AI Giants Seek Antitrust Exemption to Slow Development

Leading artificial intelligence companies are calling for permission to coordinate development timelines—a move that would require exemption from antitrust law and has sparked debate about whether safety concerns or commercial interests are driving the proposal.

In early September, Anthropic CEO Dario Amodei published an essay advocating that AI companies slow development of their most advanced models, citing inadequate safety safeguards. The position quickly gained endorsements from Sam Altman of OpenAI, Elon Musk of xAI, Demis Hassabis of Google DeepMind, and Microsoft's Satya Nadella. Altman announced OpenAI would delay its planned 2026 public offering to prioritize safety advances, according to Al Jazeera.

Why it matters

This debate extends beyond technical safety to fundamental questions of market structure and global governance. If granted antitrust exemption, a small group of American companies would gain unprecedented authority to set development standards for a transformative technology—potentially reshaping competitive dynamics while excluding international voices from decision-making processes that affect the entire world.

Commercial interests shape positions

The divide among tech leaders follows clear business lines. Companies selling AI models—Anthropic, OpenAI, xAI, and Google DeepMind—support slower development. Microsoft, which provides cloud infrastructure for these models, also backs the proposal.

Meanwhile, Nvidia CEO Jensen Huang opposes the slowdown, arguing market forces can drive safe innovation without new regulation. Nvidia sells the chips AI companies purchase before building anything, and the company has profited from intense competition that drives premium pricing. Meta's Mark Zuckerberg, whose company distributes models freely, warned the proposal could increase legal liability.

Antitrust concerns and regulatory questions

Under current antitrust law, these companies cannot hold private meetings to coordinate development pace or competitive practices. The proposal would require exemption from these rules, ostensibly for safety reasons.

David Sacks, the White House AI lead, has questioned both the competitive implications and the independence of nonprofit evaluators that would oversee compliance. His core concern: how can evaluation be truly independent when one company inspects another?

International dimensions

Amodei's proposal includes provisions to restrict China's access to advanced AI chips and semiconductor manufacturing equipment. China's foreign ministry characterized these warnings as Cold War-style fearmongering designed to hinder Chinese technological progress. Even Amodei acknowledged China presents the most difficult challenge for his plan.

Any agreement among five major American companies would not bind international competitors, raising questions about effectiveness.

Missing voices in the debate

PricewaterhouseCoopers estimates $31.6 trillion will flow into AI infrastructure globally by 2050. Countries offering abundant low-cost, reliable, low-carbon electricity stand to attract significant investment. Yet regions including the Arab world have been largely absent from discussions about AI governance.

Nations providing energy, land, and capital for AI infrastructure could leverage these resources to negotiate evaluation standards, operational conditions, and liability frameworks within their jurisdictions—challenging the assumption that development rules should be set exclusively by American companies.

Banking analogy falls short

Proponents compare the proposal to Basel III banking regulations implemented after the 2008 financial crisis. However, those rules were imposed by external regulators following demonstrated harm. The AI proposal involves competing companies coordinating before crisis occurs—a fundamentally different structure that more closely reflects industry interests than regulatory precedent.

These details were first reported by Al Jazeera in an opinion analysis examining the debate's commercial and geopolitical dimensions.