As Anthropic halts access to new models, India grapples with its AI future

Anthropic’s sudden move to freeze access to its new AI models following a US government order has raised new questions for the global tech industry. In India, the decision sparked a long-running debate over whether one of the world’s largest AI markets can afford to rely on technology developed and controlled elsewhere.

The announcement came late Friday, when Anthropic said it had received a US government order that required it to stop access to its newly launched Fable 5 and Mythos 5 models to all foreigners, including its foreign workers. The move came shortly after the company announced a partnership with India’s IT services giant Tata Consultancy Services to increase business adoption of AI in India, underscoring how closely the country’s AI ambitions are tied to technologies developed and controlled in the US.

While the wider implications are still unclear, some reports said the first security concerns were reported to the government by Amazon CEO Andy Jassy. And The Information said the White House is unlikely to extend similar restrictions to other AI companies and has privately criticized Anthropic’s handling of jailbreak risks. Anthropic objected to the way the government was working and said that this step should not have been taken.

Still, the development has sparked a debate among Indian innovators, investors, and policy experts about whether the country should accelerate efforts to build domestic AI capabilities, deepen investments in open source alternatives, or continue to rely on a handful of US border model suppliers. For some, the episode is a wake-up call to technological dependency. For others, it’s a reminder that access to critical AI systems can be shaped by decisions of a country beyond India’s control.

India has become one of the most important markets for frontier AI companies. Anthropic and OpenAI both described the South Asian country as their second largest market after the US, indicating its growing importance in the global AI race. Companies have already set up offices in India, expanded local hiring, partnerships, and business plans in recent months, betting on India’s large base of developers, startups, and businesses to accelerate the adoption of their latest technologies.

For many in India’s tech sector, Friday’s Anthropic announcement was about one AI company. It has re-opened questions about the country’s long-term AI strategy and whether India can remain dependent on a small number of external AI suppliers.

“It’s completely game-changing,” said Aakrit Vaish, founder of Indian AI venture platform Activate, referring to Anthropic’s decision. “I think this is changing the way we all have to think about dominant AI in India.”

Vaish told TechCrunch that he woke up Saturday morning “shocked and confused” by the announcement and said it strengthens the case for developing domestic AI capabilities. He expects startups to increasingly turn to open source models and plans to encourage companies in his portfolio to reduce their dependence on a small number of frontier AI providers.

For some innovators, the biggest concern was what AI’s frontier access restrictions would mean for competition. Vijay Rayapati, co-founder and CEO of Atomicwork, told TechCrunch that the episode highlights the risks facing startups whose teams span multiple countries if access to advanced AI systems is increasingly subject to country restrictions.

Atomicwork has about 25 employees in the US, although most of their product engineering team is based in Bengaluru, India.

“If your AI team isn’t made up entirely of US citizens, you’re at a competitive disadvantage,” Rayapati said, arguing that unequal access to AI models across the border could give some companies a significant edge over competitors.

The concern comes as parts of India’s tech sector are already grappling with questions about how AI can reshape the global talent economy. This week, US real estate technology firm Opendoor closed its India office less than two years after expanding into the country, with CEO Kaz Nejatian citing a push to bring operations closer to US clients and a shift to smaller indigenous AI teams.

While Opendoor didn’t specify how much of the decision was motivated by AI-related efficiencies, the move added to a broader discussion about how advances in AI could affect the future of the world’s tech workforce and what that could mean for India’s position as a hub for engineering talent.

Without Anthropic

In addition to AI startups and developers, the Anthropic episode also sparked a wider debate among India’s tech leaders about reliance on external AI infrastructure.

Sridhar Vembu, founder of Indian SaaS company Zoho, said the move shows that “technology is the ultimate weapon” and urged Indian organizations to embrace lean and open models.

“What can our government do right now? Make sure that orgs in India accept small, Indian and Chinese models,” Vembu wrote in X.

Investor and former Infosys CEO Mohandas Pai responded to Vembu on X, saying the development highlights the need for a more ambitious national AI strategy and calling on the government to significantly increase investment in AI, computing infrastructure, and deep technology.

“We are lagging behind and need a national strategy to move forward quickly,” Pai wrote, urging the government to create an annual fund of R500 billion (about $5 billion) for AI and deep technologies, and a ₹2 trillion (about $21 billion) debt guarantee program to support cloud infrastructure, hardware, and semiconductor development.

Pai’s proposal would curtail India’s existing AI efforts. In 2024, New Delhi approved the IndiaAI Mission with a budget of 103.72 billion (about $1.2 billion) over five years, aimed at expanding computing infrastructure, supporting startups, and developing indigenous AI capabilities.

Despite growing interest in AI and New Delhi’s push to develop capabilities at home, India remains a minor player in the development of the frontier model. Only a few startups are pursuing basic AI models, including Sarvam, which released open-source models earlier this year. However, another high-profile AI startup, Krutrim, is looking into cloud infrastructure and AI services after positioning itself for basic model development.

Much of India’s AI ecosystem is instead focused on applications and specialized models built on top of existing base models. Recent examples include Avataar AI, which launched a video production model earlier this week that aims to provide a low-cost alternative to offerings from rivals including Google’s Veo, Kling, Luma, and Runway.

Not everyone agrees that the biggest challenge is lack of money. In response to Pai’s comments, Lightspeed partner Hemant Mohapatra said the biggest obstacles to building globally competitive AI companies are talent, access to computing resources, and execution, rather than just the size of investment commitments.

Mohapatra estimated that training a frontier AI model could cost anywhere from hundreds of millions to billions of dollars, depending on how it’s done, but said successful AI companies have historically scaled their capital needs over time as adoption increased.

Yet for some policy watchers, the implications go beyond AI startups or model providers.

Prasanto Roy, a New Delhi-based technology policy expert who advises multinational companies, said the episode could reinforce the Indian government’s concerns about strategic independence, which he compared to the lesson many countries have learned from Russia’s loss of access to SWIFT and other parts of the global financial system following its attacks in Ukraine.

He told TechCrunch that the move is likely to spark a major nationalist backlash in India and described it as a poorly thought-out decision by Washington, with ramifications that go far beyond Anthropic itself.

“Even if this is fixed or reversed, the Anthropic episode shows that there is no such thing as a geopolitically neutral external LLM,” Roy said. “America’s AI models are tied to America’s geopolitics.”