India's deep tech indigenisation: As state-led de-risking advances, private capital must forge subsidy into compounding enterprise

This article is authored by Atharv Singh, managing partner, Purus Research.

A general-purpose technology’s most salient measure lies in the time the world takes to absorb it. In the three years since ChatGPT’s watershed release on November 30, 2022, Generative Artificial Intelligence (Gen AI) has reached one in six people worldwide, a milestone the internet took a decade to attain, and breakthroughs have arrived at a cadence unprecedented in civilisation. In its dawning, AI has already unravelled the protein folding predicament for 200 million proteins, earning its pioneers Hassabis and Jumper the 2024 Nobel Prize in Chemistry. It has formulated 3,80,000 stable new crystals, among them 528 candidate lithium-ion conductors, twenty-five times the number previously known. And it has designed drugs and moved them into human trials in under thirty months, against the three to six years conventional discovery requires. AI is the vanguard of deep technologies. Cambridge defines deep technology as “grounded in breakthrough science and advanced engineering.” Developing such technologies inherently takes painstaking decades of intense capital investment, regulatory and policy incentives, and top-echelon talent.

Deep tech

The role of capital in enabling states and the private sector to develop deep technologies cannot be overstated, yet the scale of capital alone is seldom the arbiter of their contest. Technologies that reconfigure cost curves across industries, rewrite the global military balance, and diffuse into medicine, energy, and food, while casting uncertainty over employment, are contested at the summit of global governance, and no state that grasps their paradigm-resetting ambit will leave a contender’s progress unimpeded. Between 2016 and 2022, the share of US subsidy actions aimed at dual-use and advanced technologies rose from a third to three quarters, approaching the nine tenths China has sustained since 2017. Powerful states hold proficiencies and chokeholds at different stages of the multifarious supply chains that fuel global deep technology development, and as they vie for control, the retaliatory policy measures that ensue will alter the cost curves of diffusion.

India’s stake in this struggle for technological supremacy, of which even the moderate threshold is technological sovereignty, is compounded by poverty still to be outgrown and by strategic capability too hard won to be squandered again as the world passes through the Fourth Industrial Revolution. Already the third largest economy at purchasing power parity, with the second largest military and the largest population, home to the world’s largest IT talent pool, half of its global capability centres, and a fifth of its chip designers, India unambiguously stands within hailing distance of the summit where these technologies are contested. Its participation will define its ability to manoeuvre around obstacles to its interests in a world where the space for national sovereignty is narrowing amid a great power condominium. Yet despite its scale and depth across the global economic, military, and technological complexes, and its vibrant demography, India remains a nation earning a meagre $2,813 per head, a standard of living far below what the modern age affords. Its manufacturing base is expanding, albeit gradually and unevenly, and it spends a scant 0.64% of Gross Domestic Product (GDP) on research, trailing far behind the United States, China, and even fellow emerging economies such as Brazil, a sign of an economy still given to rent-seeking and incumbent inertia. For years, India lacked the indispensable features a frontier-technology-led industrial ecosystem requires, and it had to come to terms with the consequences as other powers dictated its ability to defend and develop itself.

In May 1992, the US imposed sanctions on ISRO and Glavkosmos, the Russian space agency, voiding a transfer of cryogenic upper stage technology valued at ₹235 crore. India began anew in April 1994, and it was not until January 2014 that the indigenously built CE-7.5 reached orbit aboard GSLV-D5. In 1999, with enemy infantry assaulting the Kargil heights, the United States reportedly withheld GPS data, and India went on to build NavIC for sovereign navigation and timing at a cost of ₹3,260 crore.

Full stack sovereignty, a tall order by any measure, is by far India’s most assured pathway towards sweeping strategic capability and mass prosperity. At Purus, we reckon the cumulative investment required to forge capability across every layer of the stack at ₹47 to 66 lakh crore ($500 to 700 billion) over the decade to 2035, equivalent to 2.5 to 4% of India’s gross fixed capital formation in that period. At the ₹56 lakh crore ($600 billion) midpoint, that spending would engender 4.7 to 7.1 million direct jobs and about 3 million construction job years. In a country where merely 23.6% of workers hold regular salaried employment and female labour force participation languishes at 40%, building an Indian deep tech stack would, should it succeed, formalise and feminise the workforce on a scale without precedent.

A Purus Research audit elaborates on the centre’s groundwork by analysing approvals worth ₹5.4 lakh crore ($57 billion) since 2020 across more than a dozen schemes, from semiconductor capex subsidies and the AI mission’s compute allocations to deep tech funds of funds, supplemented by seven Production Linked Incentive (PLI) sectors that furnish the underlying hardware. While the government continues to commit risk capital, the ensuing compounding that transforms a subsidised plant into an enduring enterprise hinges on private capital.

Globally, pension funds, endowments, and insurers are the anchor investors in such intense capex cycles. India, while holding prodigious pools of long duration capital in its insurance reserves and pension corpus, is a conspicuous exception. Its regulatory architecture cordons that capital off from the illiquid, decade-long positions venture capital and private equity demand, though insurance and pension liabilities, themselves decades long, are precisely suited to carry them.

Policy aligned deep tech is already among the best performing equity themes on earth. Between January and May 2026, deep tech claimed the top five places among US equity themes, with SpaceX completing the largest IPO in history. Meanwhile, McKinsey estimates the net internal rate of return on deep tech funds since 2003 at 16%, compared with 10% for mainstream tech funds.

Moving forward, deep tech’s stellar performance is likely to be bolstered by three secular capex cycles, each policy backed, capital intensive, and set to run for the better part of a decade or longer. These are NATO’s rearmament under its pledge to spend 5% of GDP on defence and security by 2035, the global compute build, which McKinsey estimates will absorb $6.7 trillion of capital by 2030, and the clean energy transition, which the International Energy Agency (IEA) approximates at $2.2 trillion of investment a year.

India’s private markets are exhibiting early signs of deep tech momentum. IVCA logged a record $2.96 billion across 189 deals in 2025, even as the wider Indian tech startup market, all sectors combined, contracted 17% to $10.5 billion. Tracxn’s analysis pegs deep tech startup funding at $2.22 billion for January to August 2026, up a staggering 157% on the same months of 2025. Exit value, meanwhile, quadrupled from $152 million across 7 exits in 2024 to $600 million across 18 in 2025. Yet of 8,842 Indian deep tech startups, only 373 have passed Series A, a bottleneck that keeps the country’s haul under 2% of the world’s deep tech capital and leaves 62% of funds naming exits as their binding constraint. What remains open is the dynamism private capital can add, the successive innovations it may enable, and the returns it can earn.

(The views expressed are personal)

This article is authored by Atharv Singh, managing partner, Purus Research.