US Edition
Published • loading... • Updated
India Chases 'DeepSeek Moment' with Homegrown AI
Indian startups showcased AI models supporting 22 languages with government backing, aiming for sovereign AI amid global competition, experts caution against near-term breakthroughs like China’s DeepSeek.
- At the five-day AI Impact Summit in New Delhi, Indian startup Sarvam AI and the government-supported BharatGen initiative released homegrown multilingual models, with Sarvam AI claiming optimisation across 22 Indian languages.
- Policy backers argue India needs sovereign AI to reduce reliance on US and Chinese platforms and meet local data rules, driven by domestic demand, as of February 20, 2026.
- Safety experts at the summit observed that Gnani.ai introduced Vachana speech models trained on more than a million hours of audio, with Nanubala Gnana Sai of the Cambridge AI Safety Institute warning that existing models have Western biases.
- Prime Minister Narendra Modi posed with global tech leaders and highlighted Indian models, while Abu Dhabi-based G42 said the UAE will deploy an AI supercomputer in India on Friday.
- Observers said India-specific models could still influence services over the coming year, but analysts said India is unlikely to deliver a `DeepSeek moment` soon, according to Prihesh Ratnayake, head of AI initiatives at think-tank Factum.
Insights by Ground AI
40 Articles
40 Articles
Coverage Details
Total News Sources40
Leaning Left3Leaning Right5Center14Last UpdatedBias Distribution63% Center
Bias Distribution
- 63% of the sources are Center
63% Center
14%
C 63%
R 23%
Factuality
To view factuality data please Upgrade to Premium
