The Modi-Trump summit reinforced three pillars of the US-India relations — defense, trade, and energy. However, technology, too, was a central theme, with both leaders emphasising its role in shaping the future of their partnership.
The US and India have now formalised this through COMPACT (Catalysing Opportunities for Military Partnership, Accelerated Commerce & Technology), a broad framework that cements their deepening tech, industrial, and security ties.
Story continues below this ad
The summit delivered headline agreements — the Transforming Relationship Utilising Strategic Technology (TRUST) Initiative, INDUS Innovation, and the Autonomous Systems Industry Alliance (ASIA). These signal unprecedented cooperation in AI, semiconductors, and defense technology.
But one critical issue remains unresolved — access to the most advanced AI chips.
Artificial intelligence is set to accelerate in both Washington and New Delhi. The recent AI Action Summit in Paris underscored this reality. While US Vice President JD Vance and Prime Minister Narendra Modi differed on AI governance, both rejected fear-based regulation. AI, they agreed, must be embraced at full speed.
Story continues below this ad
Yet, the way the US and India pursue AI acceleration is diverging.
DeepSeek, a Chinese AI company, has alarmed US lawmakers by delivering cutting-edge models on limited computing, prompting Washington to tighten restrictions on AI chips and models. India, in contrast, sees an opportunity, announcing plans to build sovereign large language models (LLMs) and scale AI investments.
The question is no longer whether AI will define US-India relations, but whether India will be allowed the tools to compete. The joint statement explicitly mentions AI compute access as part of the US-India technology agenda.
But this does not mean unrestricted access to high-end GPUs governed by AI diffusion rules, which are necessary for training large foundational models. Under the current US framework, India falls under Group 2, meaning its AI chip access is capped and subject to licensing approvals. Meanwhile, China, despite being prohibited as per Group 3 restrictions, has exploited loopholes to stockpile embargoed AI chips through intermediaries.
This exposes a flaw: Compliance is punished, while circumvention thrives.
The Biden administration had already imposed diffusion restrictions to curb China’s AI capabilities. The Trump administration is expected to go further, potentially expanding limits not only on chips but also AI models and proprietary datasets.
India, which has played by the rules, now risks being locked out at a critical moment in its AI trajectory. If Washington wants India as a long-term AI partner, a serious conversation about unrestricted AI compute access is overdue. At the summit, India committed to long-term US energy deals, strengthened trade ties, and reinforced its strategic alignment with the West.
Its stance on global security, including Ukraine ceasefire discussions, further underscores its role as a responsible international actor. Unlike China, which hoards restricted technology and circumvents export controls, India has signalled a willingness to engage in structured AI partnerships.
Restricting AI compute for a trusted democratic partner — one that Washington itself positions as a counterweight to China — raises questions about their strategy. A stronger Indian AI ecosystem will expand US-India technology trade, deepen AI semiconductor supply chains, and boost AI-driven research collaborations.
The INDUS Innovation framework provides a natural platform to formalise AI compute access, ensuring India builds its own AI ecosystem instead of relying on third-party solutions. The ongoing diversification of semiconductor supply chains has already positioned India as a key player.
Ensuring India has the compute power to develop its AI sector would further strengthen its role as a technology hub. AI compute access must be fully integrated into the existing strategic framework.
most read
A strategic review of India’s classification under the US AI export controls could differentiate India from China in the upcoming diffusion rules, allowing for greater flexibility in GPU exports. Future agreements under TRUST and INDUS could incorporate scalable AI compute provisions, ensuring India has the resources to move from policy discussions to real-world AI deployment.
As the US-India trade partnership expands, there is an opportunity to align AI compute access with broader economic cooperation. The Modi-Trump summit laid the foundation for a deeper US-India technology alliance. But its effectiveness will depend on whether India gains the tools to build AI at scale.
The writer is the founder of think tank Policy 4.0. which has worked on AI and crypto policy with the G20 and several global regulators. She has worked previously on Capitol Hill and with the MEA.
