Roche Is Betting $1 Billion on a Startup That Barely Exists
Roche just signed a deal worth up to $1 billion with a startup that's barely a year out of stealth mode. The target: screening over 300,000 bispecific antibody combinations to find treatments that don't exist yet. It's a massive bet on discovery platforms over proven molecules.
Imagine walking into a restaurant, and instead of ordering off a menu, you ask the chef to cook every possible combination of 300,000 ingredients to find the perfect dish. That's basically what Roche just paid for.
The Swiss pharma giant signed a deal with Dualitas Therapeutics, a startup that emerged from stealth barely a year ago, to screen more than 300,000 bispecific antibody combinations in the hunt for new treatments for immune and inflammatory diseases. The deal could be worth up to $1 billion.
For a company that raised $65 million in its Series A, that's one hell of a first date.
What Roche Is Actually Buying
Let's start with the basics. Bispecific antibodies are lab-made proteins designed to grab onto two different targets at the same time. Think of a regular antibody as a wrench that fits one bolt. A bispecific is a wrench that fits two bolts simultaneously, letting you do things a single-target antibody simply can't.
The problem? Figuring out which pair of targets actually works is brutally hard. There are hundreds of thousands of possible combinations on the surface of human cells, and most of them do nothing useful. It's like trying every two-ingredient cocktail at a bar with 600 bottles behind it. Most of those drinks will taste terrible. A few might be transcendent.
That's where Dualitas comes in. The company built a proprietary platform called DualScreen, which can systematically test massive numbers of bispecific combinations for real biological activity. Instead of guessing which pairs might work based on existing science, DualScreen lets you just… try them all and see what happens.
Roche is paying $36.5 million upfront for access to that platform, with the rest of the potential $1 billion tied to research, development, and commercial milestones, plus tiered royalties. If candidates come out of the screening, Roche takes over everything downstream: preclinical work, regulatory filings, manufacturing, and sales.
Get tomorrow's biotech intelligence before your competitors.
Join thousands of biotech professionals who start their day with our free, daily briefing.
A Startup's Coming-Out Party
Dualitas was founded in 2023 by Forbes Huang and Greg Lazar, and the company stayed in stealth mode until September 2025. It's based in South San Francisco (because of course it is) and has a small but growing pipeline. Lead programs include DTX-102 for rheumatoid arthritis and DTX-103 for allergic disease.
But this Roche deal isn't really about those programs. It's about the DualScreen platform itself. Roche looked at the entire bispecific antibody landscape and decided that the ability to generate and triage combinations at industrial scale was worth a billion-dollar bet. That's a fundamentally different kind of deal than licensing a single molecule.
Think of it this way: Roche didn't buy a painting. They bought the artist's studio, easels included.
This is also Dualitas's first major partnership, which makes the validation especially meaningful. Going from a $65 million Series A to a $1 billion deal structure with one of the world's largest pharma companies in roughly a year is the biotech equivalent of getting called up from the minors straight to the World Series.
Why Pharma Is Obsessed With Bispecifics Right Now
The Dualitas deal doesn't exist in a vacuum. Bispecific antibodies have become one of the hottest modalities in drug development, with hundreds of candidates in clinical trials.
The money pouring into the space is staggering. In 2025, BioNTech and Bristol Myers Squibb struck a deal for a PD-L1/VEGF-A bispecific worth up to $11.1 billion. Pfizer partnered with 3SBio on a similar bispecific for up to $6 billion. Merck, Novartis, Gilead, and Jazz Pharmaceuticals have all signed billion-dollar-plus bispecific collaborations in the past two years.
By those standards, Roche's $1 billion ceiling with Dualitas looks almost modest. But the structure is different in an important way. Most of those mega-deals were for specific, identified drug candidates, often already in clinical trials. Roche is paying for discovery capacity: the ability to find molecules that don't exist yet.
That shift matters. According to industry data, preclinical deals accounted for 75% of bispecific antibody transactions in 2025, above the five-year average of 70%. Pharma companies are moving upstream, betting on platforms and early-stage exploration rather than waiting to buy proven assets at a premium.
The Bigger Picture: Discovery as a Competitive Weapon
For decades, drug discovery followed a relatively linear path. You picked a target based on published research, built a molecule against it, and tested it. The bottleneck was usually clinical development (getting the drug through trials and past regulators).
Bispecific antibodies have shifted that bottleneck backward. When your drug needs to hit two targets, the number of possible combinations explodes. Traditional approaches, where scientists pick pairs based on educated guesses, can only explore a tiny fraction of the space. Platforms like DualScreen promise to blow that open.
It's a bit like the shift from hand-drawn animation to computer-generated imagery. The creative vision still matters, but the tools determine how much territory you can explore and how fast.
Platform collaborations like this one are a big reason why biopharma dealmaking has surged. The companies that can generate the best bispecific combinations fastest will have a structural advantage; everyone else will be licensing from them.
The Bear Case (Because There Always Is One)
Let's be honest about what $36.5 million upfront buys you: a lottery ticket. A very sophisticated, scientifically rigorous lottery ticket, but a lottery ticket nonetheless. Screening 300,000 combinations doesn't guarantee that any of them will become a drug. Most won't. Biology is messy, and what works in a screening assay often fails in animals, let alone humans.
The milestone-heavy structure reflects that reality. Roche is limiting its upfront risk while dangling nearly $1 billion in potential payouts if things go well. That's smart financial engineering, but it also means Dualitas has to actually deliver.
And while the DualScreen platform sounds impressive, it's still early. Dualitas's own pipeline programs haven't entered clinical trials yet. The platform has generated interesting preclinical candidates, but the ultimate test of any drug discovery engine is whether its molecules work in patients.
Why You Should Care
This deal is a signal flare for where antibody drug discovery is heading. The era of one-target-at-a-time thinking is giving way to combinatorial exploration at scale. AI-guided design, high-throughput screening, and data-rich platforms are becoming the new competitive moat in biopharma.
Roche clearly believes that whoever can explore the most biological space, fastest, wins. And they're willing to bet a billion dollars on a two-year-old startup to prove it.
Whether Dualitas delivers blockbuster drugs or expensive data remains to be seen. But if you're watching biotech, remember: the next wave of breakthrough medicines might not come from a brilliant scientist's "eureka" moment. They might come from a machine that tested 300,000 options before breakfast.
4 min read
BMS Just Paid $1.5 Billion for a Company With Zero Approved Drugs
Bristol Myers Squibb is dropping $1.5 billion in cash on Orbital Therapeutics, a preclinical-stage startup with zero approved drugs and zero revenue. The bet says everything about Big Pharma's platform fever and the desperate race to fill patent cliffs.
