Anthropic’s decision to disable access to its Fable 5 and Mythos 5 AI models for foreign nationals following a US export control directive has reignited concerns over India’s dependence on foreign AI systems and sharpened calls for greater technological self-reliance.
The move, announced on Saturday, has raised questions about the risks of relying on AI capabilities controlled by overseas companies and governments at a time when businesses and public institutions are increasingly integrating advanced AI tools into critical operations. Earlier this month, a select group of Indian entities and government agencies had received early access to Anthropic’s Mythos model, which was being evaluated for cybersecurity and software security applications.
Immediate Defence Vulnerabilities
Industry executives and policy experts said the development underscores a broader challenge facing India as AI becomes an increasingly important layer of digital infrastructure. “The sudden, unilateral withdrawal of Fable 5 and Mythos 5 by the US government creates an immediate and dangerous asymmetry of defence,” Neehar Pathare, managing director and CEO of 63SATS Cybertech, told Fe. The decision highlights the risks of depending on external providers for capabilities that may become essential for securing financial systems, digital infrastructure and critical services, he added.
The episode has also revived the debate around sovereign AI, or a country’s ability to develop, deploy and govern critical AI technologies without excessive dependence on foreign providers. “If the most significant tech differentiator you are leveraging has external control loops, then you have to accept you are vulnerable,” Pratyush Kumar, co-founder and CEO of Sarvam AI, said in a post on X, adding that countries must increasingly be able to shape their own technological destinies.
Experts said that sovereign AI should not be viewed as technological isolation but as the ability to retain strategic autonomy over critical technologies. “It is imperative that India works towards a long-term goal of sovereign AI,” Kazim Rizvi, founder of technology policy think tank The Dialogue, said. According to Rizvi, investments in indigenous compute capacity, frontier research and a stronger domestic innovation ecosystem are essential to ensure that India can make independent choices regarding the development and deployment of AI systems.
Compute and Capital Bottleneck
The government has in recent months stepped up efforts to build domestic AI infrastructure under the IndiaAI Mission, including the creation of a national compute ecosystem and support for indigenous model development. Industry experts said such initiatives are necessary first steps but added that significant gaps remain in funding and computing resources.
“In recent months, the government has been laying special focus on developing indigenous AI infrastructure that can help build sovereign AI centres. This is the first step towards preventing over reliance on AI systems built outside the country. Naturally, this includes securing data infrastructure and building domestic semiconductor and cloud ecosystems. Regular incentives to encourage adoption of trusted Indian AI products will also encourage banking and financial systems to work with them closely. The key is to ensure that the domestic AI systems meet the compliance standards,” Ashish Biji, Partner for Cybersecurity, Business Advisory, BDO India stated.
“Building a Fable-class model is a compute and funding game,” Paras Chopra, founder of AI research lab Lossfunk, said in a LinkedIn post. While leading global AI labs are estimated to have access to hundreds of thousands, and in some cases over a million advanced graphics processing units (GPU), India’s compute base remains considerably smaller. Experts said that bridging this gap would require investments running into tens of billions of dollars and sustained policy support over several years.