Key Takeaways
- India’s sovereign AI model will launch by February 2026 AI Summit
- Government has deployed 38,000 GPUs, far exceeding 10,000 target
- First Indian foundational model expected by end of 2025
- 12 domestic firms being supported for AI development
India will launch its sovereign artificial intelligence model by February 2026, a top IT Ministry official announced. The announcement came during the India Mobile Congress 2025 where significant progress in the country’s AI infrastructure was revealed.
Rapid Infrastructure Scaling
Electronics and IT Secretary S. Krishnan stated that India has rapidly scaled its computing infrastructure despite being a latecomer to AI. The country has deployed 38,000 graphics processing units (GPUs), far exceeding the initial target of 10,000 units.
“Before the end of this year, hopefully, we should have our first entirely Indian foundational model. By the time of the India AI Summit, we should be able to launch India’s sovereign model as well,” Mr. Krishnan confirmed.
India AI Impact Summit 2026
The government has scheduled the two-day India AI Impact Summit for February 19-20, 2026. Officials emphasized their focus on developing inclusive AI platforms that address specific sector needs.
“The other important element is smaller, sector-specific models that enhance productivity and deliver to people who most need it in sectors where AI must make a difference,” Krishnan explained.
Continuous GPU Expansion
The government is continuously enhancing computing infrastructure through quarterly bidding processes for additional GPUs. “We continue to call for bids from time to time and practically every quarter we have bids,” Krishnan noted, adding that a bid is currently active.
Indigenous GPU Development
Developing an indigenous GPU has been identified as a crucial goal under the India Semiconductor Mission 2.0, which is currently under discussion.
Data Sovereignty and Domestic Support
Ministry of Electronics Additional Secretary Abhishek Singh confirmed that India’s sovereign AI model will be completely trained on Indian datasets and hosted on Indian servers. The government is supporting 12 domestic firms, with two expected to have foundational models ready by end-2025.
