Modelling plus lab testing to improve industrial enzymes for Indian makers
TL;DR Foldwright sells fixed-scope projects that make a client's industrial enzyme work better in its own process. Software shortlists candidates, a partner laboratory measures them and the measurements steer the next shortlist. Indian enzyme producers and biomanufacturing firms pay per project, with part of the fee tied to hitting the target.
India's bioeconomy stood at roughly US$150 billion (about ₹14.5 lakh Cr) in 2023, and bio-industrial work made up US$72.6 billion of that[1]. Detergent, food, textile and feed producers all depend on enzymes bought or made locally. Getting one to perform better in a particular process has usually meant many months of trial batches in a well-equipped lab.
Published work shows software can narrow that search. One 2025 paper reported that a model trained on measurements from 1,217 candidates picked versions performing 1.6 to 42 times better than the starting point[2]. But mid-size Indian producers rarely employ the data scientists needed to work this way, and the platforms that do it are built and priced for multinationals.
In regulatory terms, this is a contract research service, not a manufacturer. Any laboratory work involving engineered micro-organisms falls under India's Rules, 1989, so the partner laboratory must hold the required biosafety clearances before a project starts[3].
Each engagement starts with the client's existing material, the property it wants improved and whatever assay results it already holds. Foldwright's software ranks a modest shortlist for each round. The partner lab measures that shortlist in the client's own assay conditions, and those results feed the next round. Two to four rounds is the plan. The client owns the outcome, while Foldwright keeps the right to train its models on anonymised measurements.
The seed idea was a structure prediction platform. That market more or less closed in July 2022, when a public database put over 200 million predicted structures online for free[4]. Predicted structure is now an input anyone can get. What a client still pays for is a measured improvement in its own process.
The day-to-day user is the client's R&D scientist. The payer is the R&D head or founder, out of a development budget. Likely buyers include established producers such as one listed Indian enzyme maker, which serves more than 25 industries[5], newer biomanufacturing companies and contract manufacturers. There's no public price for this kind of project work, so ₹40–80 L per engagement plus a target-linked fee is our assumption to test.
One European protein-design startup raised US$73 million (about ₹706 Cr) in November 2024 for software of this kind, listing large enzyme and pharma makers among its customers[6]. So large buyers clearly pay. In India the BioE3 policy is funding biomanufacturing centres expected to back more than 250 startups and smaller firms[7], and those firms will need outside design help rather than their own teams.
No published figure covers spending on this service in India. The firmest adjacent number is the US$72.6 billion bio-industrial segment (roughly ₹7 lakh Cr)[1], which is far broader than your market. A closer signal is customer counts - that European startup reports over 21 paying customers worldwide[6]. Count the Indian firms with both the need and the budget before you raise a rupee.
The European startup sells a self-service platform to scientists[6]. A US enzyme-design firm works through commercial partners and says one programme with a consumer-goods major ran 18 months[8]. Large Indian producers keep this work in-house. The opening is a managed service, quoted in rupees, with testing done in India so samples and results stay close to the client.
| Difficulty | Hard (56/100) |
| Starting capital | ₹2.5–4.5 Cr |
| Time to launch | 9–15 months |
Much of the software is public. The hard bit is turnaround, because each round needs a lab that reports clean measurements in weeks and not months, and that usually means a signed deal with a contract lab plus your own quality checks on its data. Selling is slow too, since a client is handing over material it treats as confidential.
The design service needs no licence of its own, but lab work with engineered organisms is tightly regulated.
Building this outside India? The categories above usually have a local equivalent, so check with your own regulator.
Don't trust a model's ranking as a result. Only a measured improvement in the client's own assay counts.
Want to work out the AI layer for Foldwright? Talk to Yash →
I'd back a founder with a biochemistry background and real project-delivery discipline. If you're a pure software team then skip it, because the lab half decides whether clients come back. Prove one thing first - a measured improvement for one paying client inside four months, documented well enough that their process engineer accepts it.
Plan for roughly ₹2.5–4.5 Cr, mostly on scientists, partner-lab testing and a long first sales cycle.
The design service needs none, but lab work with engineered organisms falls under the Rules, 1989. Your partner lab must have its biosafety approvals in place.
Yes, enzyme makers everywhere want better performers. Check your local biosafety rules for engineered organisms and the permits for shipping samples.
The free Foldwright validation kit has the line-by-line capital breakdown, the effort scoring, the biosafety and contract checklist, a 14-day validation sprint with pass/fail targets, a 90-day launch plan, 10 interview questions for R&D heads, and 5 assumptions to test. Download the kit →
Thinking about building Foldwright or something close to it? The make-or-break work is a laboratory partner that returns trustworthy measurements fast enough to keep projects on schedule, and I can help you plan for that before you raise. If you need a hand, you know where to find me. Talk to Yash →
For illustration only. Names are fictional and ideas hypothetical; sourced facts are real but may be out of date. Nothing here is legal, financial or professional advice, and Yash Khatri is not personally liable for any use of it. Read the full disclaimer.