By Amit Kapoor and Sheen Zutshi
The world cannot base its AI strategy on the assumption that the world’s most advanced models will remain globally available, commercially accessible, and continuously usable. This is the signal sent by the US government’s restrictions on Anthropic’s Fable 5 and Mythos 5, which makes clear that any country can be denied access to frontier intelligence at any given moment.
The US uses export controls to slow the diffusion of strategic technologies such as encryption source code, advanced chips, chip design software, and semiconductor equipment. The recent restriction on Anthropic’s models offers other countries an early glimpse of a potential future in which the next layer, controlled by export directives, could become frontier intelligence itself. Therefore, the lesson from this episode is essentially that India and other countries cannot rely on partnerships with Foundational AI Technology companies alone, not because they are unreliable, but because they are not sovereign players.
India had recently requested access to Mythos from Anthropic, along with other countries under Project Glasswing, to understand the capabilities of frontier AI models for cybersecurity across its banking, telecom and other sectors. Due to US export controls, that access has been disrupted for India and the rest of the world. US export control directives are not unpredictable moves to begin with. If anything, other countries should have seen this coming. Historically, technology restrictions have shaped global information flows.
Anthropic operates inside the American state, so rightfully, if the White House decides that a frontier model must be restricted, even the largest AI lab will have to comply. Then it will not matter whether India is Anthropic’s second-largest consumer base, because being a customer is not the same as having control; foreign customers will remain a downstream priority in American political decisions, which is completely fair, especially in the technology market. So what are the lessons for India from this Mythos-Fable cut-off moment?
India’s own history shows that it has faced such denial before. But that denial led to pivoting its own pathway, not changing destination. The Nuclear Programme was built over decades post-independence, and its strategic capabilities developed in a world where external access could not be assumed. However, it cannot be compared with Artificial intelligence. Nuclear capabilities were anchored in physical capabilities and reached strategic thresholds, whereas AI frontier intelligence development is constantly evolving. No one knows what the future can actually look like, as LLM models are just the first visible layer of frontier Intelligence. But governing AI is like nuclear technology: AI is politically sensitive, strategically consequential, and difficult to govern through certainty alone. With AI models improving rapidly, computing chips evolving, and inference costs and applications changing across the economy, frontier intelligence will proliferate with or without India’s quest towards AI sovereignty.
Some experts would argue that India can respond to this by diagnosing it as a competitiveness problem. The issue is not only that it has a compute dependency and hasn’t developed foundational models, labs, or research ecosystems, though all of those matter. The deeper problem is that India lacks an AI continuity doctrine to preserve its agency and continuity in the world, as frontier intelligence itself is becoming a more gated and controlled strategic layer.
That’s why India cannot answer this challenge through “checklists for AI sovereignty”, i.e., buying GPUs, building data centres, funding a few startups, and announcing that some foundational models are being built. It should not expect either legacy IT services sector to build those frontier intelligence labs, because they require research depth, not thin R&D spending; a willingness to fail; and building for India and the world. These are aspects on which the Indian IT services model has never operated, but it has built its legacy on labour arbitrage.
Even though the IndiaAI mission’s compute capacity and foundational model push are important beginnings. Recent news on Sarvam’s progress in sovereign AI, compute models, and deployment is encouraging, but the job is far from done. A country of India’s scale needs many competing labs, not one or two symbolic winners. India cannot stop at one promising lab; it needs to create and sustain the conditions to develop many more that push towards the frontier. Anthropic itself emerged from OpenAI. India does not need one Sarvam; it needs conditions for ten Sarvams. We need to start building institutions and conditions that can support an AI continuity doctrine now.
One answer lies in Singapore’s journey into Biomedical science. The entrepot economy was transformed into a global biotech over the decades. It did for BioTech what India needs to do for frontier AI: build institutions before the ecosystem is fully ready. Through A*STAR and Biopolis, Singapore not only recruited global talent, built its research autonomy, and gave institutional backing to Jackie Ying, from MIT, to lead its Institute of Bioengineering and Nanotechnology, but also saw its success compound over the years.
The US may ease controls on frontier models in future to protect the interests of its AI ecosystem again, but can we do so, given that we cannot clearly stop relying on current frontier intelligence, while those institutions cannot wait in perpetuity to be developed if India doesn’t want to be next billed for importing frontier intelligence.
(Amit Kapoor is chair &Sheen Zutshi is research manager at Institute for Competitiveness. X: @kautiliya).
The article has been published with Economic Times on June 22, 2026.
