After witnessing a mega success with Apple’s local manufacturing efforts, Union Minister for Electronics and Information Technology Ashwini Vaishnaw has engaged in high-level discussions with NVIDIA executives to advance India’s ambitions for locally made high-end graphics processing units (GPUs) and domestic production of cutting-edge AI hardware. The talks focused on manufacturing advanced edge AI devices in India, hinting at a major push toward reducing dependency on imported semiconductor technology.
Vaishnaw met with Vishal Dhupar, NVIDIA’s Managing Director for South Asia, and other senior officials to explore collaborative opportunities in GPU development and localised production.
In a post on X (formerly Twitter), the minister wrote: “Discussed development of sovereign GPUs and manufacturing of edge devices like DGX Spark in Bharat. This device delivers up to 1 petaFLOP performance with secure inferencing for models up to 200 billion parameters. This compact GPU doesn’t require the Internet. Suitable for railways, shipping, healthcare, education and remote applications.”
Spotlight on NVIDIA DGX Spark and Edge AI potential
The centerpiece of the conversation was NVIDIA’s recently unveiled DGX Spark, a compact AI supercomputer showcased at CES 2026. Designed for edge computing, the device offers offline, secure AI inference capabilities – ideal for sensitive or remote environments where constant internet connectivity is impractical.
Some of the key advantages highlighted here include:
– Up to 1 petaFLOP of performance.
– Support for large language models with up to 200 billion parameters.
– Fully offline operation, enhancing data security and sovereignty.
– Applications in critical sectors like transportation, healthcare, education, and rural development.
Manufacturing such devices locally would align with India’s Atmanirbhar Bharat initiative and bolster national AI infrastructure.
Broader push for locally-made GPU ecosystem
The discussions build on India’s long-term goal of developing homegrown high-performance GPUs. Vaishnaw has previously stated that the country aims to create sovereign GPU designs within 3-4 years, potentially leveraging open-source architectures or licensed technologies rather than building from the ground up.
This effort is supported by the India AI Mission, which has significantly expanded access to subsidised GPUs, deploying 38,000 units at affordable rates (Rs 65 per hour) and selecting 12 startups to build indigenous AI models.
With global demand for AI chips surging and supply chains facing geopolitical risks, partnerships with leaders like NVIDIA – which dominates over 80% of the GPU market – could accelerate India’s entry into high-end semiconductor manufacturing. No immediate agreements were announced though.
