Anthropic Freezes Models: India Confronts Its AI Future

Anthropic's abrupt decision to halt access to its newest artificial intelligence models, following a directive from the U.S. government, has ignited significant discussion across the global technology sector. In India, this action has rekindled a long-standing debate about whether one of the world's largest AI markets can sustainably rely on technologies developed and governed by external entities.

The announcement, made late on a Friday, revealed that Anthropic had received a U.S. government mandate requiring the suspension of access to its recently launched Fable 5 and Mythos 5 models for all foreign nationals, including its own non-U.S. employees. This development occurred shortly after Anthropic publicized a partnership with Indian IT services giant Tata Consultancy Services, aimed at accelerating enterprise AI adoption in India, thereby underscoring the deep integration of India's AI ambitions with technologies originating from and controlled by the U.S.

While the full scope of implications remains uncertain, some reports suggest that the initial security concerns were brought to the government's attention by Amazon CEO Andy Jassy. The Information further indicated that the White House is not expected to impose similar restrictions on other AI companies, reportedly attributing the issue privately to Anthropic's management of alleged jailbreak vulnerabilities. Anthropic, for its part, has contested the government's portrayal of events and maintained that the action was unwarranted.

Nevertheless, this incident has sparked considerable debate among Indian founders, investors, and policy experts regarding the nation's strategic direction: should it intensify efforts to cultivate indigenous AI capabilities, increase investment in open-source alternatives, or continue its reliance on a limited number of U.S. frontier model providers? For some, the event serves as a stark reminder of technological dependence. For others, it highlights how access to increasingly vital AI systems can be influenced by geopolitical decisions beyond India's direct control.

India has emerged as a crucial market for frontier AI companies. Both Anthropic and OpenAI have identified the South Asian nation as their second-largest market globally, surpassed only by the U.S., which reflects its growing significance in the international AI landscape. These companies have recently established offices in India, expanded local hiring, forged partnerships, and launched enterprise initiatives, banking on India’s extensive talent pool of developers, startups, and businesses to accelerate the adoption of their latest technologies.

For many within India's technology sector, Anthropic's Friday announcement transcended the actions of a single AI company. It reopened critical questions concerning the country's long-term AI strategy and the sustainability of India's dependence on a small group of foreign frontier AI providers.

“It completely changes things,” remarked Aakrit Vaish, founder of the Indian AI venture platform Activate, in reference to Anthropic's decision. “I think this materially changes the way all of us should be thinking about sovereign AI in India.”

Vaish conveyed to TechCrunch his "shock and confusion" upon learning of the announcement on Saturday morning, asserting that it significantly strengthens the argument for developing domestic AI capabilities. He anticipates a growing trend among startups to adopt open-source models and plans to advise companies within his portfolio to reduce their reliance on a few dominant frontier AI providers.

For some founders, a more profound concern centered on the potential impact of restrictions on frontier AI access on competitive dynamics. Vijay Rayapati, co-founder and CEO of Atomicwork, informed TechCrunch that the incident underscored the risks confronting startups with globally distributed teams if access to advanced AI systems increasingly becomes subject to geopolitical constraints.

Atomicwork employs approximately 25 individuals in the U.S., though a substantial portion of its product engineering team is based in Bengaluru, India.

“If your AI team is not made up entirely of U.S. citizens, you are at a competitive disadvantage,” Rayapati stated, contending that unequal access to frontier AI models could furnish certain companies with a notable advantage over their competitors.

This apprehension arises as segments of India's tech sector are already grappling with how AI might redefine the economics of global talent. This week, U.S. real estate technology firm Opendoor closed its India office less than two years after expanding into the country. CEO Kaz Nejatian cited a strategic move to bring operational functions closer to U.S. customers and a pivot towards smaller, AI-native teams.

While Opendoor did not explicitly detail the extent to which AI-driven efficiencies influenced its decision, the move contributed to a broader discussion about how advancements in AI could impact the future of global technology work and its implications for India’s standing as a prominent engineering talent hub.

Beyond startups and AI developers, the Anthropic situation also catalyzed a wider discourse among India's technology leaders concerning the nation's reliance on foreign AI infrastructure.

Sridhar Vembu, founder of the Indian SaaS company Zoho, asserted that the move demonstrated “technology is the ultimate weapon” and strongly encouraged Indian organizations to increasingly adopt smaller and open-source models.

“What can our government do right now? Ensure that orgs in India embrace smaller models, both Indian and Chinese open source ones,” Vembu posted on X.

Investor and former Infosys executive Mohandas Pai responded to Vembu on X, arguing that the development highlighted the imperative for a far more ambitious national AI strategy. He called upon the government to significantly escalate investments in AI, computing infrastructure, and deep technology.

“We are way behind and need a national mission to get going quickly,” Pai wrote, proposing that the government establish an annual ₹500 billion (approximately $5 billion) fund for AI and deep tech, complemented by a ₹2 trillion (around $21 billion) credit guarantee program to bolster cloud infrastructure, hardware, and semiconductor development.

Pai’s proposed financial commitment would substantially overshadow India’s current AI endeavors. In 2024, New Delhi approved the IndiaAI Mission with an allocation of ₹103.72 billion (about $1.2 billion) over five years, specifically aimed at expanding compute infrastructure, supporting startups, and fostering indigenous AI capabilities.

Despite growing interest in AI and the Indian government's push for domestic capabilities, India remains a relatively minor participant in frontier model development. Only a handful of startups are actively pursuing foundational AI models, including Sarvam, which released open-source models earlier this year. Conversely, another high-profile AI startup, Krutrim, shifted its focus towards cloud and AI infrastructure services after initially aiming for foundational model development.

Much of India’s AI ecosystem has instead concentrated on developing applications and specialized models built upon existing foundation models. A recent example is Avataar AI, which launched a video-generation model this week, positioning it as a lower-cost alternative to offerings from competitors such as Google’s Veo, Kling, Luma, and Runway.

However, not everyone concurs that a lack of capital is the primary obstacle. In response to Pai’s remarks, Lightspeed partner Hemant Mohapatra contended that the most significant constraints to building globally competitive AI companies are talent, access to computing resources, and effective execution, rather than merely the magnitude of investment commitments.

Mohapatra estimated that training a frontier AI model could range from hundreds of millions to several billion dollars, depending on the methodology, but noted that successful AI companies have historically scaled their capital requirements incrementally as adoption has grown.

Nonetheless, for some policy observers, the implications of this incident extend well beyond AI startups or model providers.

Prasanto Roy, a New Delhi-based technology policy expert who consults for multinational corporations, suggested that the episode would likely intensify concerns within the Indian government regarding strategic autonomy. He drew a parallel to the lessons many countries learned from Russia’s loss of access to SWIFT and other components of the global financial system following its invasion of Ukraine.

He informed TechCrunch that the move is likely to provoke a significant nationalist backlash in India and characterized it as a poorly considered decision by Washington, with consequences far exceeding Anthropic itself.

“Even if this is corrected or reversed, the Anthropic episode shows there’s no such thing as a geopolitically neutral foreign LLM,” Roy stated. “American AI models are bound to American geopolitics.”