India–US Plan To Ramp Up GPU And Data Centre Infrastructure Trade

India’s interim trade deal with the United States is expected to improve access to advanced technology hardware such as graphics processing units (GPUs) and data centre equipment.

According to the India–US joint statement outlining the framework for the agreement, both countries aim to increase trade in high-end technology products and expand cooperation in critical digital and data infrastructure.

Tech Trade Expansion To Support India’s AI Push

Union Minister for Electronics and Information Technology Ashwini Vaishnaw said data centres will be one of the biggest growth drivers in the coming years. He said India has already seen committed investments of around $90 billion in the data centre ecosystem, and this figure is expected to exceed $200 billion in the near future.

Vaishnaw observed that the growth of the computing infrastructure had the potential to introduce new capabilities along the artificial intelligence (AI) value chain and assist startups in developing and expanding AI-led solutions based in India to international markets.

Vaishnaw also emphasised the fact that India is currently running approximately 38,000 GPUs on its common compute infrastructure.

He further claimed India will then increase its total GPUs by 20,000 more, yet a bigger increase will be added as a part of the AI Mission by the government, which is the wider effort that the country is making to establish domestic computing capability at scale.

High Import Duties Have Been A Key Cost Barrier

Piyush Somani, founder, chairman, and managing director of ESDS Software Solutions, pointed out that enterprise GPU servers currently attract import duties of 20%-28% in India, sharply increasing the cost of AI compute infrastructure.

He noted that an NVIDIA H100 server, which costs around $250,000 in the US, can end up costing nearly $325,000 in India after duties, making GPU-as-a-service offerings roughly 40% more expensive compared to competing global hubs such as Singapore and the UAE.

According to him, this cost disadvantage has been one of the biggest obstacles limiting India’s competitiveness in the AI infrastructure race.

Data Centre Costs May Fall If Duties Are Cut

Somani said if import duties on enterprise GPU servers are reduced toward zero under the trade framework, the economics of large-scale AI infrastructure could improve significantly.

He estimated that the cost of setting up a 10 MW GPU-ready data centre could fall by around 14%, dropping to ₹2,518 crore from ₹2,926 crore.

He added that such cost reductions could unlock large-scale expansion capacity and potentially enable tens of thousands of crores in additional AI infrastructure investment at a national level.

Qualcomm Event Showcases India’s Chip Design Progress

Vaishnaw made remarks at a press conference held at semiconductor major Qualcomm’s campus in Bengaluru, where he highlighted India’s progress in semiconductor design.

He said Qualcomm had unveiled a two-nanometre chip designed in India, pointing to India’s growing role in advanced chip development.

Vaishnaw added that global firms are increasingly designing complete products in India, citing examples such as AMD and Qualcomm.

He noted that modern semiconductor designs involve extreme complexity, with each die containing around 20 to 30 billion transistors, calling this a major milestone for India’s technology ecosystem.

On intellectual property (IP), Vaishnaw clarified that ownership will depend on company-level policies. Nevertheless, he restated the promise of co-development and co-creation of India and adherence to international standards of IP.

What Does This Mean For Investors?

For investors, the interim trade framework signals a possible shift in how India builds its AI infrastructure. If duties are eased and access to GPUs and data centre equipment improves, project costs could come down, and timelines could shorten.

Alongside steady investment activity and wider use of high-end computing, this keeps the long-term case intact for India’s AI and data centre ecosystem. Companies linked to digital infrastructure, cloud services and chip design stand to benefit the most.

Sources:

Economic Times

Business Standard