The idea of “sovereign AI” has often been dismissed as techno-nationalist rhetoric, an idea which is considered expensive, duplicative and largely unnecessary in a world where the best models are already available.
The recent developments around Anthropic may have changed that calculation. The sudden restrictions on access to some of its frontier models are a reminder that artificial intelligence infrastructure is subject to the laws of the countries where it is built.
On June 12, Anthropic said it would “abruptly disable” its most advanced AI models, including Fable 5 and Mythos 5, for all users after the US government ordered it to suspend access for foreign nationals, citing national security concerns.
For India, this has become one of the strongest arguments yet for investing in domestic capabilities. Under the Rs 10,371.92-crore IndiaAI Mission, the government has shortlisted 12 organisations and consortia to build foundational models tailored to Indian needs.
The projects span large language models, small language models, voice systems and multimodal AI capable of handling text, images and video.
IIT Bombay-led BharatGen consortium
At the centre of the effort is the IIT Bombay-led BharatGen consortium.
With government support of Rs 1,058.52 crore, BharatGen has emerged as the programme’s flagship project, accounting for nearly 60 percent of the total approved funding under the foundational models pillar.
The consortium is building open-weight multilingual and multimodal models intended for public-sector applications, including translation, document understanding and citizen services.
IIT-B-led consortium, includes IIT-Madras, International Institute of Information Technology-Hyderabad, IIT-Kanpur, IIT-Hyderabad, IIT-Mandi, IIM Indore, Indraprastha Institute of Information Technology, Delhi and IIT-Kharagpur. In November 2025, IIT- Bombay registered BharatGen with the Registrar of Companies (ROC) as a non-profit organisation.
The company unveiled one of its large language models at the IndiaAI Impact Summit in February.
Sarvam AI
Among startups, Sarvam AI has received the largest allocation. Backed by Rs 246.72 crore in compute support, it is focused on Indic-language models that can navigate the complexities of Indian speech patterns, dialects and code-switching.
Sarvam was also among the first companies to publicly unveil its sovereign model during the IndiaAI Impact Summit.
Talking about the Anthropic controversy, Sarvam co-founder and CEO Pratyush Kumar said the episode highlighted why nations and companies need greater control over the AI systems they depend on.
“Fable ban is a good instigation for more people to engage in recognising the need for sovereignty,” Kumar wrote on X. “But while many opine, Sarvam is at work making the Fable of India’s first consequential AI company a reality.”
GnaniAI (voice), GanAI and Avataar AI (video)
Gnani AI, allocated Rs 177.27 crore, is taking a different approach. Rather than text-first systems, it is building voice-native models aimed at users who interact with technology primarily through speech.
The government is also betting on multimodal systems.
Gan AI, backed by Rs 110.03 crore, is developing video-generation capabilities for enterprise applications such as education and personalised communication.
Avataar AI, which received Rs 16.10 crore, is working on visual AI systems and recently introduced Varya, a distilled video-generation model designed to reduce the cost of producing high-quality content.
SoketAI, Zenteiq and Genloop
Several projects are focused less on scale and more on efficiency.
Soket AI, allocated Rs 177.08 crore, is building small language models that can run on local servers and edge devices without relying heavily on cloud infrastructure.
Zenteiq, backed by Rs 206.49 crore, is researching optimisation techniques to improve model performance while reducing computational demands. Genloop, which received Rs 2.61 crore, is exploring methods to create lightweight models through iterative fine-tuning and model adaptation.
Fractal and Tech Mahindra
Others are targeting specific sectors.
Fractal Analytics, with support of Rs 137.91 crore, is developing enterprise-grade systems for analytics and reasoning. Tech Mahindra’s Maker’s Lab is focused on domain-specific models for sectors such as telecom and logistics.
Intellihealth and ShodhAI
Intellihealth, backed by Rs 49.50 crore, is building healthcare AI trained on medical datasets to support diagnostics and decision-making.
Shodh AI, allocated Rs 9.40 crore, is developing benchmarks and evaluation frameworks to test indigenous models for safety, bias and performance.
Taken together, the 12 projects look less like an industrial policy experiment and more like an insurance policy against a future where access to intelligence is no longer guaranteed.
Source: moneycontrol.com