Full Breakdown
AI Leaders Call for a Global Slowdown; U.S. and China Push Back
By Drooid · · How we work
The Proposal: “Pace the Frontier”
On September 12 Anthropic chief executive Dario Amodei published an essay urging “frontier” AI labs to collectively slow model development. He proposes embedding independent safety evaluators, creating shared safety standards among democratic nations, and eventually negotiating a global framework that includes authoritarian governments.
Background and Context
The call follows a narrative of a U.S.–China AI arms race. Silicon-valley executives have warned that losing the race could be disastrous, and recent resignations—most notably a former Anthropic researcher—have amplified safety concerns. The U.S. has imposed export controls on advanced Nvidia chips and pursued legal actions against alleged illegal model distillation by Chinese labs. Chinese authorities, through the Standardization Administration of China and the Cyberspace Administration of China, have announced monitoring of AI development, acknowledging some of the risks raised by Amodei.
Data and Statistics
- Anthropic is preparing a Nasdaq IPO; analysts expect a single quarter of adjusted operating profit but note the company has not yet achieved steady profitability.
- OpenAI has postponed its IPO, citing safety work after a recent hack of a third-party service.
- Token prices for frontier models have fallen, with users shifting to cheaper alternatives such as DeepSeek V4 Pro and Kimi 3.
- A sell-off in AI-semiconductor stocks followed Amodei’s essay, but Lynx Equities analyst KC Rajkumar found no evidence of reduced GPU orders or supply-chain disruptions.
Official Statements & Responses
Chinese state media China Daily framed the proposal as an attempt to blunt China’s AI advance.
China’s Foreign Ministry spokesperson Guo Jiakun dismissed the discourse as “fear-mongering.”
Kevin Xu, founder of Interconnected Capital, said the AI safety community in China is growing and shares many concerns with American labs, but warned the vision could “subjugate and restrict China first.”
Criticism & Opposition
Industry leaders opposing a coordinated slowdown include:
- Aidan Gomez, CEO of Cohere, who warned a pact among a handful of labs could act as a “soft cartel” raising barriers for smaller competitors and open-source projects.
Conflicting Reports & Gaps
Analysts differ on market impact. While some cite a sharp decline in AI-semiconductor valuations as evidence of reduced demand, Rajkumar reports no observable slowdown in hardware orders, describing the sell-off as “macro-driven.”
Feasibility is disputed. Sandra Wachter of the Oxford Internet Institute called global coordination “highly unrealistic” given current geopolitical tensions, whereas Amodei’s essay assumes “government regulation and intervention” could eventually align democratic and authoritarian actors. No concrete enforcement mechanism has been presented.
Verbatim Quotes
- “Global pacing will require cooperation with China, the autocratic country with by far the most advanced AI capabilities,” — Dario Amodei
- “The AI safety community in China is growing and shares many of the same concerns that American labs do,” — Kevin Xu
- “Slowing the rate from ‘extremely fast’ to ‘only somewhat fast’ gives up relatively little strategic advantage, while potentially greatly improving safety,” — Dario Amodei
- “Nowhere in the Dario essay is there a call to slow down the pace of hardware roll-out,” — Rajkumar
What’s Next
- The Federal Open Market Committee meets on September 16-17; analysts expect the decision to influence AI-semiconductor valuations more than Amodei’s essay.
- U.S. and Chinese leaders are slated to meet later in September for a bilateral AI dialogue, though no timeline for a formal agreement has been set.
- Upcoming IPO filings from Anthropic and OpenAI will reveal whether the “pacing” strategy materially affects corporate finances and valuation.