Evolution's Unread Pharmacy: Why We Backed Venomics AI
Somewhere in the Western Ghats, a snake is hunting. When it strikes, it delivers a cocktail of molecules that evolution has been refining for millions of years, each one built to find a specific target in its prey and act on it almost instantly.
Compare that with how drugs usually get found. Teams spend years, and a lot of money, screening molecules that were never designed to do anything inside a living body. Most of them don't.
Venom is different. Every peptide in it has already been tested, under the harshest possible conditions, on whether it can find its target, work quickly and hold up long enough to matter. We've come to think of it as one of the most patient chemistry experiments on Earth, and the world has barely started reading the results.
That is why Speciale Invest is leading the $2.7M pre-seed round in Venomics AI. The company is building the World Venom Atlas, a proprietary map of global venom peptides and what they do in the body and using it to train AI that turns them into medicines.
Evolution's head start
A predator's survival depends on hitting the right target in its prey, fast, and nothing else. Spiders, scorpions, snakes and cone snails have spent hundreds of millions of years refining molecules to do exactly that. That is the brief for a good drug.
Most of those targets are ion channels, the gates that control electrical signaling in every cell and govern pain, heartbeat and insulin release. Small-molecule drugs bind deep inside these receptors, where subtypes look nearly identical, so they often hit the wrong one. Venom peptides sit on the outside, precisely where subtypes differ.
Blocking Nav1.7 silences pain; blocking its near-identical cousin Nav1.5 can stop the heart. Tarantula and centipede peptides have shown more than 1,000-fold selectivity between the two. Roughly a dozen approved drugs already come from venom, including captopril (Brazilian pit viper), exenatide, the first GLP-1 drug (Gila monster), and ziconotide (cone snail).
A library we've barely opened
An estimated 20 million venom peptides exist, yet fewer than 0.1% have been studied. All the world's public venom databases together hold somewhere around 40,000 to 60,000 entries, scattered across repositories that don't talk to each other.
Until recently, tiny samples and days of work per test made mapping impractical. Today, sequencing and mass spectrometry catalogue a venom from nanograms, and automated systems test hundreds of samples in parallel. The library was always there. We can finally read it with better tools.
Why India
India is one of 17 megadiverse countries, and its venomous animals live everywhere from desert scrub to rainforest to coral reef. Venom can even differ within the same species from one region to the next - a cobra in West Bengal carries different toxins from one in Rajasthan.
Access is tightly governed. Under the amended Biological Diversity Act, foreign-controlled entities need prior approval to use Indian biological resources, including isolated peptides. Add the permits and field networks that take years to build, and this is a resource even the best-funded drug companies cannot simply buy.
Venom is the data; AI is the reader
AI models are only as good as the data they learn from. And the data that matters most in drug discovery- whether a molecule does something in a living system is scarce. Most of it is missing, scattered across the literature, or locked inside pharma's private libraries. Generating candidate molecules is getting cheap. Knowing which ones will work is not.
Venomics AI starts from the other end. Every venom it collects is sequenced, characterized and screened against human drug targets in its own wet lab. The result is the Venom Atlas: a proprietary, multi-omics map of Indian venom peptides, tied to experimentally measured activity.
Trained on that data, models built on open foundations such as ESM and AlphaFold learn to rank hits, flag toxicity and eventually redesign venom scaffolds to be safer and more selective. Each screen sharpens the models, and the models pick the next screen. Instead of asking AI to find a needle in a haystack of billions of theoretical molecules, Venomics AI hands it needles that evolution has already validated.
The team
Prof. Kartik Sunagar has spent nearly two decades studying venom. He runs the Evolutionary Venomics Lab at IISc, Bangalore and his group's work on next-generation antivenoms with Scripps Research and the Technical University of Denmark was published in Science Translational Medicine. Vinay Prabhu leads the computational side. He has a dual PhD from Carnegie Mellon and was previously Chief Scientist at UnifyID in Silicon Valley.
Plenty of startups in this space are strong in biology or strong in ML. It's much rarer to find both in the founding team, working on the same system from the start. That was a big part of why we wanted to back them.
What's next
This round will help the team grow the Venom Atlas as well as take the first steps toward lead programs. Honestly, nobody knows yet which medicines are hiding in these venoms. Maybe a gentler way to treat pain, maybe a new way to steady the heart. But we're convinced the search has never been better equipped, and that some of the most promising places to look are right here in India.
Venomics AI also fits how we think about AI more broadly at Speciale Invest. We are looking for companies whose models learn from proprietary data that is hard to gather and highly differentiated, in verticals where the opportunity is large. As models become more widely available, the data they learn from is increasingly what sets companies apart. The Venom Atlas is a dataset that takes years of fieldwork, permits, and experiments to assemble on a platform proven by evolution.
We're excited to be part of it from day one.
We are thankful to South Park Commons and Operators Studio for their shared conviction in Venomics AI.
Speciale Invest is an early-stage deep tech venture fund based in India, now deploying from its $70M Fund III. It invests in frontier technologies at the pre-seed and seed stages across aerospace, defence, space tech, semiconductors, robotics, quantum, biotech, advanced materials, energy and AI, amongst others.