India must build deep tech for the world, not just for itself

Deep tech founders building for India problems or gaps is a very narrow version of the opportunity in most areas. Our deep tech ambition must not be building for Indian customers.

Semiconductors, advanced materials, industrial robotics, aerospace components, defence hardware, scientific instruments - none of these businesses work at India-only scale. Their economics need global customers to work. So, the real question isn't whether India can build deep tech for itself. It's whether India can become the place where the world's frontier technology gets built, tested and scaled - even when India accounts for a fraction of its buyers.

We have already seen this playing out in payments - when NPCI launched UPI to fix a domestic problem: too much cash, too little interoperability between banks. A decade later, UPI handles roughly half the world's real-time digital payment transactions, according to the IMF. It's now live for payments or remittances in 11 countries, from Singapore to France.

UPI isn't deep tech. But it proves that India can solve a problem at home, at massive scale, and then export the solution to markets that never had that problem to begin with. The question is whether India can do this in semiconductors, materials, robotics, aerospace, scientific equipment. The early signs are positive – contrary to what many would assume.

Two things need to be true for India to build world-class deep tech products/platforms: enough people who know how to build it, and enough capital to fund the years of iteration it takes to get there.

Talent and capital are converging

For a long time, India's engineering prowess revolved around cost, not invention. That's changing. India now has more than 3,600 deep-tech startups, and the government expects that to be closer to 10,000 by 2030.

The quieter shift is who's building them. Engineers and scientists who trained at global chipmakers, aerospace companies and research labs abroad are coming back to start or run their own companies. Global capability centres have become the training ground for this: three in four of the world's top 1,000 R&D centres already have a base in India, and they've moved well past support work into real engineering for aerospace, defence, semiconductors and manufacturing.

Capital is following. The government approved a ₹1 lakh crore research and innovation scheme in mid-2025, aimed squarely at AI, quantum, semiconductors, defence and space. That's a bet on original research, not just execution. India has seen record VC deployments in deep tech & AI already - ~$3 billion in 2025 alone.

India can be the world's deep-tech test bed

Building the technology is one problem. Proving it works is another - and this is where India has an advantage almost nobody talks about. India isn't just one market. It's several markets stacked on top of each other: dense cities next to villages with patchy infrastructure, extreme heat and extreme cold, some of the world's best private hospitals a few hundred kilometres from places with barely any healthcare, factories running on decades-old equipment next to ones running the newest robotics.

A product that works reliably across all of that - through heat, dust, unreliable power, patchy internet, users with wildly different comfort levels with technology - has been through truly rigorous testing.

Pharmaceuticals already proved this model works. India has more US FDA-compliant manufacturing plants than any country outside the US itself. Global pharma companies don't just manufacture there - they run large clinical trials there, using India's diversity as a testing ground before a drug reaches the rest of the world. That's the exact model deep tech needs: validate it in India, to a standard the world already trusts.

The raw testing environment already exists - the geography, the climate, the range of infrastructure and demand. What's missing is the formal machinery to turn that into something the world will accept: certification that carries weight internationally, procurement processes that function as real early revenue, and capital patient enough to survive the slow climb from prototype to a certified, sellable product.

Testing alone isn't validation. What actually proves a technology works is a buyer demanding it perform under real commercial pressure, not in a lab. India increasingly has these buyers too - across manufacturing, automotive, energy, healthcare and defence, companies that want globally competitive performance, not a version that's "good enough." That build-test-validate loop is exactly what took UPI from a domestic fix to infrastructure now running in a dozen countries.

None of this happens without capital built for this specific job, and Indian venture capital still has work to do here.

Fund the gap directly. The stretch between a working prototype and a certified, sellable product is where equity investors often get cautious. Indian VC already has investors built for 0-to-1 and investors built for 1-to-10 - deep tech doesn't sit neatly in that narrative. The product development cycles and milestones look different, and so does what counts as progress. Investors are solving for this by structuring milestone-based funding, combining equity with debt, or co-invest alongside government grants/incentives.

Another gap is the testing infrastructure itself, limited to select research institutions today. With certification delays being a real constraint, backing the private labs and testing facilities that many hardware startups will depend on is a critical need.

If engineers and scientists coming back from global chipmakers and aerospace companies really are a tailwind, funds that build relationships with this group directly, rather than waiting to be found, will see the best of it before everyone else does.

Talent and capital efficiency are already falling into place. Real-world conditions are harsh enough to stress-test what gets built. Buyers are demanding enough to validate it. What's still missing is patient money — capital willing to sit through the years between a working prototype and a paying customer, instead of pulling back at the first delay. That's the piece venture capital can help control, and right now, it's the piece missing from the lab.

India doesn't need to be the end customer for every one of these technologies to matter. Semiconductors, materials, robotics, aerospace components and scientific instruments will always need global markets to make economic sense. What's still open is where they get built, tested and refined before they get there. India doesn't need to win every category. It needs to become the place every category has to pass through.

(The writer is Lead - Investments, EquirusInnovateX Fund. Views are personal.)