US engages in high-stakes talks with India to regulate Anthropic AI models

The US Commerce Department dropped a regulatory bomb on June 12, issuing an export control order that restricts foreign nationals from accessing Anthropic’s most advanced AI models, Fable 5 and Mythos 5. By the next day, Anthropic had disabled those models globally, cutting off users worldwide before the company even had detailed guidance from authorities on how to comply.

The fallout has been swift, particularly in India, where Anthropic had been actively building enterprise partnerships. Now Washington and New Delhi find themselves in tense discussions over how, and whether, Indian firms and developers can regain access to the technology they were just starting to integrate into their operations.

What happened and why it matters

The export control order from Commerce Secretary Howard Lutnick’s department cited national security concerns. Specifically, the government flagged risks around potential jailbreaks of advanced AI systems, where bad actors could manipulate models into bypassing their safety guardrails.

Anthropic’s status as a Pentagon contractor added another layer of sensitivity. The company’s ties to US defense infrastructure made its most capable models a natural target for tighter controls, particularly as those models grew more powerful and more widely deployed abroad.

For India, the timing was particularly painful. Anthropic had recently partnered with Tata Consultancy Services to expand enterprise AI offerings in the region. TCS is one of India’s largest IT services companies, and the partnership signaled Anthropic’s serious intent to capture a share of India’s rapidly growing AI market.

The diplomatic picture

Prior discussions between Washington and New Delhi in May 2026 had already touched on access to Anthropic’s Mythos model. But those earlier talks were focused on access arrangements, not the kind of regulatory framework negotiations that the current situation demands. No confirmed reports indicate ongoing high-stakes bilateral negotiations between the US and India regarding the regulation of Anthropic’s AI models as of now.

Washington’s broader approach to AI governance has favored bilateral partnerships over multilateral frameworks. The US has largely rejected the idea of global AI governance bodies, preferring to cut deals country by country where it can leverage its position as the world’s dominant AI exporter.

India, for its part, has been championing a different vision. New Delhi has increasingly emphasized AI sovereignty, the idea that countries should develop homegrown AI capabilities rather than depending on foreign providers who can, as this episode demonstrates, pull the plug overnight.

What this means for investors and the tech landscape

For Indian AI startups and infrastructure companies, the situation could be a catalyst. Government policy in India has already been trending toward supporting domestic AI development, and the spectacle of a US company cutting off access to its models overnight gives policymakers a compelling argument for why that investment needs to accelerate.

The broader signal here is about platform risk. Any enterprise, in any country, that builds critical workflows on top of a US-based AI model now has to price in the possibility that geopolitical decisions could sever access with minimal warning.

The key variable to watch is whether Washington and New Delhi can agree on a regulatory framework that restores access for Indian enterprises while satisfying US national security concerns. If they can, the Anthropic-TCS partnership and others like it could resume with new guardrails in place.

Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our Editorial Policy.