The hunt for the “Figma of Biotech” has a new frontrunner.
Chai Discovery, a San Francisco-based startup aiming to turn drug discovery into a software engineering discipline, has raised $130 million in Series B funding. The round values the company at $1.3 billion, securing its spot as the industry’s latest unicorn.
General Catalyst and Oak HC/FT co-led the deal. They were joined by a heavy-hitting roster of returning backers, including OpenAI, Thrive Capital, and Menlo Ventures, alongside newcomers Emerson Collective and Glade Brook.
Moving biology from the bench to the screen
The pitch is distinctively Silicon Valley: Biology is just code that hasn’t been debugged yet. While most AI-biotech startups spend millions building massive wet labs to validate their models, Chai is taking a different approach. They are betting on a “pure play” software model.
The founding team—CEO Joshua Meier, Jack Dent, Matthew McPartlon, and Jacques Boitreaud—started with the premise that drug discovery remains too artisanal.
“We realized that AI could generate molecular structures no human would even think to sketch,”
Meier has told investors. The goal isn’t to be a pharmaceutical giant, but to be the architectural software—the CAD suite—that pharmaceutical giants rely on.
By separating the computational design from the physical validation, Chai avoids the capital-intensive trap that sinks many biotech upstarts. Instead of building labs, they build APIs that connect to contract research organizations (CROs) and automation providers. You design it in the cloud; the API triggers the synthesis elsewhere.
The Stripe and OpenAI pedigree
The founding team represents a convergence of large language model (LLM) research and infrastructure reliability.
Meier brings experience from Facebook AI and OpenAI, where he worked on the scaling laws that define modern AI. Co-founder Jack Dent comes from Stripe, bringing an infrastructure mindset to a field often plagued by one-off experiments. They are joined by McPartlon and Boitreaud, who provide the necessary computational chemistry expertise.
Investors view this lack of traditional “lab coat” DNA as an asset.
“Chai looks less like a traditional biotech and more like an AI-native design company that happens to operate in biology,”
noted one investor close to the deal.
The economics of design
Chai’s platform allows researchers to input biological targets and constraints—manufacturability, binding pockets, ADMET profiles—and receive fully generated molecular candidates. It doesn’t just search existing databases; it hallucinates new viable structures.
The business model reflects this software-first approach. Rather than owning the entire drug pipeline and its associated binary risks, Chai operates on a licensing and royalty structure:
- Platform fees reported around $2–3 million annually.
- Royalties on successful programs ranging from 5–10%.
This creates a recurring revenue stream that looks more like enterprise SaaS than a volatile biotech stock. The company is initially targeting large pharmaceutical companies working on “undruggable” targets—proteins that have historically resisted small molecule intervention.
Validation in the wild
This raise comes shortly after a $70 million Series A in August 2025. The rapid valuation step-up suggests the technology is working in practice, not just in theory.
According to data shared with investors, early pilots with top-tier pharma partners demonstrated a 40% acceleration in lead optimization cycles. More significantly, the company claims a 70% reduction in failed compounds during that phase. Their models reportedly achieved binding affinity prediction accuracy near 85%, a significant jump over the 60–65% standard in conventional methods.
With over $200 million in total funding and a burn rate kept under $40 million annually, the company has secured a runway extending beyond two and a half years.
To shore up its industry credibility, Chai has added Mikael Dolsten, the former CSO of Pfizer, to its board. His presence signals a serious intent to navigate the regulatory and IP frameworks that will inevitably surround AI-generated therapeutics.
