The US just forced Anthropic to cut off access to Indian users. That’s why India needs its own lawfare

The dispute between Anthropic and the US government may yet be resolved. But the vulnerability exposed by the episode will return, in other technologies, through other legal instruments and under other flags, because the underlying logic now favours it

By Aditya Kashyap and Kartikey Singh

The vulnerability exposed by Anthropic’s forced suspension of advanced AI models was not primarily technological. It was jurisdictional. India’s response must be legal as much as technological. An American export-control order switched off a service for users around the world in a matter of hours. The instrument was legal. On June 12, the United States government directed Anthropic to suspend access to two of its most advanced AI models (Fable 5 and Mythos 5) for every foreign national, whether inside or outside American territory. The order even covered Anthropic’s own foreign-national employees. Subsequently, Anthropic complied, taking the models offline for affected users, including customers in India. While access to its other models was left untouched, the company has publicly disputed the basis of the directive.

The mechanism matters here. The tool used here was neither a tariff nor a sanctions list. It was a domestic export-control directive: A piece of American administrative law, operating within America’s own statutory framework, that nevertheless reached across borders and disabled a commercial service almost overnight. That is the part India should study closely. The episode demonstrates that globalisation now runs not only on infrastructure and markets, but on law. Thus, India needs a lawfare doctrine, not merely a sovereign-AI strategy.

The reaction in India was led, predictably, by Sridhar Vembu. The Zoho founder described technology as the “ultimate weapon”, arguing that national sovereignty and security are now inseparable from technological capability, and suggested that globalisation is effectively over. He urged Indian institutions to rely more heavily on domestic and open-source models rather than foreign frontier systems.

That argument is persuasive as far as it goes. But it captures only half the exposure. The other half is jurisdictional. And that is a problem for lawyers and policymakers before it is for engineers. What happened on June 12 is best understood as lawfare: The use of legal instruments to achieve strategic ends that once required harder tools. Globalisation was often presented as a political infrastructure, with capital, goods and code flowing beyond the reach of any one sovereign. That story was always incomplete.

The chokepoints of the digital economy — advanced semiconductors, frontier AI models, cloud infrastructure, payment rails and technical standards — sit physically and legally within a small number of jurisdictions. Whoever controls the jurisdiction controls the off-switch.

The Anthropic directive is simply a clear illustration of that structural reality. For India, the policy conclusion is not merely that we should build our own models, though we should. It is that technological self-reliance without a corresponding legal doctrine is an incomplete shield.

A serious Indian lawfare doctrine would contain four elements. The first is defensive insulation. The European Union’s Blocking Statute provides a useful example. It restricts compliance with certain specified foreign extraterritorial measures and limits the legal effect of related foreign judgments within the Union. China has developed its own counter-sanctions and blocking framework. Notably, India has the Digital Personal Data Protection Act and the IT Rules, but no equivalent instrument addressing the question raised by the Anthropic episode: When a foreign government directs a company to alter conduct affecting Indian users, what status does that directive possess within Indian law?

At present, the answer is unclear. That uncertainty is itself a vulnerability. India does not need a sweeping or confrontational blocking regime. But it does need a narrowly tailored framework establishing that foreign legal commands affecting Indian users, Indian infrastructure or Indian commercial rights do not automatically acquire force within India. They must first be tested against Indian law.

Further, the second requirement is redundancy. India does not need to win a hundred-billion-dollar frontier-model race to reduce its exposure to abrupt external cut-offs. It needs to eliminate single points of failure. The country has already demonstrated through UPI, Aadhaar and ONDC that digital public infrastructure can be built at national scale. The same logic should now extend to compute capacity, open-weight models, cloud diversity and procurement strategy. Vembu’s preference for smaller models fits naturally within this approach, not as the entire solution but as one component of it.

The third element is leverage. Lawfare is not only defensive. India’s domestic market is itself a source of jurisdictional power. Any platform operating here accepts Indian law as the price of entry. Reciprocity should therefore become part of regulatory thinking. Market access cannot mean that Indian users remain exposed to abrupt service withdrawal because another government changes its administrative position. A provider capable of being ordered to disconnect Indian customers without notice is, in commercial terms, an unreliable counterparty. Our procurement frameworks and sectoral regulations should recognise and price that risk rather than ignore it.

The fourth element is the forum dimension. Trade law, investment treaties, the WTO and emerging technology governance arrangements are not side stages. They are arenas in which legal argument frequently substitutes for harder forms of pressure. India already understands this logic in other domains. The challenge is to bring the same seriousness to technology, data and AI governance, where our posture has too often been reactive rather than strategic.

None of this requires the language of decoupling. Indeed, the argument is stronger without it. The point is not that India should isolate itself from global technology markets. The point is narrower and more defensible: No state should allow essential services to depend entirely on the continuing discretion of another country’s administrative machinery.

The dispute between Anthropic and the United States government may yet be resolved. But the vulnerability exposed by the episode will return, in other technologies, through other legal instruments and under other flags, because the underlying logic now favours it.

Globalisation promised a world in which borders mattered less. What this moment reveals is that borders did not disappear. They migrated into law, code, and infrastructure, where they are harder to see and easier to weaponise. India’s challenge is to stop being merely the object of other states’ lawfare and to begin acting as the author of its own.

Kashyap is a practising advocate at the Supreme Court of India and central government counsel, and Singh is a lawyer based in New Delhi