India’s first homegron semiconductor chip to launch by the end of 2025: Ashwini Vaishnav
Union Electronics and IT Minister, Ashwini Vaishnav, has confirmed that India’s first indigenous semiconductor chip will be ready by the end of 2025.
Listen to the story
Union Electronics and IT Minister, Ashwini Vaishnav, has confirmed that India’s first indigenous semiconductor chip will be ready by the end of 2025. Speaking for money control, the minister said that 5 units required for chip production are running and the project remains on time. In collaboration with Tata Electronics and PSMC in Taiwan, a semiconductor will be developed under the government’s ambitious Semicon India program, and will be built at the upcoming Dholera Fabrication Plant in Gujarat.
This announcement is a milestone in a trip to India, which is towards the chip self -sufficiency, for decades the areas with global veterans. The one -time Dholera Fab will serve as the foundation stone of India’s semiconductor manufacturing ecosystem – an initiative that was a greenlight in December 2021 with a budget outlay of 76,000 crore. Vaishnav said that the chip is expected to be unveiled by September or October 2025.
Apart from hardware success, Vaishnav also unveiled India’s widespread push in Artificial Intelligence (AI). He revealed that India is not behind in the AI race and it will be from the top few countries to exploit AI powers. Recently, he revealed that India is actively working on its big language model (LLM), which would especially complete the diverse linguistic and cultural landscape of the country. The AI model is expected to be ready within the next 10 months.
“In the last 1.5 years, our teams are working closely with startups, researchers, professors, etc., today, we are calling for proposals to develop our own fundamental models. The model will take care of Indian references, languages and culture, which is devoid of prejudice,” Vaishnav said during the conclusion of Utcar.
The project is being supported by an advanced computing setup, with more than 18,000 high-performance GPUs available in India. These include NVIDIA H100 and H200 Chips, as well as Mi325 units, which keeps India on solid ranks compared to global AI leaders. For reference, China’s Dipsek AI was trained on 2,500 GPUs, while Chatgate required around 25,000.
To make this computing power more accessible, the government has created a general feature that developers, researchers and startups can use without heavy upfront costs. About 10,000 GPUs are already operational in this feature, which has been more integrated.
The government has allocated 334 crores to IISC Bengaluru to strengthen research in Gallium Nitride (GAN) technology – a significant component in modern telecommunications and power electronics.
India’s homegron AI model is being developed by six teams and can come between 4 to 10 months from now. With the rapid growth in the place and infrastructure, Vaishnav is confident that India will be one of the top nations globally to successfully exploit the AI’s transformative ability.
