Anthropic's Model Suspension Highlights India's AI Dependence Dilemma

Anthropic's recent decision to suspend access to its Fable 5 and Mythos 5 models for all foreign nationals, including its own foreign employees, following a U.S. government directive, is more than a temporary access issue. It is a clear example of how geopolitical decisions can directly impact the availability of critical AI infrastructure. The move, announced late Friday, came shortly after Anthropic partnered with Indian IT services giant Tata Consultancy Services to expand enterprise AI adoption in India. This timing underscores how tightly India's AI ambitions have become linked to technologies developed and governed in the United States.

The immediate cause, as reported, involves security concerns first flagged to the U.S. government by Amazon CEO Andy Jassy. According to The Information, the White House is unlikely to extend similar restrictions to other AI companies and is privately blaming Anthropic's handling of alleged jailbreak vulnerabilities. Anthropic has disputed the government's characterization and argued the action should not have been taken. Regardless, the episode has triggered a broader discussion among Indian founders, investors, and policy experts about the country's long-term AI strategy.

For many in India's tech sector, the announcement was a wake-up call on technological dependence. Aakrit Vaish, founder of Indian AI venture platform Activate, described waking up "shocked and confused" and said the decision "completely changes things." He argued it strengthens the case for developing domestic AI capabilities and plans to encourage companies in his portfolio to reduce reliance on a small number of frontier AI providers. This sentiment echoes a growing concern that access to increasingly critical AI systems can be shaped by decisions made far beyond India's borders.

The concern extends to startups with global teams. Vijay Rayapati, co-founder and CEO of Atomicwork, highlighted the risks facing companies whose teams span multiple countries. With about 25 employees in the U.S. and product engineering based in Bengaluru, Atomicwork could face competitive disadvantages if access to frontier AI models becomes subject to geopolitical restrictions. "If your AI team is not made up entirely of U.S. citizens, you are at a competitive disadvantage," Rayapati said. This anxiety is compounded by recent events like Opendoor shutting its India office, with CEO Kaz Nejatian citing a shift toward smaller AI-native teams, adding to questions about AI's impact on global talent hubs.

Beyond individual startups, the episode has prompted calls for a more robust national AI strategy. Sridhar Vembu, founder of Zoho, stated that "technology is the ultimate weapon" and urged Indian organizations to embrace smaller and open-source models, both Indian and Chinese. Mohandas Pai, an investor and former Infosys executive, called for the government to create an annual ₹500 billion (about $5 billion) fund for AI and deep tech, alongside a ₹2 trillion credit guarantee program. This proposal would significantly expand India's existing efforts, such as the IndiaAI Mission, which has an outlay of ₹103.72 billion over five years.

However, not everyone agrees that capital is the primary constraint. Lightspeed partner Hemant Mohapatra argued that the biggest hurdles to building globally competitive AI companies are talent, access to computing resources, and execution. He noted that training a frontier model can cost hundreds of millions to several billion dollars but that successful companies typically scale their capital needs over time. This perspective suggests that simply increasing funding may not be sufficient without addressing ecosystem challenges.

India's current position in frontier model development remains limited. Only a few startups, like Sarvam, are pursuing foundational models, while others like Krutrim have pivoted toward cloud and AI infrastructure. Much of the ecosystem focuses on applications and specialized models built on existing foundations, as seen with Avataar AI's recent video-generation model. This application-layer strength may offer resilience but also highlights dependence on underlying models developed abroad.

The broader implication, as noted by technology policy expert Prasanto Roy, is that the Anthropic episode reinforces concerns about strategic autonomy. He compared it to the lesson many countries drew from Russia's loss of access to SWIFT after its invasion of Ukraine, suggesting that no foreign LLM can be considered geopolitically neutral. "American AI models are bound to American geopolitics," Roy said, predicting a nationalist backlash in India and describing the move as a poorly considered decision by Washington with consequences extending far beyond Anthropic itself.

This debate reflects a pattern seen globally, where nations are increasingly questioning reliance on foreign technology providers for critical infrastructure. For India, one of the most important markets for frontier AI companies, the challenge is to balance immediate access to advanced AI with long-term goals of technological self-reliance. The outcome will likely influence not only India's AI trajectory but also how other emerging tech markets navigate similar dependencies.