Opinion Our digital infrastructure is built on systems from China and the US. But India has all the ingredients to close this gap

Consider a typical success story: A leading Indian health-tech app, trusted by urban professionals and rural patients alike, storing millions of sensitive medical records on foreign clouds. Its interface taps global APIs for payments, messaging and analytics. Behind every diagnostic suggestion hums a machine-learning model trained in Silicon Valley or Beijing, funded by cross-border capital. On the surface, it is “Made in India.” Under the hood, it is assembled from borrowed parts.

This isn’t mere academic quibbling over “digital sovereignty.” When your codebase lives in distant data centers, you inherit their risks — sudden regulatory lockdowns, supply-chain shocks and opaque pricing. Compliance becomes a game of whack-a-mole across jurisdictions. A whiff of geopolitical tension can threaten your IP, disrupt your uptime, and turn your customers into collateral damage. Ask any entrepreneur who’s scrambled to replace a deprecated API or fought to rehost terabytes of data after a cloud-provider price spike. The cost is not just financial. It’s strategic fragility.

The case for an AI Commons

The irony is that India has all the ingredients to close this gap — exceptional engineers, universities producing hundreds of thousands of computer-science graduates, patient capital from pension funds and sovereign reserves, and a proven track record of digital public goods. UPI and Aadhaar exploded because they were designed as open, interoperable platforms, built with public-sector backing but open to private innovation. ISRO showed us that resource-constrained genius can outpace many wealthier rivals. Why shouldn’t we apply the same formula to AI and critical software infrastructure?

The answer is the AI Commons: A public-private open-source ecosystem for foundational models, platforms and code libraries, seeded with Indian capital and governed by a consortium of academia, industry and civil society. Think of it as the UPI for AI. An interoperable, transparent, sovereign resource that encourages experimentation and shared ownership. The Commons would host core language, vision and speech models trained on diverse, India-centric data, including regional languages, local dialects, and domain-specific corpora to ensure cultural relevance and mitigate bias. It would provide standardised APIs for developers at all scales, from solo founders to enterprise giants, with clear licensing and audit trails. And it would be underpinned by Indian-run data centres, secured against foreign coercion.

Skeptics will worry about “reinventing the wheel” when the likes of OpenAI, Google and Meta pour billions into model development. But the wheel we need is different: Tuned to our regulatory context, our multilingual complexity, our social fabric. Global models absorb Western norms and then blast them through a worldwide megaphone. They treat Indian English as an accent, not a distinct variant with its own idioms. They gloss over rural healthcare realities and small-town commerce. Worse, they slap us onto carbon-hungry GPUs farmed in carbon-cheap but geopolitically risky locales. A Commons model, by contrast, would weigh local energy mixes, incentivise green compute, and align with India’s net-zero pledge.

The AI Commons isn’t isolationist. It would interoperate with international repositories and academic research, enabling two-way exchange rather than one-way dependency. Just as UPI joined hands with Singapore’s PayNow and Bhutan’s e-payments, the Commons could federate with Europe’s GAIA-X and other national AI lighthouse projects, so that innovations flow freely, governed by shared standards and mutual trust.

Financing a venture of this scale isn’t science fiction. Our National Infrastructure Pipeline has already mobilised trillions into roads, rails and renewable energy. Why not a similar corpus for digital sovereignty? A Digital Public Goods Fund, seeded by the RBI’s reserves or a dedicated sovereign wealth vehicle, could co-invest in computer clusters, open-source labs and talent pipelines. Pension funds, with their decades-long horizons, would find the steady governance and defensive moat of public-good IP an attractive anchor for their allocations.

Beyond money, success demands governance and culture. The Commons would be governed by a charter that guarantees open access, rigorous auditability and equitable contribution. A rotating board of technologists, ethicists and users would oversee licensing, security audits and socio-economic impact assessments. And a national academy of digital commons would foster skills at the grassroots, offering fellowships and hackathon grants to students, start-ups and NGOs.

Closing India’s code deficit isn’t just about resilience; it’s about shaping the global narrative. We’ve shown the world that a middle-income country can leapfrog entrenched systems — banking without branches, identification without paperwork, space missions on shoestring budgets. Can we now pioneer a model of collaborative AI that balances innovation with public interest, sovereignty with solidarity? If so, India won’t merely digitise. It will define the next generation of digital public goods.

We’ve built the highway. Now we must build the engine.

The writer is a vice president at a global investment firm. Views are personal