Anthropic, OpenAI, and DeepMind call for AI development brake; Cohere CEO cries 'cartel'

Anthropic CEO Dario Amodei published an essay requesting a brake on the pace of large language model development, warning that the technology poses looming dangers including bioterrorism, cyberattacks, and economic disruption. [2] The leaders of three other major artificial intelligence laboratories—OpenAI CEO Sam Altman, Google DeepMind chairman Demis Hassabis, and SpaceXAI CEO Elon Musk—expressed support for the essay, with Musk publicly agreeing on the social media platform X. [2] This coordinated messaging represents a shift for the industry, considering that OpenAI and Anthropic have historically competed in a winner-takes-all race. [2]

Cohere co-founder and CEO Aidan Gomez published a response on September 13, 2026, stating that artificial intelligence requires evidence-based standards rather than rules dictated by a cartel. [3] Gomez argued that a small group of market-dominant Silicon Valley companies should not define global safety standards while simultaneously dictating the speed of technological progress for the rest of the world. [3] He accused these leading firms of operating as an oligopoly that uses public fear to justify softening competition rules and dictating terms for everyone else. [3] Despite his firm opposition to the proposed development brake, Gomez acknowledged that offensive cyberattacks are becoming cheaper faster than defensive measures can improve. [3]

The public calls for an artificial intelligence development slowdown from Amodei triggered an immediate negative reaction in global financial markets, causing major equity markets across Asia and New York to sink. [5] South Korea's KOSPI index experienced significant declines by closing down 3.3 percent, while the United States Nasdaq index also fell by 0.8 percent on the news. [5] During the ensuing market selloff in the United States, semiconductor stocks took the heaviest losses among technology shares, whereas software company names rallied sharply during the exact same trading period. [5]

The sudden alignment among rival frontier labs suggests a coordinated effort to shape impending regulation before governments unilaterally impose it. By framing themselves as the responsible stewards of dangerous technology, the largest players may inadvertently lock out smaller competitors, precisely as Cohere’s Aidan Gomez warns. This dynamic mirrors historical regulatory capture, where dominant firms use safety standards to raise the barrier to entry for challengers. If governments adopt Amodei’s proposed braking mechanisms, chipmakers could face sustained headwinds as compute demand softens, explaining the immediate semiconductor market selloff. What the sources don't address: How exactly an international development brake would be enforced without inadvertently shifting the center of AI innovation to jurisdictions that refuse to participate.