Chai Discovery AI Drug Discovery: Timeline, Partners and Results

Chai Discovery builds AI models that predict and design molecules, starting with Chai-1 and then Chai-2 for antibody design. OpenAI backed its first funding round, and Eli Lilly, Novartis and Pfizer now use its models.

Presented By Our Partners

What Is Chai Discovery?

Chai Discovery is an AI company that builds foundation models for molecules: software that predicts how proteins, drugs and other molecules fit together and, in its newer models, designs new ones. It came out of stealth in September 2024, about six months after it was founded, with a seed round backed by OpenAI and Thrive Capital, as Bloomberg via Daily Maverick reported. TechCrunch reported that the company started out in OpenAI’s offices.

Its first model, Chai-1, predicts the 3D structure of molecular complexes, the same job as Google DeepMind’s AlphaFold 3. By mid-2025, unrelated research groups were already using Chai-1 in published work. Its second model, Chai-2, announced on June 30, 2025, designs antibodies from scratch for a chosen target, with no starting antibody to copy. Coverage of Chai’s technical report put its hit rate at roughly 16 to 20 percent of designs tested, a company-reported figure from a preprint that has not been peer reviewed.

Chai’s public business is giving drugmakers access to its models; it has not disclosed a drug pipeline of its own. Fierce Biotech reports that Eli Lilly, Novartis and Pfizer use its models, Chai announced a collaboration with Lilly in January 2026 to train custom models for biologics design, and it signed a collaboration agreement with argenx in immunology. After its seed, Chai raised a $70 million Series A led by Menlo Ventures in August 2025, a $130 million Series B at a $1.3 billion valuation in December 2025 and a $400 million Series C at a $3.8 billion valuation in July 2026. Its valuation nearly tripled in seven months without a disclosed clinical asset. Funding and deployments show demand, not that the designed molecules work in people.

Chai Discovery Healthcare Timeline: Every Move Since 2024

Entries are newest first.

Chai Discovery’s Biggest Results, and How Far Each Has Been Proven

Chai’s strongest evidence is outside use of Chai-1; its headline antibody result is still company-reported.

Chai-2 antibody design. In June 2025, Chai reported that Chai-2 designed antibodies from scratch for chosen targets, with coverage putting the hit rate at about 16 to 20 percent of designs tested. Evidence level: Company-reported (preprint, not yet peer reviewed). Stage: Molecule design. What would prove it next: peer-reviewed publication and an outside lab reproducing binders from Chai-2 designs.

Chai-1 in outside hands. Unrelated groups have run Chai-1 and published the results. A University of Texas Southwestern team used Chai-1 with AlphaFold 3 and RoseTTAFold All-Atom to model 14,178 drug-bound human proteins carrying chemical modifications, and concluded that the methods can predict these effects but need a larger benchmark to judge accuracy (PMID 40320551). Evidence level: Peer-reviewed. Stage: Research tool.

Featured Partner

Invest in the Infrastructure Behind Modern Medicine

As healthcare expands beyond hospital walls, the buildings and campuses supporting that shift are generating compelling returns for investors who move early. The Healthcare Real Estate Fund offers qualified investors direct access to a curated portfolio of medical office, outpatient, and specialty care facilities.

Learn More →

Chai-1 on antibodies: a mixed independent test. A Johns Hopkins benchmark in mAbs found that with one seed, Chai-1 achieved 0% high-accuracy docking for antibodies and 3.33% for nanobodies, against 10.2% and 13.3% for AlphaFold 3. With twenty seeds, Chai-1 predicted the hardest antibody loop (CDR H3) to a median of 2.71 Å, better than AlphaFold 3’s 2.9 Å, but did worse on nanobodies (PMID 40814020). Evidence level: Peer-reviewed. Stage: Research tool. This is the kind of outside check every model on this tracker should face.

Pharma deployments. Lilly, Novartis and Pfizer use Chai’s models, according to Fierce Biotech, and argenx signed a collaboration agreement (Chai release). A deployment is a deal, not a result. What would prove value: a partner naming a development candidate designed with Chai, then a clinical trial.

Chai scientist Zhuoran Qiao also wrote a 2025 Science perspective on how structural foundation models can help explain and reprogram molecular biology (PMID 40674480). It is a viewpoint article, not new data.

Who Chai Discovery Works With

OpenAI is Chai’s only tier 1 tie: it backed the September 2024 seed round alongside Thrive Capital, and coverage of later rounds still describes Chai as OpenAI-backed. Menlo Ventures led the Series A. We found no funding or partnership with Anthropic, Google, Microsoft or Meta.

On the pharma side, Chai announced a collaboration with Eli Lilly in January 2026 to train custom models for biologics design, signed a collaboration agreement with argenx in immunology, and counts Novartis and Pfizer among the companies using its models. Chai trains its own models; it has not said publicly whether it runs its own wet lab or relies on contract labs to test designs.

What to Watch Next

• Peer-reviewed publication of the Chai-2 antibody design results.
• An outside lab reproducing Chai-2 binders on new targets.
• The first development candidate a partner credits to Chai’s models.
• A first clinical trial for a molecule designed with Chai.
• New independent benchmarks of Chai models against AlphaFold 3 and other open models.

Related: our AI healthcare breakthroughs tracker, the OpenAI and ChatGPT healthcare timeline, the AI lab biotech deal tracker, how Isomorphic Labs approaches drug discovery and our HYPE Index methodology.

What This Means for You

If you are a patient, Chai has no drug in a disclosed trial, so it does not change your treatment today. Antibody drugs designed with tools like Chai-2 would still need years of testing before reaching a pharmacy.

If you work in research, Chai-1 is a model outside groups already use and test, which lets you judge it on your own targets. Read the independent benchmarks before relying on it for antibody docking, and wait for peer review of Chai-2’s hit rates before treating them as settled.

Compiled by Healthcare Discovery from company announcements, named news reports and PubMed. Every PMID on this page was checked against PubMed on October 28, 2026. Evidence levels follow the scale explained in our AI healthcare breakthroughs tracker. Healthcare Discovery uses AI research tools from several companies on this page, including OpenAI, Anthropic and Google models; no company listed reviewed or paid for this page.

Sources

1. Hitawala FN, et al. What does AlphaFold3 learn about antibody and nanobody docking, and what remains unsolved? mAbs, 2025. PMID 40814020
2. Medvedev KE, et al. Leveraging AI to explore structural contexts of post-translational modifications in drug binding. Journal of Cheminformatics, 2025. PMID 40320551
3. Qiao Z. AI to rewire life’s interactome: Structural foundation models help to elucidate and reprogram molecular biology. Science, 2025. PMID 40674480
4. The Pharma Letter. Chai Discovery raises $400 million to advance AI-designed drug discovery, July 2026. The Pharma Letter
5. Dealroom. Chai Discovery raises $400M Series C at $3.8B valuation for AI drug discovery, July 2026. Dealroom
6. Business Wire. Chai Discovery Announces $400M Series C to Advance AI-Driven Molecular Design, July 2026. Business Wire
7. Fierce Biotech. Chai brews up $400M series C to fuel AI molecule models used by Lilly, Novartis and Pfizer, July 2026. Fierce Biotech
8. Chai Discovery via BioSpace. Chai Discovery Enters Into Collaboration Agreement With argenx to Advance AI-Driven Immunology Breakthroughs, 2026. BioSpace
9. TechCrunch. From OpenAI’s offices to a deal with Eli Lilly: how Chai Discovery became one of the flashiest names in AI drug development, January 2026. TechCrunch
10. Business Wire. Chai Discovery Announces Collaboration with Eli Lilly and Company to Accelerate Biologics Discovery, January 2026. Business Wire
11. BioPharmaTrend. AI unicorn Chai Discovery and Eli Lilly to train custom AI models for biologics design, January 2026. BioPharmaTrend
12. TechCrunch. OpenAI-backed biotech firm Chai Discovery raises $130M Series B at $1.3B valuation, December 2025. TechCrunch
13. PharmExec. Menlo Ventures led $70 million Series A fundraiser for Chai Discovery to transform molecular design, August 2025. PharmExec
14. Business Wire. Chai Discovery Unveils Chai-2 Breakthrough Achieving Fully De Novo Antibody Design With AI, June 2025. Business Wire
15. Chai Discovery. Introducing Chai-2, June 2025. Chai Discovery
16. BioPharmaTrend. Chai Discovery releases foundation model for zero-shot antibody design with 16 to 20% hit rates, 2025. BioPharmaTrend
17. bioRxiv. Zero-shot antibody design in a 24-well plate, July 2025. bioRxiv
18. Bloomberg via Daily Maverick. OpenAI, Thrive Capital back six-month-old AI drug discovery startup, September 2024. Daily Maverick
19. ReadWrite. OpenAI funds drug discovery startup with almost $30 million raised, September 2024. ReadWrite
20. Chai Discovery. Seed announcement, September 2024. Chai Discovery

Free Daily Briefing

The Latest Longevity Science.
Delivered Every Morning.

Join researchers, physicians, and health professionals getting daily breakthroughs in AI-driven medicine, epigenetics, and longevity research.

Support the research that powers this editorial

Subscribe Free

No spam. Unsubscribe anytime. We respect your inbox.

Keep Reading on Healthcare Discovery

• Sleep Regularity Beats Duration: Why When You Sleep May Matter More Than How Long
• The CBT-I Question: What 40 Years of Insomnia Research Reveal About the First Line Treatment Doctors Still Underprescribe
• Young Blood Revisited: Inside the 2026 Trials Testing Plasma Exchange and Rejuvenation Factors