India Must Not Compete on Borrowed Foundations. Sovereign AI Capability Is Now an Economic Necessity.

‍India must maintain its strategic lens on technological sovereignty, but discard its mistrust of its own innovators.  The state must urgently act as mission setter, risk absorber and anchor customer for private champions across the full AI stack.

‍The off-switch is out of our reach

Recently, Washington ordered America's leading artificial intelligence laboratory to cut every foreign national off from its most advanced models, and the company complied within a day. The lesson is not about one company or product. It spans from chips and rare earths to advanced materials, and AI is now its sharpest instance.

Technology and the supply chains beneath it have become instruments of statecraft, switched on the moment trade or diplomatic tension rises. Any nation that runs its AI economy on borrowed foundations inherits that exposure, and it inherits it at a layer far deeper than hardware ever reached.

When countries restricted advanced chips or rare earths, the dependency was visible and physical. With AI it sits inside the layer where economic value, talent specialisation, security infrastructure and technical intellectual property will compound for a generation.

A country that rents its intelligence does not merely pay a licence fee. It cedes the ground on which its future industries, defence systems, digital security, and public services will be built. 
‍

India had the right lens but the wrong instrument

India has heard this warning before. Ashok Parthasarathi, science and technology adviser to Indira Gandhi and a principal architect of post independence technology policy, diagnosed three things early and correctly. Technology is sovereignty. Dependence on foreign suppliers in strategic sectors is a security risk and not merely a question of efficiency. A nation must invest for the long term in science and technology rather than wait for the market to do it. The technology embargoes that followed India's nuclear and space programmes in the 1960s and 1970s made that fear rational rather than ideological.

His error lay in the instrument he chose. Parthasarathi reached for state dominance and a reflex of suspicion toward Indian private enterprise, rather than a model in which the state sets the mission and attracts private capital and indigenous innovation in behind it. State control managed a few islands of capability in atomic energy, space, and basic electronics, but at a high systemic cost of protected monopolies, gacial diffusion into the economy, constrained iterative research, nanny-state over-regulation, and chronic under investment in quality. 

His costliest premise was the one about equity. Parthasarathi held that technology development should serve the bottom of the economic pyramid first, and that view, applied to every domain, misread how technology actually matures in an emerging economy. Technologies of mass impact in areas such as energy, agriculture, health, and transport must reach as wide a section of society as fast as possible. Frontier technologies are fundamentally different. Semiconductors, advanced computing, defence systems, advanced materials, robotics, and space are niche by nature and may not touch the whole population at once. They demand separate budgets engineered for cutting edge innovation and global competitiveness, judged on the capability they create rather than on their immediate reach.

The opening after 1991 then unleashed Indian firms building businesses of global scale in software, telecom services, banking, and pharmaceuticals far faster than any plan had anticipated. It became clearer to all idealogues of state control that we had lost an entire generation of potential innovation to other Asian economies that had opened up before us.
‍

The world's largest economies fund frontier technology intentionally 

India need not reinvent what works, because the economies above USD 5 trillion have already demonstrated outcomes. The enthusiasm of the Indian state for deep technology is now substantial in rhetoric, yet the mechanisms that deliver real support remain trapped in bureaucratic cholesterol. 

The United States earned its technological lead over decades, and it sustained it through bipartisan investment that attracted the best minds in the world. Through DARPA, BARDA, the National Science Foundation, NASA, the Department of Defense and more than twenty other grant giving agencies, it has funded trillions of dollars of national research across universities, corporations and startups, then backed that research with sustained procurement to pull young technology to commercial maturity. 

Oracle's first customer for its relational database in the late 1970s was the Central Intelligence Agency. More recently, Anduril, a defence technology startup founded only in 2017, took over a United States Army contract worth up to USD 22 billion (about Rs 1.8 lakh crore) for an augmented reality combat system, won a Marine Corps counter drone contract worth up to USD 642 million (about Rs 5,300 crore), and secured a separate air defence production award of about USD 250 million (about Rs 2,100 crore). Famously, NASA granted over USD 15 billion of revenue to SpaceX since its founding, indisputably laying down the foundation for what eventually became the largest IPO in economic history. US State investment and anchor demand has consistently evolved young companies into globally relevant innovators every decade.

China built a competitive playbook within three decades through state sponsored intellectual property and process onshoring on a scale no economy had attempted before, ranging from luxury goods and electric vehicles to smartphones and high speed rail. The methods were aggressive. The Federal Bureau of Investigation has estimated that the theft of intellectual property, in which China is the principal infringer, costs the American economy between USD 225 billion and USD 600 billion each year through counterfeiting, piracy and trade secret theft. A Wall Street Journal review found that Huawei received about USD 46 billion in loans and credit lines from state lenders, alongside roughly USD 25 billion in tax breaks between 2008 and 2018. Those methods may breach global norms on intellectual property, yet the outcome is not in dispute. China is now a genuine competitor across the global technology economy.

India, by contrast, pursued a public sector led model from independence onward that never built an indigenous innovation ecosystem at the scale a USD 4 trillion dollar economy now requires.
‍

The Indian state must set missions, absorb risk, and buy first

The way forward is to keep Parthasarathi's strategic lens and discard his mistrust. The Indian state should act as mission setter, risk absorber and platform builder, working with Indian entrepreneurs, corporates, and capital providers rather than manage exceptions and what-if scenarios around them. It should designate Indian innovators as National Champions and have public sector undertakings cultivate a rainforest of private companies and startups beneath them, developing them by serving as the anchor customer, cutting the bureaucracy around qualification criteria, and prizing velocity over completeness.

The state has begun this work, and the past decade has shown real progress, including the indigenous systems that performed during Operation Sindoor in May 2025. The intent is no longer in question. The scale and the speed of execution are. 

The Research, Development and Innovation (RDI) scheme of Rs 1 lakh crore (about USD 12 billion), approved in 2025, is a welcome start, yet it is structured through multiple qualification criteria, step-down allocators, and suffers from constant personnel changes that restart evaluation processes. It is dwarfed by the scale, velocity, and predictability of American and Chinese spending. The ceiling on a single Anduril contract, about USD 22 billion, is over 1.5 times of India's annual allocation of USD 14.6 billion (₹1.39 lakh crore) earmarked for procurement from domestic industry in the defence budget. 

The IndiaAI Mission is a genuine step forward, launched in 2024 with an outlay of Rs 10,371 crore (about USD 1.2 billion), onboarding more than 38,000 graphics processing units to a subsidised common compute facility and backing over 20 teams to build indigenous foundational models. Initiatives such as iDEX and IN-SPACe have attempted to open defence and space procurement to startups, with initial success.

However, the velocity of execution still trails ambition. These missions and schemes must absorb more qualified personnel from industry, such as the Aadhar or UID program did last decade. They must be evaluated on efficiency of transmission and enrolment, and be managed by accountable professional asset allocators, not government committees that rotate membership every quarter. 

Furthermore, winning a pilot or grant is no guarantee of scale contracts, and most programmes stop at tiered grants or uncertain integration. The private sector cannot make multi decadal allocations on a technology that has no commercial market yet. That is precisely the gap the state exists to close. It must become the grant driver and the anchor client, giving domestic deep technology companies both the grant capital and the demand certainty they need to innovate.
‍

Sovereign AI is a full stack and not a handful of indigenous models

Sovereign capability in AI must be defined precisely, or it will collapse into a vanity project. It is not one national large language model unveiled at a summit. It is a full stack, large frontier models alongside small language models and vertical models tuned for Indian sectors, all trained on data that reflects how Indians actually live, work and speak across our regional languages. A model that cannot serve a farmer in Karnataka or a doctor in Rajasthan in their own language is not sovereign capability. More importantly, indigenous models that do not have Government agencies and departments as customers, after receiving State grant support, will leave more questions unanswered.

The principle to apply is whole chain sovereignty. Map the dependency across compute, models, and data, then close it systematically, layer by layer, in partnership with Indian innovators and enterprises. The economic case is as vital as the strategic one. Owning the core technology means India captures the value it creates, builds the depth of talent that compounds over time, and exports finished products to the world rather than renting intelligence back from it at a price set elsewhere. 
‍

Animal spirits return only when the bureaucratic bullying ends

None of this will move at the required pace without a deeper change in how the state treats those who build in India. The reform agenda is familiar: ease of doing business, faster customs, lowered friction from regulations and filings, safe harbour provisions that nurture multi-decadal investments and innovation, and directed incentives. However, these must be pursued with far greater urgency than it has been. 

For decades the operating ideology inside government, the same instinct that Parthasarathi's contemporaries carried, treated private producers with suspicion and supressed their drive to take risks. The missing complement to India's supply side ambition is State demand. Anchor purchasing and grant support, directed at Indian AI companies building core technology indigenously, would do more to revive the animal spirits of Indian producers than any statement of intent.
‍

The window closes this decade, and India has already proven it can compete

The cost of moving slowly is best read in what successful economies have already built. BYD, the Chinese automaker, now employs close to 110,000 engineers (1.1 lakh), the largest research workforce in the global automotive industry, and files around 32 patents on every working day. That depth let it leapfrog established Western incumbents. China built more than 30 million passenger vehicles (3 crore) in 2025, against roughly 10 million vehicles (1 crore) of all kinds produced in the United States. The lesson is straightforward. markets reward those who invest in proprietary capability, build scale, and then compete globally. 

The window is finite. Over the next ten years the world will reorganise its supply chains in energy, technology, AI, and manufacturing, and the nations and companies that claim share now will hold it long into the future. In AI the lock in is harder still, because advantage compounds through data and talent that accumulate where the IP is held.

India has already proven it can build national category leaders under constraint. Many of its firms have created global-scale businesses without the heavy subsidies and guaranteed state procurement that fuelled the rise of American and Chinese technological corporates, and often while competing against public sector undertakings on unfair terms. 

If the State is able to redirect its growing revenues into supporting manufacturing scale, focused research, and international expansion, pair private ambition with a state machinery that finally trusts it, then India can own the rails it runs on. The alternative is to keep building on foundations that someone else can switch off on a morning of their choosing.
‍

DISCLAIMER

_{The views expressed herein are those of the author as of the publication date and are subject to change without notice. Neither the author nor any of the entities under the 3one4 Capital Group have any obligation to update the content. This publications are for informational and educational purposes only and should not be construed as providing any advisory service (including financial, regulatory, or legal). It does not constitute an offer to sell or a solicitation to buy any securities or related financial instruments in any jurisdiction. Readers should perform their own due diligence and consult with relevant advisors before taking any decisions. Any reliance on the information herein is at the reader's own risk, and 3one4 Capital Group assumes no liability for any such reliance.Certain information is based on third-party sources believed to be reliable, but neither the author nor 3one4 Capital Group guarantees its accuracy, recency or completeness. There has been no independent verification of such information or the assumptions on which such information is based, unless expressly mentioned otherwise. References to specific companies, securities, or investment strategies are not endorsements. Unauthorized reproduction, distribution, or use of this document, in whole or in part, is prohibited without prior written consent from the author and/or the 3one4 Capital Group.}

‍