AI Slowdown Call by Top Tech Titans Creates New Bind for India

A call by Anthropic Chief Executive Officer Dario Amodei and backed by top-tier tech titans like Elon Musk and Sam Altman for a more deliberate approach to the development of frontier artificial intelligence is opening a new debate in Washington — one that could have consequences well beyond the US technology industry.

India specifically is facing a difficult choice over how closely to align its artificial-intelligence ambitions with the US and China as Washington debates tighter controls on frontier AI while Beijing seeks to spread Chinese-developed models and infrastructure across the developing world.

The issue is no longer just how quickly AI should develop. It is also who will supply the models, chips and computing infrastructure on which India's AI economy will depend. This is a crucial geopolitical and technological junction. The country wants deeper technology ties with the US, particularly in semiconductors, cloud computing and advanced AI, while continuing to use BRICS as a platform for the interests of emerging economies. The challenge is to expand those relationships without becoming locked into either technology sphere.

The debate in Washington over the pace of frontier AI has added urgency to that calculation.

Amodei and other technology leaders have argued that the development of increasingly capable AI systems should proceed with stronger safeguards, including independent evaluations and clearer thresholds for high-risk capabilities. The proposals come amid growing evidence that AI agents can operate with greater autonomy and interact with external systems.

The argument has also exposed a divide inside the US over how much weight should be given to AI safety compared with maintaining America's lead over China.

Washington's restrictions on advanced semiconductor exports are already limiting China's access to some of the computing power required to develop cutting-edge AI. If the US and its allies also coordinate more closely on the development and deployment of frontier systems, the result could be a technology divide extending beyond chips to the models and infrastructure that sit on top of them.

China and the BRICS Intention

Beijing is taking a different route.

At the BRICS summit in New Delhi, President Xi Jinping promoted a BRICS AI open-source community and a digital ecosystem cloud platform intended to expand access to AI models, computing resources and technical expertise across developing economies.

The appeal for China is clear. Open-weight models can be distributed more widely than proprietary systems, allowing Chinese companies and institutions to build relationships with developers, governments and businesses in markets where access to Western technology is restricted or expensive.

For Beijing, that could turn AI into another channel for expanding its technological influence across the Global South.

India is an important test of that strategy.

New Delhi wants access to American technology and investment, but it also wants greater control over critical digital infrastructure. Chinese open-weight models could give Indian developers another source of capable AI systems and reduce dependence on a small number of US providers.

The trade-off is strategic. Greater reliance on Chinese technology in a sensitive sector could introduce a different form of dependence.

India has encountered versions of this dilemma in other technology markets: the cheapest or most accessible option is not always the one that leaves the country with the most room to manoeuvre.

Wriggle Room for India

The uncertainty could nevertheless give India's technology industry some time to adjust.

The country's roughly $280 billion technology-services sector is under pressure from AI tools capable of automating software development, testing, research and other work traditionally performed by large engineering teams.

For companies such as TCS, Infosys and Wipro, a slower pace of frontier-model development could make the transition less abrupt. They could use the period to retrain employees, upgrade enterprise data systems and move further away from billing primarily for headcount and hours.

The opportunity is less about building the next frontier model than about putting existing models to work inside banks, manufacturers, retailers and government departments.

India's startup market faces a similar adjustment. Companies built largely around thin layers on top of third-party AI models can be vulnerable whenever a major provider releases a more capable or cheaper model.

A more stable development cycle could give founders time to build products around proprietary data, industry-specific workflows and customer relationships rather than competing on access to the latest general-purpose model.

Compute Capacity Remains Problem

India's AI ambitions remain small compared with the clusters being assembled by the world's largest technology companies. The IndiaAI Mission has expanded access to GPUs, but the country's available capacity is still a fraction of what leading US laboratories can deploy.

That matters beyond the ability to train a domestic foundation model. Researchers and companies need computing power to fine-tune models, run evaluations and develop applications at scale.

For New Delhi, that makes access to compute an industrial-policy issue as much as a technology issue.

One option under discussion is a larger national pool of AI computing capacity, potentially supported by requirements for major data-center operators to make a portion of their capacity available to Indian startups, researchers and public institutions.

Such a system would not make India independent of foreign technology. It would give Indian companies more choice over which models they use and reduce the risk of being dependent on a single cloud or AI provider.

Keeping Both Doors Open

India is unlikely to become independent of US and Chinese technology in the foreseeable future. Nor does it need to.

A more practical objective is to ensure that both ecosystems remain accessible while India builds capabilities of its own.

That means expanding domestic compute, maintaining open technical standards, developing indigenous models where there is a commercial or strategic case and building applications around India's existing strengths in software and enterprise services.

The geopolitical stakes are broader than India's technology industry.

If the US retains an overwhelming lead in frontier models and advanced chips while China succeeds in making open-weight AI widely available across emerging markets, the global AI market could develop along two distinct lines.

India's size makes it an important swing market. Its software companies are major users of Western technology, while its BRICS membership gives Beijing a platform for promoting Chinese alternatives.

The immediate priority is therefore more practical than trying to win the frontier-model race: securing enough computing capacity, maintaining access to competing models and building domestic expertise so that no single supplier becomes indispensable.

The AI contest is increasingly about who controls the infrastructure through which models are trained, deployed and accessed. For India, preserving the ability to choose may prove more important than choosing a side.