The government’s ambitious plan to build sovereign AI models is running into a familiar hurdle: a shortage of private capital. While IndiaAI Mission-backed startup Sarvam recently raised $234 million at a $1.5 billion valuation, most other startups selected under the programme have attracted only a fraction of that amount.
Launched in 2024 with an outlay of Rs 10,000 crore, the IndiaAI Mission aims to support the development of indigenous foundational models through a combination of grants, subsidised compute and infrastructure support. The programme has shortlisted 12 teams, including Sarvam AI, Soket AI, Gnani AI, Gan AI, Avataar AI, BharatGen, GenLoop, Zentieq, Intellihealth, Shodh AI, Fractal Analytics and Tech Mahindra Maker’s Lab. The goal is to create homegrown AI models that can reduce dependence on foreign platforms and strengthen India’s technological sovereignty. Yet, despite growing policy support, private investors have largely remained selective.
Capital Disparity
The contrast is most visible in funding. Sarvam has raised nearly $300 million across funding rounds, making it India’s best-funded foundational AI startup. Avataar AI, another participant in the programme, has raised about $55 million. Enterprise voice AI startup Gnani.ai has secured less than $20 million, while Soket AI and Gan.ai have raised only a few million dollars each through seed and early-stage rounds. BharatGen, meanwhile, is a government-backed academic consortium and relies primarily on public funding. The disparity suggests that while the government has backed a broad set of contenders, venture capital has so far concentrated behind a handful of companies.
The funding gap mirrors broader trends in the market. According to CB Insights, global private equity and venture capital investment in AI touched $225.8 billion in 2025, with the United States accounting for nearly $189 billion. Indian AI startups raised about $1.34 billion during the period, representing a small share of global funding. Much of that capital flowed into application-layer companies rather than firms attempting to build large foundational models from scratch.
Revenue Timeline Lag
“The IndiaAI Mission laid important groundwork, but public compute infrastructure and research grants alone don’t immediately create fundable companies. There’s a pipeline lag,” Abhishek Srivastava, general partner at Kae Capital, said. Research-led AI startups take longer to demonstrate commercial traction, while most investors continue to favour businesses that can show revenue visibility within a shorter time frame, he said.
Industry analysts say the preference for application-layer companies is driven by lower capital requirements and clearer paths to monetisation. By leveraging existing foundation models, such startups can focus on specific enterprise and consumer use cases without incurring the enormous costs associated with training large models. This has helped companies such as Uniphore, Innovaccer, CommerceIQ and Eightfold attract significantly more investor interest than deep-tech AI ventures.
“These companies offer clearer commercialization pathways, lower capital requirements, and faster revenue generation. By leveraging existing foundation models, application-layer startups can focus on solving specific enterprise and consumer use cases, making them more attractive to investors seeking quicker returns and lower execution risk,” Neha Singh, co-founder of Tracxn stated.
Anushree Verma, Senior Director Analyst at Gartner shared that from a funding point of view, frontier models will always attract more funding than just the sovereign players. “Overall, the limited funding in homegrown AI startups does not necessarily indicate weak investor interest in AI sovereignty. Rather, it reflects the relatively nascent nature of India’s foundational AI ecosystem, where only a few companies have reached the scale, technical maturity, and strategic positioning required to attract substantial capital,” Singh added.
That dynamic, however, needs to change as AI sovereignty becomes a strategic priority for governments and large enterprises, analysts said.