Public Resources are used to create AI Wealth. Should the Public Get a Share?

By Nivedita Pandey

On Independence Day, PM Modi announced AI-skills training for one crore young Indians. He identified AI, data centres, quantum computing, and robotics among emerging technologies in which India should build leadership. This is strongly tied to Viksit Bharat 2047, India’s developed-nation roadmap.

Globally, governments and companies are pouring unprecedented resources into AI R&D and commercialisation. Gartner estimates that worldwide AI spending across infrastructure, software, and services will reach US$2.59 trillion in 2026.

But this wave of AI-led transformation raises a question that is hard to brush under the carpet: who will own the gains, and who will bear the costs?

AI is not built in isolation. Large AI models are built using vast bodies of human-created material accumulated over decades-including books, research papers, software code, images and other digital content, including public domain and copyrighted material, without the original creators necessarily being compensated for its use in AI training.

Governments also contribute through publicly funded universities, research institutions, digital public infrastructure and taxpayer-funded datasets. If AI companies generate enormous profits using this collective foundation often funded by public and taxpayer money, should all the rewards flow only to a handful of investors?

Global approaches to answering this question differ.

Singapore is not treating AI as a private-sector story alone. It is building national capability: committing more than S$1 billion to public AI research between 2025-2030. The Singapore government has made significant investments towards making citizens, universities, start-ups, and public institutions globally competitive in the AI economy. It has also emerged as an early adopter of AI in government operations and public-service delivery workflows, including through Pair, its internal productivity assistant for civil servants.

In June 2026, US Senator Bernie Sanders introduced the American AI Sovereign Wealth Fund Act. Under the bill, qualifying AI companies would make a one-time issuance of new shares sufficient for the US Treasury to hold 50% of all outstanding equity immediately after issuance; these holdings would form the public sovereign wealth fund. The bill authorises annual appropriations equal to 5% of the fund’s average market value, after administrative costs, for direct payments and other public purposes determined by Congress, towards the benefit of American citizens. This bill may not pass in its present form. Still, its core argument is difficult to ignore: if AI creates extraordinary wealth from public research, shared human knowledge and public infrastructure, should the rewards flow only to a handful of shareholders?

The UK has committed up to £2bn through 2030 to build a public compute ecosystem, including more than £1bn to expand its AI Research Resource 20 times.

The EU’s InvestAI aims to mobilise €200bn, including support for up to seven AI gigafactories. Neither is an ownership model of the kind Sanders proposes. But both recognise that AI infrastructure is too important to national strength to leave entirely to private companies.

India’s Aadhaar and UPI show that India can build public digital infrastructure at scale. The IndiaAI Mission, backed by ~₹10,372 crore over five years, extends the DPI legacy into AI. The next step is to turn AI access into shared value.

That means better public services, AI capability in universities and public institutions beyond a few elite campuses; quality jobs and investment across states; and genuine opportunities for start-ups beyond the largest cities. It also means guarding against public data, subsidised computing and government procurement becoming an opaque subsidy for a small circle of firms.

AI’s physical infrastructure matters too. Data centres require land, electricity and, depending on their cooling systems, water. In regions facing power or water stress, these demands must be measured and disclosed. Companies receiving public land, grid access, water allocations or incentives should report resource use and deliver tangible benefits to host communities.

Public support should not mean public risk and private reward. India should therefore establish clear public-value conditions for publicly supported AI: transparent access rules for public compute and datasets, measurable service outcomes, regional inclusion, and fair sharing of benefits where public resources and assets create exceptional private value.    

This is not an argument against private enterprise. India needs investment, competition and innovation. But when public resources help generate private AI wealth, the public deserves a transparent, durable and meaningful return.

The author is an Applied AI & Digital Transformation leader, and a Wharton alumna with experience across US, UK, India, and Singapore. She can be reached at NiveditaPandey@alumni.upenn.edu