India Fast‑Tracks Homegrown 2nm GPU to Rival Nvidia, Targets 2030 Launch

India is fast‑tracking the development of a homegrown 2‑nanometre graphics processing unit (GPU), targeting completion by 2030. Spearheaded by experts at the Centre for Development of Advanced Computing (C‑DAC) in Bengaluru, the project has been allocated $ 200 million in funding.

According to a senior C‑DAC official, industry roadmaps from market leader NVIDIA indicate that by 2028 the most advanced GPUs will be manufactured on a 2 nm process node. By 2030 these chips will set the standard for data‑centre and AI‑training workloads, the official explained. India’s new GPU aims to match that performance while cutting production costs by half.

Currently Apple’s A‑series processors used in the iPhone are fabricated on a 3 nm process, making them the most advanced commercial chips in use today. After completing the design phase C‑DAC plans to partner with Taiwan Semiconductor Manufacturing Co (TSMC) to handle mass production of the 2 nm GPUs at roughly 50 percent lower cost than comparable offerings.

GPUs have become indispensable for artificial intelligence, powering the machine‑learning operations in data centres worldwide. Since the launch of ChatGPT in late 2022 demand for high‑performance GPUs has soared, driving NVIDIA’s valuation to increase tenfold and elevating it to the position of the world’s second‑most valuable company.

Despite its considerable chip‑design talent, India currently holds no native GPU patents and depends on US suppliers such as Intel, AMD, Qualcomm and NVIDIA for critical AI hardware. An executive order issued by then US President Joe Biden last year underscored this vulnerability by warning that in a conflict scenario exports of advanced chips to countries such as India could be restricted.

“That executive order was a wake‑up call,” the C‑DAC official said. “Since then we’ve been committed to building an indigenous GPU. By 2030 we’ll deploy these chips in our cloud servers and supercomputers, giving academia, researchers and startups the tools they need to develop sovereign AI models and cloud platforms.”