Ashwini Vaishnaw holds talks with Nvidia on sovereign GPU development in India

Union Minister for Electronics and Information Technology Ashwini Vaishnaw met senior executives from Nvidia on Thursday to discuss India’s plans to develop a sovereign graphics processing unit (GPU) and expand domestic manufacturing of advanced data-processing hardware.

Sharing details of the meeting on social media platform X, Vaishnaw said the discussions covered the development of indigenous GPU capabilities as well as the production of high-performance edge devices in India. The minister also posted a photograph from the meeting with Nvidia’s managing director for South Asia, Vishal Dhupar.

According to Vaishnaw, the talks included Nvidia’s DGX Spark, a compact AI computing system designed to deliver up to one petaFLOP of performance. He noted that the device supports secure inferencing for AI models with up to 200 billion parameters and can operate without internet connectivity, making it suitable for applications in sectors such as railways, shipping, healthcare, education and other remote or critical environments.

Also read: PM Modi holds closed-door AI review with IndiaAI startups ahead of February summit

Nvidia currently dominates the global GPU market, controlling more than 80 percent of the segment. Its chips have become central to artificial intelligence development worldwide, driving strong demand from governments and technology companies building large-scale AI models.

The meeting follows Nvidia’s recent announcements at the Consumer Electronics Show (CES), where the company showcased its DGX Spark and DGX Station deskside AI supercomputers. These systems are designed to allow developers to run advanced AI models locally, ranging from 100-billion-parameter models on DGX Spark to trillion-parameter models on DGX Station.

Vaishnaw has earlier stated that India aims to build its own GPU within the next three to four years, as part of a broader push to reduce reliance on imported AI hardware and strengthen domestic capabilities.

Also read: Indian AI startups should aim for global leadership, says PM Modi

Under the IndiaAI Mission, the government has been expanding access to high-end computing infrastructure for researchers and startups. While the initial plan targeted the deployment of 10,000 GPUs, the programme has since scaled up to 38,000 GPUs, which are being offered to developers at a subsidised rate of ₹65 per hour.

The government has also shortlisted twelve startups and research groups to develop homegrown AI models and engines. These include Sarvam AI, Soket AI, Gnani AI, Gan AI, Avaatar AI, the IIT Bombay-led BharatGen consortium, Zenteiq, Gen Loop, Intellihealth, Shodh AI, Fractal Analytics and Tech Mahindra Maker’s Lab.