. After spokesperson Guo Jiakun said Monday that "fomenting various threats, engaging in confrontation and malicious competition will only disrupt the process of global governance of artificial intelligence," the state-backed Global Times called Amodei's essay a "Cold War playbook" for AI — "hypocritical and short-sighted," aimed at curbing China "through technological barriers and regulatory monopolies" and excluding it from AI governance. The targeting is precise: the essay urges the US to keep chip export controls and crack down on distillation by Chinese labs, which is why the Asia Society's Lizzi C. Lee frames Beijing's sticking point as one question — what frontier-safety cooperation is supposed to look like when one side is simultaneously restricting your access to frontier compute. We tracked Monday's first market pricing of the slowdown debate — SoftBank sheds 10% as the AI slowdown call hits the tape — and the diplomatic track runs through the US-China AI safety talks scheduled before the summit.
While Beijing calls Amodei's scenario hypocrisy, its regulators have written that scenario into law for two years. A Reuters analysis published today details the trail: the Cyberspace Administration's September 2024 national AI safety framework flagged an explicit future loss-of-control case — AI autonomously acquiring resources, replicating itself, developing self-awareness and seeking power — and the expanded September 2025 version sharpened it to a sudden intelligence "leap," adding a governance principle translated as "trusted application, preventing loss of control." China's May agent guidelines require developers to detect, intervene in, block and recover from misbehaving agents, with users keeping final decision authority; Xi told the World Artificial Intelligence Conference in July that AI must "always remain under human control." Beijing stops short of Amodei's embedded-evaluator model but permits third-party safety assessments and outside audits of open models. Read together with the state security minister's essay naming Claude Mythos and GPT-5.5-Cyber as infrastructure threats, the posture is coherent: reject the pacing proposal, legislate the risk.
What to watch: whether the palace meeting produces any text at all, and whether AI governance appears as a deliverable — or a ceasefire demand — at the September 24 summit.
One government says the risk is exaggerated, one says the warning is a weapon, one hosts a chat about it. If you could hand the pacing decision to exactly one of those rooms, which door do you knock on?
Sources: Reuters — King Charles to host AI executives · New York Times — King Charles to convene AI leaders · Reuters — Global Times blasts Anthropic's calls as 'Cold War' tactic · AP via NPR — Beijing hits back at Anthropic CEO · Reuters — How China is preparing for the risk of AI escaping human control · Modern Diplomacy — China's AI safety strategy · Bloomberg — China rejects AI fearmongering
