AI-designed antibodies land a $3.8 billion bet
Two years after it was founded, Chai Discovery has become one of the most richly valued startups in AI drug discovery — and, more unusually, one whose software is already inside the labs of the world's largest pharmaceutical companies. On July 14, the San Francisco company said it had raised $400 million in a Series C round led by Index Ventures, tripling its valuation to $3.8 billion just seven months after it was worth $1.3 billion.
The round, which the company described as heavily oversubscribed, brings Chai's total funding to roughly $630 million since its 2024 launch. Kleiner Perkins, Sequoia Capital and Dimension joined as new backers, alongside a roster of returning investors that reads like a who's-who of Silicon Valley capital: Bain Capital Ventures, Thrive Capital, Menlo Ventures, General Catalyst and OpenAI, whose participation underscores how tightly the frontier-AI and drug-discovery worlds have converged.
What separates Chai from the crowded field of AI biotech startups is not the size of the check but where its models have ended up. Chai's software is already being used by Pfizer, Novartis and Eli Lilly — three of the largest drugmakers on earth — moving the company past the demo stage that has trapped many of its peers.
From promise to deployment
"AI drug discovery has moved from promise to deployment, and Chai's models are already unlocking progress for our partners — enabling them to design better molecules, move faster against difficult targets, and take on challenges that traditional discovery methods have struggled to solve," said Joshua Meier, Chai's co-founder and chief executive.
Meier, who founded the company with Jack Dent, Matthew McPartlon and Jacques Boitre, framed the raise as fuel for a longer-term ambition. "Tomorrow's medicines should be designed with the precision, speed and scale of modern engineering, and this support helps us move faster towards that future," he said.
At the center of the pitch is Chai-3, the company's latest model. Chai says it materially improves target success rates and binding affinity over its predecessor, Chai-2 — which in 2025 became the first zero-shot generative platform for fully de novo antibody design to reach double-digit experimental success rates. According to the company, Chai-3 roughly doubles the success rate of that earlier model and can produce antibodies that bind up to 100 times more tightly to their intended targets. In practical terms, that is the difference between an antibody that theoretically works and one a drug company would actually advance toward the clinic.
The pharma partnerships that matter
Chai's commercial traction has arrived in a rapid sequence. In January 2026 it announced a research collaboration with Eli Lilly that included a bespoke AI model trained on Lilly's proprietary data. In June, Pfizer signed a licensing agreement granting it early access to Chai-3 plus a model trained on Pfizer's own data. And in the run-up to the funding, Novartis said it would use Chai's models, including Chai-3, to discover antibodies across multiple therapeutic programs.
For an incumbent drugmaker, the appeal is straightforward: antibody discovery has historically meant screening enormous physical libraries and grinding through wet-lab iterations. A generative model that proposes viable candidates in silico compresses that timeline and widens the range of targets worth attempting.
"AI is already reshaping how we discover medicines," said Fiona Marshall, president of biomedical research at Novartis. "By applying advanced models to challenges like antibody design, we can explore a broader range of possibilities, make better decisions earlier, and focus our efforts on the most promising candidates."
That endorsement from a big-pharma research chief is the kind of validation venture investors have been waiting for across a sector that has attracted billions but produced few approved drugs.
Why this matters
The AI-for-drug-discovery boom has been long on funding and short on proof. Chai's raise is notable because its models are not merely promising in benchmarks — they are being licensed and deployed by Pfizer, Novartis and Lilly, companies with the resources to test whether AI-designed molecules survive contact with real biology. A tripling of valuation in seven months signals that investors believe the deployment story is real, not that they are chasing another speculative platform.
It also reflects the deepening entanglement of the AI industry with pharma. OpenAI's presence on Chai's cap table, next to Sequoia and Kleiner Perkins, shows frontier-model money flowing into biology as a proving ground for generative design beyond text and images. If antibodies can be engineered like software, the economics of early-stage drug R&D — long the industry's most expensive and failure-prone phase — could shift meaningfully.
The caution is equally real. No Chai-designed molecule has yet been approved, and the gap between a tightly binding antibody on a screen and a safe, effective medicine in patients remains vast and littered with failures. Valuations in AI biotech have run well ahead of clinical evidence before.
What to watch
The next test is whether any of the Pfizer, Novartis or Lilly programs built on Chai's models advance into formal preclinical development or the clinic — the first true measure of whether AI-designed antibodies translate into drugs. Watch, too, for how Chai deploys $630 million in cumulative capital: whether it stays a model-and-software supplier to pharma, or begins building its own therapeutic pipeline, a move that would put it in direct competition with the very partners now funding its ascent.
"AI drug discovery has moved from promise to deployment, and Chai's models are already unlocking progress for our partners, enabling them to design better molecules and take on challenges traditional discovery has struggled to solve."-- Joshua Meier, Co-founder and CEO, Chai Discovery