Parsa Venkateshwar Rao Jr | Govt-driven AI Might Not Be Right Strategy For India

India under the Narendra Modi government is determined to make a place for itself in the emerging sector of Artificial Intelligence, which is blazing and booming in the American markets, with the prospect of mega Initial Public Offering (IPO) by Starlink-Space X-AI owner Elon Musk, which is set to break the trillion-dollar mark, even as other AI heavyweights like Anthropic, Open AI, Alphabet and Meta are looking to the market to raise the huge funds to boost and expand AI operations through research and data centres.

There was a mini-storm that shook the AI stocks from Seoul to New York, and where Elon Musk’s $600 billion were wiped out on June 23, and Samsung Electronics and SK Hynix shares fell by nine per cent each on the morning of June 26, and trading had to be stopped. But volatility is the name of the beast – or market.

India’s AI gambit is government-driven so far, moving along the track laid down by AI Mission announced in March 2024 with its various bureaucratic-sounding pillars like IndiaAI Compute Capacity, IndiaAI Innovation Centre (IAIC), IndiaAI Datasets Platform, IndiaAI Applications Development Initiative, IndiaAI FutureSkills, IndiaAI Startup Financing, and Safe & Trusted AI. The fund allocated for the mission for a five-year period is Rs 10,300 crores, around $1.24 billion.

The differences in the scale and mode of the AI take-off in the United States, China, South Korea and Taiwan, compared to India, are obvious. In the case of China, South Korea, Taiwan, the backing of the respective governments in creating a conducive atmosphere is a given, but even in these countries, private initiative and enthusiasm marks the AI project. In India, the markets are quite silent, even somnolent, over AI. It is as though AI is a non-marketable technology/commodity. There are thousands of AI ventures, mainly small ones, hoping to make it big, but they are all looking to the government to dole out projects. There is no market as yet for AI products in the country. Ironically, many of the technology institutes are offering courses in AI, including online ones, and computer technology companies are hiring those with AI qualification certificates. Things are happening in small circuits, and it will take time to become the mega projects envisaged by global tech behemoths like Google’s Alphabet, Facebook and WhatsApp’s Meta.

India’s AI Mission is justified in its own sphere because the government is thinking of ways to make AI work for public services, governance and public digital infrastructure, as also to adapt the Large Language Models (LLMs) of AI to the Indian situation, and looking to build Small Language Models (SLMs) for specific activities. The government is also right in its desire and efforts to make AI trustworthy and safe. The regulatory framework would be necessary to make the AI ecosystem reliable. Logically, the regulatory framework would work better when the AI ecosystem is in place and the challenges that the uses of AI throw up in practice would enable both legislators and experts to think of responsive regulations.

The white paper issued by the Office of the Principal Scientific Adviser in December 2025, titled “Democratising Access to AI Infrastructure”, by Animesh Jain, senior policy fellow (adjunct), gives a lucid picture of the status of AI in the country, with a section on “Industry Initiatives”. The paper notes: “Industry is also investing heavily in storage solutions for AI. Companies like Yotta Data Services, NTT, CntrlS and Adani ConneX have made large investments in hyperscale and sovereign cloud storage facilities. Yotta Data Services operates Asia’s largest single-building data centre in Navi Mumbai (72 MW IT load). CntrlS Data Centres operates 19 facilities with a combined load of 250 MW.”

It also shows that the Indian AI ventures are focusing on AI storage capacities and not in new generative AI models like Open AI’s GPT-5.5, Anthropic’s Claude Opus, Google’s Gemini 3.1 Pro, and these are just retail AI models. The deeper works lies at the level of LLMs to deal with subjects like advanced physics. And that would require higher-level research teams with higher academic qualifications operating at an AI company’s research and development units.

In a white paper from the Principal Scientific Adviser’s Office released in March titled “Advancing Indigenous Foundation Models”, the bureaucratic approach is laid bare. It says: “The IndiaAI Mission released a Call for Proposals for foundational AI models in January 2025 to invite startups, researchers, and entrepreneurs to develop large multimodal models, large language models, and small language models aligned to India specific needs.”

This brought in 506 proposals, of which four initiatives were selected in the first phase in April 2025, and eight initiatives in September 2025. Ideally, many those who responded to the government’s call should have approached the private investors, and the private investors in turn should have taken up parallel projects, and supported the ideas. If any of these proposals had substance, and they showed initial success, then there could have been further rounds of investment. It is how technology-driven business and business-driven technology should work. But India's is a far cry from a forward-driven market. The Indian entrepreneurs need government hand-holding because they are essentially risk-averse private investors. But once the government enters the picture, it is inevitable that it comes with a large public interest-oriented vision, which is as it should be. Then it becomes a millipede, and it cannot move forward fast enough. It cannot drive the market. There is no excitement about how lucrative the new technology is.

The government’s IndiaAI Mission has some uses, of course. The government is likely to be a big user of AI. That has its advantages, as it will be of help in the public health system, in matters of administration, in security networks and policing. The AI in the hands of the government, and in the hands of mega corporations, also has its ominous and negative dimensions. Ways have to be found to check their use of AI. Something like “ethical hackers” version of watching over the uses of AI will be necessary.

But the more important thing is that for AI to prosper in India and boost the economy, it should not be government-guided. It does not matter which party is in power. It should not be cornered by the “big bucks” monopolists either. AI should unlock the imagination and freedom of hundreds of entrepreneurs. The market, and not the government, should drive India’s entire Artificial Intelligence project.