Anthropic curbs ignite AI debate in India

India is rethinking its artificial intelligence strategy after the U.S. government issued export-control directives that led Anthropic to disable access to its Fable 5 and Mythos 5 models for foreign nationals. This shift highlights India’s reliance on foreign foundational models and a growing push for “sovereign AI” to ensure technological independence, according to industry reports.

Why is India rethinking its AI strategy?

India previously aimed to become an AI innovation hub by building applications on top of foreign foundational models. This strategy faced a setback last week when Anthropic suspended access to its newest models, Fable 5 and Mythos 5, for foreign nationals to comply with U.S. export controls.

Saket Dandotia, co-founder and CEO of Onetab.ai, told CNBC that the ability for frontier access to “vanish overnight on a foreign government’s order is the whole problem.” Dandotia noted that while diversifying across models buys time, it does not provide independence.

An ADP Research report released Thursday found that 41% of Indian workers use AI nearly every day. This is higher than 26% in China and 19% in the U.S., though experts say this adoption reflects a heavy reliance on foreign technology.

Did You Know? According to an ADP Research report, 41% of Indian workers use AI almost daily, significantly outpacing adoption rates in both China (26%) and the U.S. (19%).

What is the “Sovereign AI gap” in India?

India currently lacks domestic production of cutting-edge chips and does not possess a frontier-scale foundation model comparable to those in the U.S. or China. Data center capacity also lags behind these two nations, despite rapid growth.

The Indian government has launched a semiconductor mission, an AI mission, and offered tax breaks to global hyperscalers to build data centers. In the private sector, Sarvam AI raised $300 million at a $1.5 billion valuation on Monday, with investment from HCL Technologies.

However, Manish Agarwal, co-founder of Humyn Labs, told CNBC that India lacks the deep-tech investment capital available in the U.S. and China. He warned that without a strong government drive, the pursuit of sovereign AI risks “ephemerality.”

Expert Insight: Samantha Carter notes that India’s current position creates a strategic vulnerability. By building the “application layer” on foreign foundations, India’s tech ecosystem remains subject to the regulatory whims of foreign governments, making the transition to a full sovereign stack a matter of national economic security.

Why is deep-tech investment lagging?

Last year, Indian startups raised $10.5 billion, the third-highest total globally after the U.S. and U.K., according to December data from Tracxn. Most of this capital flowed into retail, fintech, and enterprise applications rather than deep-tech firms.

Private investors remain conservative regarding disruptive technologies. For example, HCL Tech’s 14.27 billion rupee ($151 million) investment in Sarvam was less than 10% of the dividends the company paid to shareholders in the financial year ending March 2026.

Venture capitalist Mohandas Pai has urged Prime Minister Narendra Modi to expand the national AI mission. Pai described existing government programs as “too slow, way too small to make any large impact.”

What happens next for Indian AI?

Experts suggest that building a foundational model capable of avoiding hallucinations would require a few trillion parameters, necessitating massive capital and computing power. For comparison, Sarvam’s flagship model has slightly over 100 billion parameters.

US Expands AI Export Controls as Anthropic Halts Foreign Access to Advanced Models on Government …

India’s current sovereign AI models rely on Nvidia architecture. Neil Shah, vice president of research at Counterpoint Research, told CNBC that India could be left “helpless” if the U.S. restricts access to Blackwell chips as it did with China.

Sridhar Vembu, co-founder of Zoho and founder of Arattai, stated on X that “Technology is the ultimate weapon” and that India must find its own path. Future stability may depend on the government’s ability to secure sovereign computing power and increased deep-tech capital.

Frequently Asked Questions

Why did Anthropic disable access to its models in India?
Anthropic disabled access to Fable 5 and Mythos 5 for foreign nationals to comply with an export-control directive from the U.S. government.

How does India’s AI investment compare to its dividend payouts?
In the financial year ending March 2026, HCL Tech’s $151 million investment in Sarvam AI represented less than 10% of what the company paid out to shareholders as dividends.

What is the primary technical risk for India’s AI development?
According to Neil Shah of Counterpoint Research, the primary risk is the potential for the U.S. to restrict access to Nvidia Blackwell chips, which would leave India without the necessary hardware for its sovereign AI models.

Do you believe government-led investment is the only way for India to achieve AI independence?