Deep tech startup investment in India: checking the science

Deep tech asks you to judge science you can't fully check yourself. The work is in finding who can, and in asking them the right questions.

A founder shows you a new battery chemistry, a satellite sensor or a protein built in a lab in Hyderabad. The deck says it works. You've got a finance degree, maybe an engineering one from twenty years ago. So what do you actually check? Start with the six checks in our guide to evaluating early-stage startups; founder, cap table and legal work the same way here. The science needs a different method, and India's own definition of deep tech gives you one.

What makes a startup deep tech in India?

DPIIT's notification G.S.R. 108(E), published on 4 February 2026, defines a deep tech startup by 4 attributes: a solution built on new scientific or engineering knowledge, a high share of spending on R&D, significant novel IP owned or in progress, and long timelines with high capital needs and large technical uncertainty. Each attribute is a diligence question. You'll use all four.

The Gazette notification also gives deep tech companies more time. A recognised deep tech startup keeps its status for up to 20 years from incorporation, with a turnover limit of ₹300 crore. Other startups get 10 years and ₹200 crore. The government is telling you, in writing, that these companies take longer. Your diligence should plan for it.

How do DPIIT's deep tech attributes become diligence questions?

Use the four DPIIT attributes as your diligence frame. Ask what new knowledge the product rests on and who else has it, what share of spending goes to R&D, what the company owns in patents or filings, and what has to go right, in what order, before anyone pays. A founder who can answer all four plainly understands their own risk.

1. New knowledge

Ask the founder to explain, in 2 sentences, what they know that others don't. Then ask where it came from. A doctoral project? A university lab? A previous employer? Whoever owned the lab may own the knowledge. If the work started at an IIT or a CSIR lab, read the technology transfer or licence agreement before anything else.

2. R&D spend

Pull R&D out of the accounts as its own line. Don't accept a percentage on a slide. Salaries of scientists, lab equipment, NABL-accredited testing, prototypes, cloud compute for model training. Compare it with the founder's claims. Then open the audited accounts filed with the MCA. A company that calls itself deep tech but spends most of its money on sales? It's probably something else, and that's fine, but you'd price it differently.

3. Intellectual property

Search the Indian Patent Office records on InPASS yourself, and the PCT filings on WIPO Patentscope too. Check who the applicant is. It should be the company, not a founder or a university. Then check what's actually been granted and what's merely filed. They aren't the same thing. A filed application is a claim. A granted patent is a right, and only in the countries where it's granted.

4. Timelines and uncertainty

Ask for the technical milestones between today and the first paying customer. Lab prototype. Field pilot. BIS or CDSCO approval, where it applies. First purchase order. Put a cost against each one. Then ask which milestone the founder is least sure about. If the answer is "none", you've learned something important about the founder.

Who can check the science for you?

You don't need to understand the science. You need someone independent who does, and a way to read their verdict. Good sources are a technical reviewer with no stake in the company, the patent record, earlier grant reviews and raw test data. Combined, they tell you whether the science is real, owned and close enough to a product.

Start with a paid technical review. Pick a domain expert from an IISc, IIT or CSIR lab, or a retired DRDO or ISRO scientist, who's got no equity, no advisory role and no past with the founders. Send them the technical documents, the patent filings and your questions under an NDA. Ask for written answers. Then sit in on their call with the founder. You'll learn as much from how the founder answers hard questions as from the reviewer's report.

Grants help too. A startup that's won a BIRAC BIG grant, a DST NIDHI award or support from another public body has had its science read by a review committee. Ask for the sanction letter. That's no proof the product will sell. It's evidence someone qualified looked.

Research on venture capital backs this up, with a caveat. In the survey of 885 venture capitalists by Gompers, Gornall, Kaplan and Strebulaev, 47% named the team as the single most important factor, but healthcare investors were the one group that didn't overwhelmingly put the team first. Those are mostly American investors. The pattern still fits: where the science is hard, investors weigh the technology itself more.

You don't have to understand the science. You have to find someone who does and has nothing to gain from your yes.

How long do deep tech timelines run?

Deep tech companies often need several rounds before revenue. So your diligence includes a question about you. Can you follow on? Who else will? Map the money each milestone needs against the investors likely to fund it, including government-backed funds with long lives.

DPIIT's operational guidelines for the ₹10,000 crore Startup India Fund of Funds 2.0 allow AIFs that back deep tech startups a life of up to 18 years, against up to 12 for sector-agnostic AIFs. Contribution to a single deep tech AIF can reach 40% of its corpus, subject to ₹500 crore, and those AIFs must invest 1.5X the amount committed to them. Our explainer on Fund of Funds 2.0 covers how that money reaches startups.

What goes on a deep tech diligence checklist?

• The two-sentence version of the new knowledge, and where it came from.
• Licence or assignment agreements with any university, lab or past employer.
• R&D spending as its own line in the accounts.
• Patent filings and grants, with the applicant's name, by country.
• Technical milestones to first revenue, with the money each one needs.
• An independent technical review you commissioned and paid for.

What should you do before the second meeting?

Pick your reviewer before your second meeting. It's the single step that turns a pitch you can't judge into one you can.

Deep tech claims can run ahead of the evidence for years, so two checks matter early: a founder background check and an honest test of market size claims.

Questions

What is a deep tech startup under DPIIT rules?

Under DPIIT's notification G.S.R. 108(E) of 4 February 2026, a deep tech startup works on a solution from new scientific or engineering knowledge, spends a high share of revenue or funding on R&D, owns or is creating significant novel IP, and faces long timelines, high capital needs and large technical uncertainty.

How long can a deep tech startup keep DPIIT recognition?

Up to 20 years from incorporation or registration, with a turnover limit of ₹300 crore in any previous year, under the February 2026 notification. Other startups keep recognition for up to 10 years with a turnover limit of ₹200 crore.

How do investors check deep tech science they cannot judge?

Find people who can. Commission an independent technical review from someone with no stake in the company, search patent records yourself, read earlier grant reviews, and ask for raw test data rather than summary charts. Then judge the founder's answers to the reviewer's questions.

Does Fund of Funds 2.0 treat deep tech differently?

Yes. DPIIT's operational guidelines for the ₹10,000 crore Startup India Fund of Funds 2.0 allow deep tech AIFs a life of up to 18 years, against up to 12 for sector-agnostic AIFs, and cap contribution at 40% of the AIF's corpus subject to ₹500 crore.

Sources: DPIIT, Gazette notification G.S.R. 108(E) on recognition of startups and deep tech startups, 4 Feb 2026 · DPIIT, Operational guidelines for Startup India Fund of Funds 2.0, Apr 2026 · Gompers, Gornall, Kaplan and Strebulaev, How Do Venture Capitalists Make Decisions?, NBER Working Paper 22587, 2016

For more on how startup funding works in India, read the MyStartupValley blog or visit mystartupvalley.com.