Nationalising AI? Why Govts Want a Piece of OpenAI and Sarvam

Governments don’t usually take equity stakes in private companies, let alone startups, but when they do, it usually signals trouble. Historically, the US government took majority equity in companies like General Motors and AIG to bail them out during the 2008 financial crisis. The UK, too, has a ‘golden share’ in Rolls-Royce to prevent hostile takeovers. 

Now, curiously, governments are looking to be shareholders in AI companies, but for very different reasons.

The Indian government is considering acquiring a 1–2% stake in homegrown AI startup Sarvam AI, valued at around $1.5 billion, according to a report by The Economic Times. 

Similarly, The Financial Times reported that OpenAI has discussed giving the US government a 5% stake as the $852 billion AI giant seeks to secure political support from the Donald Trump administration. OpenAI also reportedly proposed a broader framework that would allow a government-backed vehicle to acquire minority stakes in other leading US AI companies, including Anthropic, Google, and Meta.

While political support and national security seem to be the defining factors, as the US government has dilly-dallied when it comes to restricting both AI chips and models, there are lots of other considerations at play.

“I think the biggest factor behind these trends is sovereignty and the ability to build and ensure access to frontier AI models,” Anwesha Sen, Assistant Programme Manager, The Takshashila Institution, tells AIM.

She argues that technology has always been an instrument of national power. With AI, however, the geopolitical discourse is increasingly framed as an arms race, influencing how states think about capability, dependency, and leverage.

Similarly, TV Mohandas Pai, Chairman of Aarin Capital and former CFO of Infosys, observes that the United States increasingly views leading AI companies as strategic assets that strengthen its technological and geopolitical position.

“The US considers critical AI companies as its force multiplier to dominate the world in technology,” he says. “They are very clear in what they want to do and what they want to achieve.”

This is not the first time the US government has bought a stake in a critical tech company. In August last year, the Trump administration acquired a 9.9% stake in Intel, purchasing 433 million shares for $8.9 billion.

Adarsh Ranjan, Research Analyst at Carnegie India, believes the US has been acquiring stakes in companies it deems critical to national security across sectors such as nuclear energy, critical minerals, AI, semiconductors, and quantum. “While most government incentives have traditionally taken the form of grants, loans or procurement, direct equity ownership is a new investment vehicle that is now being explored,” he says.

One vehicle is the US International Development Finance Corporation (DFC), which has pledged equity investments across sectors including critical minerals, energy, infrastructure and quantum. In June, it approved $2.5 billion in strategic investments to access critical minerals and strengthen US supply chains.

Benefit to Taxpayers 

Pai says governments believe that investing public money in frontier AI companies could eventually generate returns for taxpayers, while also protecting strategic industries.

OpenAI wants to take that route through a proposed Public Wealth Fund. It wants such a fund to invest in diversified, long-term assets that capture growth in both AI companies and the broader set of firms adopting and deploying AI. 

“Returns from the Fund could be distributed directly to citizens, allowing more people to participate directly in the upside of AI-driven growth, regardless of their starting wealth or access to capital,” the policy document reads.

In a similar vein, Ranjan says government equity ownership creates a mechanism for distributing the gains from AI to people through dividends.  

“This is about building trust and showing that AI companies care about people, their labour, and their creations, thereby mitigating some of the harm to their image arising from the perception that they have built these systems on people's work without compensating them,” he says.

IndiaAI Mission and Strategic Partnerships

Compared to the US, India has embraced a more structured approach through its ambitious IndiaAI Mission, approved with a budget outlay of ₹10,371.92 crore. The government has already committed significant funding towards AI compute infrastructure, datasets, and indigenous foundation models. 

Taking an equity stake in Sarvam would translate into treating AI as strategic national infrastructure.

“I believe some of the funds that are given have the right to convert into equity at a particular price depending on future rounds,” Pai says.

He believes this approach should be adopted more widely.

“I think the government should spend more money on giving grants and taking equity positions in many frontier tech companies,” Pai says. “It provides capital to those companies so they can invest in deep technology and grow. The government will also benefit.”

That said, he acknowledges that some startups would inevitably fail. 

National security concerns remain central to this approach. The US government’s restrictions on Anthropic and OpenAI’s models, and China also now considering a similar tactic, leave Indian AI startups without any leverage. Sen says a government that lacks a frontier AI model, or depends entirely on another country's models for its critical systems, is in a position of structural dependency. 

“As a result, it’s not surprising that governments are moving away from the older playbook of merely offering incentives toward something more direct, i.e., seeking partial ownership,” she notes.

Competitive Advantage

Sen acknowledges the risk of governments effectively picking winners instead of letting markets determine outcomes through competition. That said, she notes that governments also fear falling behind in the global AI race.

“Given the strategic importance of building and accessing frontier AI models, coupled with the FOMO governments feel about being left behind in both innovation and diffusion, it can be argued that states first need to ramp up development and acquire these assets,” she explains

Pai says governments increasingly see AI in the same category as semiconductors, defence, and other critical technologies where market forces alone may not be sufficient. Public spending then takes the form of production-linked incentive (PLI) schemes, defence procurement and support for manufacturers.

However, Pai warns against concentrating public money in only a handful of companies. “The government must not take stakes in one or two companies,” he says. “They must take stakes in five or 10 companies in every vertical.”

Meanwhile, Ranjan says the challenge is making sure government ownership doesn't come at the cost of innovation, focusing on minority stakes rather than large ownership positions. 

Yet, he acknowledges that India is still quite a way behind. “We are only just beginning to figure out how government investments and patronage in these companies should work.”

China Plays a Different Game

India’s and the US’ approach to backing AI startups stands in sharp contrast with China’s long-term industrial strategy.

“China gives a lot of grants, soft loans and long-term loans at low cost because they want these companies to become sector champions,” Pai says.

According to researchers from Stanford University and the National Bureau of Economic Research, China's central and local government venture capital funds channelled about $912 billion into strategic industries over the past decade. Around 23% of that investment was directed toward AI-related firms, a category the researchers estimate includes roughly 1.4 million companies nationwide. 

China has continued to build on that approach. In March 2025, Beijing launched a 1 trillion yuan ($138 billion) National Venture Capital Guidance Fund to support early-stage technology startups, including AI.

“China’s policy is one of domination,” Pai says. “They will continue spending for a long period of time.”