U.S. directive on Anthropic sparks sovereign AI debate in India

Anthropic announced late Friday that it has suspended access to its Fable 5 and Mythos 5 models for foreign nationals following a U.S. government directive. The directive follows a partnership between Anthropic and Tata Consultancy Services to expand enterprise AI adoption in India. While Anthropic disputes the government’s characterization of the situation, the White House is privately blaming Anthropic’s handling of alleged jailbreak vulnerabilities. According to some reports, security concerns were reported to the government by Amazon CEO Andy Jassy. The White House is unlikely to extend similar restrictions to other AI companies, according to The Information. The restriction on these frontier AI models—defined as advanced AI models—has affected the South Asian nation, which represents a major market for these systems.

The move has triggered a debate among Indian founders and investors regarding sovereign AI, which refers to the concept of domestic control over AI capabilities. For startups with distributed teams, the restriction introduces operational hurdles. Vijay Rayapati, the co-founder and CEO of Atomicwork, noted that the situation creates a competitive disadvantage. Atomicwork, which has around 25 employees in the U.S. but keeps its engineering teams in India, faces exposure to these policy shifts. Rayapati warned that if an AI team is not made up entirely of U.S. citizens, companies are at a competitive disadvantage. Aakrit Vaish, the founder of Indian AI venture platform Activate, reacted to the announcement on Saturday morning, stating that the directive completely changes the landscape and alters how the country must approach sovereign AI.

In response, Indian technology leaders are urging the country to accelerate its domestic capabilities. Sridhar Vembu, the founder of Indian SaaS company Zoho, warned that technology is the ultimate weapon and urged Indian organizations to embrace smaller and open-source models. Investor and former Infosys executive Mohandas Pai argued that India is way behind and needs a national mission to get going quickly. To address this, Pai proposed a funding overhaul that dwarfs India’s current initiatives.

The proposed financial measures include:

• An annual fund of ₹500 billion (about $5 billion) for AI and deep tech.
• A credit guarantee program of ₹2 trillion (around $21 billion).

By contrast, New Delhi approved the IndiaAI Mission—a government initiative for AI development—in 2024 with an outlay of ₹103.72 billion (about $1.2 billion) over five years. Prasanto Roy, a New Delhi-based technology policy expert, noted that “American AI models are bound to American geopolitics.” He added that even if the decision is reversed, the episode demonstrates that foreign large language models are never geopolitically neutral, reinforcing the need for strategic autonomy.

Why it matters

The U.S. government directive forcing Anthropic to restrict access to its models for foreign nationals has sparked a debate in India regarding technological sovereignty and the risks of relying on foreign AI providers.