India’s Biggest Startup Backer Is Hiding in Plain Sight

The most consequential backer of Indian venture funds over the past decade has not been a venture capital firm. It has been the Government of India, and it has spent that decade going to considerable lengths to make sure nobody could accuse it of picking companies.

The numbers are larger than the ecosystem tends to assume. SIDBI’s annual report records gross commitments of ₹11,808 crore across 153 alternative investment funds under a single scheme, a target it passed ahead of schedule. Those funds had in turn put roughly ₹25,548 crore into 1,371 startups by the end of 2025. Few investors of any kind have backed as many Indian venture funds. Yet in all of it, the selection decision, the judgement about which founder and which company, sat with a professional fund manager rather than with a ministry.

That was a design choice rather than an accident, and it has a specific Indian history behind it. It also stands in sharp contrast to what the two largest economies did with public risk capital over the same years. Washington has increasingly converted industrial-policy support into direct equity positions in strategically important companies. Beijing built guidance funds at a scale no other state has attempted and spent years struggling to keep their decisions commercial. India built the intermediation and stayed off the cap table.

That position is now shifting, and it is shifting quietly, across several instruments rather than in a single announcement. The Research, Development and Innovation Scheme permits equity of up to 25 percent in the companies it backs. Under the semiconductor programme cleared in July, the state will match private venture rounds directly, on the same terms and in the same share class as the investors leading them. The ₹1,000-crore spacetech initiative still preserves the old intermediary model through a professionally managed fund. RDI and Semicon 2.0 go further: for the first time, the architecture explicitly allows public capital to travel all the way onto company cap tables.

So the question worth asking in 2026 is not whether the government should be investing in startups. It already is, at considerable scale, and the arm's-length model has largely worked at building fund managers and crowding private capital into the ecosystem. The question is whether the discipline that made it work survives the move closer to ownership.

How India became a limited partner

The Fund of Funds for Startups, unveiled in January 2016 with a ₹10,000 crore corpus, established the template that everything since has followed. It does not invest in startups. It commits capital to SEBI-registered Alternative Investment Funds, which are then required to invest a multiple of that commitment into recognised startups, and which make their own selection decisions. The scheme is, by construction, barred from choosing a company. Of the ₹25,548 crore those funds had deployed by the end of 2025, ₹3,803 crore went to 205 women-led ventures.

The effect on the market was structural rather than merely financial. Anchor commitments from a sovereign-backed institution gave first-time fund managers the credibility to raise from private limited partners in a market that had almost no domestic institutional LP base. SIDBI became, and remains, the largest single limited partner in Indian venture capital. It did so without ever picking a company. That separation was not bureaucratic caution for its own sake. It was the product.

Startup India Fund of Funds 2.0, notified on April 13, 2026 with operational guidelines out on April 25, refines the same idea rather than replacing it. A fresh ₹10,000 crore is segmented into four buckets: deep tech funds with no corpus cap and tenures of up to 18 years, micro venture capital funds capped at ₹400 crore, technology-led manufacturing funds, and sector-agnostic funds. The private capital multiplier is dialled by segment, set lower for deep tech at 1.5 times and higher for sector-agnostic funds at 2.5 times, which is a precise way of saying the state will subsidise hard science more heavily than it subsidises the categories private capital already likes. Distributions, net of up to five percent earmarked for ecosystem capacity building, return to the Consolidated Fund of India. This is public money structured as a revolving investment, not a grant.

The ₹1 lakh crore Research, Development and Innovation Scheme, cleared in July 2025 and launched that November, uses the same instinct at a larger scale. Capital sits in a Special Purpose Fund under the Anusandhan National Research Foundation and flows through a two-tier structure to second-level fund managers, initially the Technology Development Board and BIRAC, which run their own calls and appraisals. Support takes the form of long-tenure loans at three to four percent, equity of up to 25 percent, or a hybrid. By March 2026, more than a hundred Indian venture firms had applied to participate. The first five cheques went out in May 2026, to startups in space, drones, energy storage, medical robotics and research instrumentation. Manish Kheterpal of WaterBridge Ventures put the parallel plainly: RDI is to the deep tech ecosystem what SIDBI was to the startup ecosystem over the last decade.

Even the space bet followed the pattern. When the government committed ₹1,000 crore to spacetech, it did not write cheques to launch companies. IN-SPACe anchored the Antariksh Venture Capital Fund, a Category II AIF with a ten-year life managed by SIDBI Venture Capital, which reached a first close of ₹1,005 crore in November 2025 against a ₹1,600 crore target.

The pattern across all of it is consistent. For most of the architecture built over the past decade, a professional intermediary has sat between the taxpayer and the cap table.

This was a choice, and India had made the opposite one before

The arm's-length design reads as caution. It is better understood as institutional memory.

India’s venture capital industry was not born in a garage. It was created by the state. IDBI started a venture fund in 1986. In January 1988, ICICI and the Unit Trust of India jointly founded the Technology Development and Information Company of India, the country’s first institutional venture capital firm, alongside the Risk Capital and Technology Finance Corporation under IFCI. The World Bank selected six Indian institutions to begin venture investing. Guidelines issued that November defined venture capital so narrowly, restricting it to innovative technologies from first-generation entrepreneurs, that the category became difficult to practise commercially. State-controlled development finance institutions at the centre and in the states were, for roughly a decade, the whole of Indian venture capital.

That generation of institutions carried a structural problem rather than a personnel one. A development finance institution appraising long-gestation technology projects directly held selection risk, sectoral risk and institutional accountability on a single balance sheet, with far less institutional tolerance for the portfolio-style failure that venture capital requires. Venture returns depend on a portfolio in which most positions do not work; an institution that must justify each individual decision cannot easily run one. The failures of that era were failures of instrument design.

Fund-of-funds intermediation is the direct answer to that problem. The state supplies capital and sets the mandate; a professional manager with skin in the game, a defined tenure and a carry structure absorbs selection risk and is measured on the portfolio rather than the position. India did not stumble into being a limited partner. It arrived there having already tried the alternative.

The rest of the world went the other way

What makes this worth writing about now is that between 2025 and 2026 the two largest economies moved decisively toward ownership, and India’s design started looking less like caution and more like a considered position.

In the United States, the Department of Commerce converted roughly 10 percent of Intel into a non-voting federal equity position out of unpaid CHIPS Act grants. The Department of Defense took a 15 percent stake in MP Materials, making the Pentagon the largest shareholder in the country’s only integrated rare-earth producer. Other transactions followed across strategic industries, turning what might have looked like an exception into a broader policy direction. Larry Summers characterised the pattern as deals-based capitalism rather than rules-based capitalism, and the substance of that critique is about consistency: case-by-case negotiation is harder to hold to a standard than a rule applied to everyone.

China took a different route to a related place. Government guidance funds, launched in 2005, numbered more than 1,800 by 2021 with an aggregate target of about 1.52 trillion dollars.  But scale did not automatically produce efficient capital allocation. Research published in The China Quarterly found only 26 percent had met their target capital size, while audits and subsequent reforms exposed difficulties raising private capital, deploying funds and creating exits. In December 2025, Beijing announced a National Venture Capital Guidance Fund targeting a trillion yuan, with a twenty-year life and at least 70 percent directed to seed and early-stage companies.

That last detail is the interesting convergence. Beijing concluded that state venture capital needs tenures roughly double the private norm. India’s FoF 2.0 permits deep tech funds to run up to 18 years and extends deep tech startup recognition to 20 years under the February 2026 gazette that defined the category for the first time. Two very different systems arrived at the same insight about patient capital from opposite directions. The difference is who holds the pen on the investment decision.

What the model bought, and where it runs out

The honest assessment of the arm's-length decade is that it did the job it was designed for and is now bumping against a constraint it was not designed for.

It built a domestic institutional LP base where none existed, seeded a generation of fund managers, and helped reduce the ecosystem’s dependence on foreign capital. Those were the stated objectives in 2016 and they were materially advanced. 

Deployment has been the persistent friction, and the gap is documented across years rather than resting on any single estimate. In 2023, SIDBI’s deputy managing director put sanctions at around ₹9,500 crore against disbursements of roughly ₹4,500 crore. Policy reviews covering the preceding period found a similar ratio, with disbursals running at about 43 percent of commitments. Reporting in late 2025 suggested the proportion had improved but not closed. Drawdowns run on fund managers’ own deployment schedules, which is a feature of the structure rather than a fault in it, but it means committed capital and working capital are different quantities. SIDBI has since replaced the fixed drawdown formula with a graded structure that lets funds call larger amounts as they actually deploy, which is the system responding to exactly this friction.

The larger point is that the market has changed underneath the instrument. Indian startups raised 5.2 billion dollars in the first half of 2026, down 9 percent year on year, with deal count up 7 percent to 501 and late-stage funding down 29 percent to 2.2 billion dollars. Yet over the same six months, according to EY and IVCA, Indian private equity and venture fundraising more than doubled to 21.2 billion dollars across 48 funds, while actual investment fell 36 percent to 20.5 billion dollars. Money is being raised into funds faster than it is being put to work. Dry powder is high and rising.

Which means additional limited partner capital, on its own, is pushing on a door that is already open at least at the aggregate fund-raising level. The binding constraint has moved. It is now the willingness to underwrite hard, long, capital-hungry categories, and the visibility of an exit at the far end. That is precisely the gap the equity turn is aimed at. This is the strongest argument for why the state is changing instruments rather than simply increasing the size of the old one.

Why the equity turn is defensible, and what will decide it

Judged as instrument design, the Semicon 2.0 structure is careful. The government invests pari passu with private venture funds, on identical terms and in the same class of shares. It takes no board representation and disclaims any role in day-to-day operations. It states an intention to exit at prevailing valuation and recycle proceeds. Founders may buy the stake back. For larger recipients, the royalty clawback of 1.5 times substitutes for equity altogether. RDI applies a comparable logic with its cap of 25 percent. These are the terms a thoughtful co-investor would write, not the terms of a state building a portfolio of national champions.

Three things will determine whether it ages well.

The first is pricing. A matching investor on identical terms has outsourced valuation to the private lead, which is elegant and efficient, but works only if the private lead is genuinely independent and has done real price discovery. Where a government cheque is the reason a round clears, the reference price it is matching may be partly its own reflection.

The second is exit. States are historically better at entering positions than leaving them, because the moment to sell is rarely politically convenient. Amitesh Kumar Sinha, Additional Secretary at MeitY and chief executive of the India Semiconductor Mission, has been explicit about exiting at prevailing valuation. The real test arrives the first time the prevailing valuation is a loss, and the institutional answer to that question is worth building before it is needed rather than after.

The third is neutrality. Once the state holds equity in one firm in a sector, every regulatory, standards and procurement decision touching that sector acquires a second reading. The convergent view at the World Bank, the IMF and the OECD is that state ownership can work under specific conditions: professional management, competitive neutrality, a clear mandate, transparency, and separation of the ownership function from the regulatory one. India’s current structures satisfy most of these by construction. Keeping them satisfied as the portfolio grows is an active task, not a settled one. And this may be the hardest test of all. The financial risk of a bad investment is visible. The institutional risk of appearing to favour a company in which the state owns shares is harder to measure, but potentially more damaging.

The strongest case against all of this is the general one: that public ownership has a poor long-run record, that it tends to entrench the problems it was meant to solve, and that arm's-length regulation achieves the same ends more reliably. That critique deserves to be taken seriously rather than waved through. The reasonable response is that India is not proposing ownership as a destination. Every instrument in the stack has an exit written into it at the design stage, which is the difference between an investment programme and a nationalisation. 

The lever that is already working hardest

There is also a simpler way for the state to de-risk a startup without owning any of it: become its customer. By volume, India is already doing far more of this than it is investing. The Government e Marketplace completed ten years on 9 August 2026, with cumulative gross merchandise value crossing ₹20 lakh crore across more than 3.78 crore orders. Within that, 42,242 startups have won orders worth over ₹65,633 crore. Startups on GeM are exempted from the prior-experience, prior-turnover and earnest-money requirements that would otherwise exclude them. Defence has run a parallel version through iDEX, which has moved from prototype grants into procurement orders and acceptance-of-necessity approvals worth thousands of crore.

Measured against the public money actually disbursed to funds under the first Fund of Funds over a comparable decade, which has run in the low thousands of crore, the state has moved substantially more capital to startups as a customer than as an investor. The two are not the same unit; an order is revenue and a commitment is capital, and GeM’s startup cohort includes companies well past the early stage. But the comparison holds where it matters. Revenue does not dilute anyone, it validates a product in a way no term sheet can, and it is the single most powerful de-risking signal a government can send to private investors looking at a category they do not yet understand. That suggests an important principle for the equity experiment: ownership should remain the instrument used when procurement, grants, debt and fund-of-funds capital cannot solve the financing problem—not the default simply because the state now has the ability to invest directly.

The version India is building

Put the pieces together and something coherent appears, even though the individual programmes have largely been discussed separately.  India has a fund-of-funds layer that builds managers, a research financing layer that reaches technologies too early for those managers, a demand layer that converts capability into revenue, and now a co-investment layer for the categories where none of the first three is sufficient on its own. Each instrument has professional intermediation, defined tenure, an explicit private capital multiplier, and a written exit.

Neither Washington nor Beijing has assembled that particular combination. One moved to ownership without the intermediation. The other built the intermediation at enormous scale and struggled to keep the decisions commercial. India has spent a decade getting the intermediation right and is only now, cautiously and on pari passu terms, adding the ownership piece.

The step into equity is a genuine increase in risk, and it deserves to be watched with the same seriousness that went into designing it. But it is being taken with unusual care about terms, on the back of an architecture that has already demonstrated it can hold public capital at arm's-length from political selection. If that instinct holds, and if the demand-side lever stays the primary de-risker rather than an afterthought, India will not have to choose between the American model of state capitalism and the Chinese one. It may instead be building a third model: one in which the state takes more risk without assuming that it must also make more investment decisions. Whether that distinction survives the move onto the cap table is now the experiment worth watching.