How the government can help deep-tech startups reach their full potential

For a decade, the binding constraint on deep-tech in India has been patient capital. Science-based ventures take years to gestate before anything can be sold, and venture funds built around business models such as e-commerce, consumer products and similar tech found them too risky.

The government has addressed this directly. The ₹1 trillion Research, Development and Innovation Fund (RDIF) is one of the most transformative initiatives of the government. RDIF is expected to catalyse private sector investments in deep-tech.

In February this year, the department for promotion of industry and internal trade gave deep-tech startups a category of their own, with a recognition window of 20 years. The capital problem is getting solved.

What is needed next now that the RDIF is in place to solve the capital problem? Capital is only one piece of the puzzle that must be solved for the ecosystem to thrive and sustain. Addressing the patient capital issue produces a predictable consequence. Capital produces products. But it does not produce markets.

In the coming years, India will likely have a cohort of deep-tech firms with working prototypes in sensors, materials, diagnostics, propulsion and chips. Private capital at the growth stage will not scale a company whose product no one has bought.

The India Venture Capital Report 2026 by Bain and Company shows seed deals rising from 756 in 2024 to 803 in 2025 while deals at the Series D funding stage and beyond fell. Once a market exists, private capital automatically follows it.

There are several startups that have developed cutting-edge innovative products and technologies across sectors such as defence, railways, logistics, telecom, health and energy.

Many of these startups face a strange procurement paradox. Their products work in real life conditions. Agencies have tested the products. It is not that the customer does not like the product or does not want the technology. But often the officer is not sure how to manoeuvre the procurement through the existing rules.

How can the government help? Deep-tech products must eventually compete globally on performance and price, and a protected buyer teaches nobody to compete.

This view may indeed be right about the destination but may not be correct about the road.

In railways, health and defence, the government is the market. No private buyer will test a product such as a track sensor before Indian Railways does. The first order then serves to validate the product under real operating conditions. It lets the firm go abroad with a reference customer. Price falls with scale, and scale begins with one order. The state’s role becomes one of a facilitator and not patron.

This is how the US did it. In 1962, every integrated circuit sold in the country went to the government. David Mowery’s paper for the National Bureau of Economic Research records that share falling to 55% by 1965 and 10% by 1978. The government was the first customer, but never the only or even a permanent one.

The General Financial Rules (GFR) that govern Indian procurement were written for goods whose specifications and standards are known, and for those goods they work well. But a deep-tech product has no standard to rest on and can prove itself only by being tried. Grafting exceptions into existing rules will not do.

Innovation procurement needs a separate enabling chapter in the GFR, with its own definitions, its own sequence and its own accountability so that an agency buying something new is not judged by the rules written for buying something widely available and for standard products and technologies.

The character of that chapter matters more than its clauses. It should enable and not necessarily mandate. A mandate, whether as a quota, a set-aside, a reservation or other protection eventually ends up being anti-market. An enabling clause does the opposite.

Where a government agency judges that it can be an active buyer, where it faces a problem no catalogue product solves, it may choose this route. The choice stays with the buyer, and the competition stays among the sellers. It also enables agencies to try out innovative products from startups.

Within that chapter, a handful of principles could help. Buy outcomes, not specifications. The buyer agency should state the problem and how success will be measured and leave the solution to those who compete to supply it.

Proceed in two steps, a short paid pilot under real conditions and then an order for those who pass so that price is compared only among those who have shown they can deliver.

Separate failure of process from failure of physics so that an officer who followed due process answers for the process and not for the uncertainty of technology.

Let the inventor keep the invention since intellectual property that stays with the supplier turns one government order into a product for many markets. Pay on milestones, promptly.

Every instrument India has built since 2016 has lowered the cost of inventing. The innovation procurement intervention will create an enabling provision for the deep-tech startups to sell their products to the government more effortlessly and eventually contribute to making India an innovation hub.

The authors are, respectively, member, Niti Aayog, and associate director of strategy, FAST-India.