Metanova Labs / Nova
Metanova Labs runs Nova, the world's first decentralised drug-discovery engine. Instead of one pharma giant grinding through molecules behind closed doors, a global network of AI developers and scientists competes around the clock to screen billions of compounds against disease targets - with the best candidates rewarded in tokens and the most promising ones taken straight into wet-lab testing. Think of it as open-source drug hunting at a fraction of the cost of Big Pharma.
The Thesis, One Year On
Just over a year ago Metanova Labs was born - the team behind Nova, Subnet 68 on Bittensor - on a simple thesis: if you could turn drug discovery from a closed, $2 billion-per-molecule guessing game into an open, globally crowdsourced competition, you'd end up with something that looks very different from pharma as we know it. A year on, the proof-of-concept has quietly turned into a platform, and CEO Micaela Bazo's recent Hash Rate interview is a good moment to take stock.
Why Drug Discovery Is Stuck
Traditional drug discovery is often described as searching for a needle in a stadium-sized haystack, blindfolded, on a $2 billion budget per approved drug. Ten years from lab to pharmacy shelf. More than 90% of candidates fail somewhere in the pipeline. Micaela puts the scale of the dysfunction bluntly:
But the economics alone don't capture why the industry is creatively broken. Perhaps 90% of what Big Pharma ships is built on existing scaffolding - tweaks to compounds with already-known safety and efficacy profiles. Shareholders get predictable revenues, regulators get familiar molecules, and R&D teams get to keep their jobs. What nobody gets are truly game-changing cures. The incentive structure of listed pharma rewards incremental, near-certain returns over breakthrough, binary-outcome bets.
Nowhere is this more obvious than in mental health - which happens to be Nova's initial area of focus, and not by accident. Mental illness is the largest and fastest-growing category of unmet medical need on the planet, and also the one where Big Pharma has almost completely given up.
That last number is the most damning. Prozac, approved in 1987, introduced the SSRI class. Every major antidepressant marketed since - Zoloft, Paxil, Lexapro, Cymbalta, Effexor - is a variation on the same serotonin-reuptake mechanism. Four decades. No fundamentally new target, no new pathway, no new idea. The same is true across depression, anxiety, ADHD, and Parkinson's: near-zero mechanistic progress despite an industry-wide annual R&D spend of roughly $120 billion. Meanwhile prevalence is rising, suicide remains the third leading cause of death in 15–29 year-olds, and only 2% of global health budgets touch mental health at all - a figure unchanged since 2017.
So when Metanova's own mission statement talks about "therapeutics that regulate mental states and restore critical biological processes" - targeting the neurosignaling circuits that shape reward, focus, appetite, sleep, and social connection - they are pointing directly at the biggest and most neglected opportunity in modern medicine. Nova exists to break the incentive lock that created this vacuum. And the prize for doing so, Bazo argues, is uncapped:
Decoupling the derisking from the capital structure of a listed pharma company is the move. When the global miner network bears the compute cost and only the best hits hit the wet lab, the arithmetic changes completely - and the targets nobody else will touch suddenly become viable.
And here is where the economics really compound. Virtual screening - the stage Nova dominates - is relatively cheap. The real money in drug discovery gets burned after that, in the wet lab: synthesis, assays, animal studies, early toxicology. A smarter front-end means every dollar spent downstream is spent on a molecule that already has a much higher probability of working. Better screening doesn't just save money at the top of the funnel - it compounds: fewer dead ends pursued, less wet-lab budget wasted, shorter cycles, and meaningfully more drugs potentially discovered in the same time and for the same capital. That is the real unlock.
What the Team Has Shipped
Since launch in March 2025, Nova has screened around 11 million molecules against nine different biological targets, has three live incentive mechanisms running (small-molecule discovery, chemical search algorithms, and a brand-new nanobodies competition where the top fifty submissions go directly into wet-lab validation), and pays roughly $2,000 a day in $NOVA alpha tokens to its global miner network. The database of synthesizable compounds has grown from one billion to sixty-five billion through combinatorial chemistry. Promising hits are being synthesised and tested through a Shanghai-based partner (Yalotain, founded by a Yale PhD), after which Metanova files provisional patents and decides whether to develop in-house, license, or co-develop. Mark Jeffrey, host of Hash Rate, put the monetisation model elegantly: it's a bit like a screenplay - you might produce it yourself, or you might sell the option to a studio that produces it for you. Either way, once you have the IP, you have something valuable.
That jump in search space, combined with the incentive structure, is what's delivering roughly 418% better hit quality over baseline methods. And here's the part that really vindicates the model: miners with no background in chemistry are beating published methods. The clearest example, in Bazo's own words:
An energy-grid optimisation algorithm, repurposed by an outsider, beating a peer-reviewed drug-discovery benchmark. That is the whole thesis of decentralised drug discovery working exactly as designed - fresh eyes, properly incentivised, finding what closed labs can't.
The conviction signal is also showing up in miner behaviour. A PhD scientist Metanova recently began working with was given the choice of being paid in fiat or in $NOVA tokens - and chose $NOVA. Miners historically churn straight out of subnet tokens; holding them is a very different kind of signal. Small data point, but it's the kind of thing that tells you the people closest to the platform are starting to believe what it could become.
Bitcoin Mining, But For Cures
The Metanova model strips the drug-discovery problem back to its essence - a pure search optimisation problem - and lets the global network's smartest participants (and increasingly, autonomous AI agents) compete to surface the best candidates. Think of it as Bitcoin mining, except instead of hashing blocks, miners are rewarded for finding promising drug molecules. A 24/7, back-to-back hackathon, as Bazo herself describes it. The company calls itself a virtual biotech: screen virtually, validate physically, patent, then partner or build.
Trading at 1/100th of Its Web2 Peers
This is where it gets interesting for us as investors. Nova currently sits at a network-level market cap of roughly $17 million* one year into live operation. Against the closest old-world comparators, the asymmetry is hard to miss.
Recursion Pharmaceuticals trades at about $2 billion, having burned several hundred million on proprietary compute. AbSci is around $400 million on the back of partnerships with AstraZeneca and Merck. Isomorphic Labs, DeepMind's drug-discovery spinout, is valued at $4 billion having raised just $500 million - essentially on a promise. Big Pharma itself spends $1–2.6 billion to bring a single drug to market and is happy to pay $10 billion to acquire a late-stage asset (Pfizer's recent GLP-1 move is the latest example).
So Nova is trading at roughly 1/100th to 1/400th the valuation of these peers while already running three live competitions, closing the wet-lab loop, and paying real miners real money every day. Even a single licensing deal on one preclinical asset would, on its own, be transformative. Bazo is direct about the typical ranges:
And unlike the Web2 comparators, Nova isn't betting on a single drug - it's building a platform that can continuously generate licensable assets. That's the part we keep coming back to. Which frames the asymmetry neatly, as Hash Rate host Mark Jeffrey put it to Bazo on the show:
The Forward Look
Three things on our radar for the next six months: the first wet-lab results making it back into the subnet as training signal (the real feedback loop is what compounds the edge); the first provisional patents filed and any licensing interest they attract; and whether the nanobodies competition produces a standout candidate.
Any of those milestones, on their own, meaningfully moves the conviction needle. Taken together, they are what turns Nova from a promising experiment into something the big biotech houses will have to reckon with. A word of caution: Nova is early stage and speculative and our rigorous risk control framework at Dragonfly means we only allocate a relatively small position. We'll keep you posted as results drop.