BigHat Biosciences Raises $75M for AI-Designed Biologics

Once a drug candidate enters a patient, the burden of proof around an AI platform changes. Model quality still matters, but safety, manufacturing, pharmacology, and clinical execution start deciding whether the design deserves to become medicine.

BigHat Biosciences has raised a $75M Series C to keep working on that gap. DFJ Growth and Premji Invest co-led the financing, with a broad group of new and returning investors participating. BigHat says the round brings its total capital raised to $223M.

The timing is more important than the number. BigHat has moved BHB810, an AI-designed antibody-drug conjugate, into a first-in-human study and is preparing a second internally developed program for the clinic. The company now has to show that its design-and-experiment loop can produce useful clinical evidence, not simply convincing laboratory output.

What Happened

BigHat's September 24 announcement identifies Catalio Capital Management, LG Technology Ventures, and Sigmas Group as participants alongside existing investors 8VC, Alexandria Venture Investments, Amgen Ventures, Andreessen Horowitz, Discovery Ventures, GRIDS Capital, Intermountain Ventures, Eli Lilly and Company, Merck Global Health Innovation Fund, Quadrille Capital, and Section 32. BigHat plans to use the financing to expand its protein-design platform and move BHB810 and BHB299 toward clinical readouts. The announcement does not disclose a valuation, individual check sizes, ownership terms, or board changes. The $223M total is company-reported, and historical descriptions of the 2022 Series B differ across BigHat's own announcements, so the latest total is more reliable than reconstructing every prior tranche from inconsistent labels.

The company was founded in 2019 by Peyton Greenside and Mark DePristo. Greenside is now CEO and co-founder, while DePristo serves as an adviser. BigHat's current leadership also includes Chief Scientific Officer Stefan Weigand and Chief Development Officer John Corbin; the company's official leadership page does not list a current CTO.

Why BigHat's Feedback Loop Matters

BigHat's platform connects computational design with automated biological testing. The company describes Milliner, Reccy, and RADS as its wet-lab, data-control, and AI/ML layers. BigHat reports a roughly one-week turnaround from design to data, more than 2,000 molecules produced and characterized per week, and more than 20 assays spanning developability, affinity, and function.

Those are company-reported operating metrics, but they illuminate the business thesis. Access to frontier models will spread. A system that can rapidly create task-specific biological data, record why a molecule failed, and feed that result into another design cycle is harder to reproduce. In protein therapeutics, the negative result may be more useful than another binder that looked promising under one narrow test.

Premji Invest's published rationale focuses on this design-make-test-learn loop. The investor argues that BigHat's advantage is the ability to optimize several properties on the same purified molecule, including characteristics that affect safety, manufacturing, and clinical development. That remains an investor thesis, not an independent guarantee, but it explains why the round is tied to both the platform and the pipeline.

From Platform to Clinic

BHB810 is a CDH17-directed antibody-drug conjugate being studied in advanced gastric and gastroesophageal cancers. BigHat announced the first patient dosed on September 1, 2026. A clinical-trial record for NCT07529808 identifies a recruiting Phase 1/2 study designed to examine safety, tolerability, pharmacokinetics, and preliminary antitumor activity.

That milestone confirms that a BigHat-designed candidate reached human testing. It does not establish safety or efficacy, and no human outcome data was available in the reviewed sources. The distinction matters because the AI drug-discovery market has spent years pricing platform promise before patients could supply the harder evidence.

BHB299 is an avidity-driven T-cell engager designed for CEACAM6-expressing solid tumors. An AACR abstract describes the program's preclinical design and testing, while BigHat says it plans to begin human studies in 2027. That timing remains forward-looking, and the preclinical package should not be read as a clinical result.

The Partnership Signal

BigHat has disclosed collaborations with Amgen, Merck, Johnson & Johnson, AbbVie, and Eli Lilly. The company says it completed a strategic collaboration with J&J, completed three project collaborations with Merck, and expanded work with Lilly through TuneLab. Those relationships place BigHat's platform against protein-engineering problems beyond its own pipeline, although most commercial terms and downstream outcomes remain undisclosed.

This dual model gives BigHat two evidence streams. Partner programs can test whether the platform solves difficult external briefs, while wholly owned assets force the company to carry its own design choices through manufacturing and clinical development. The first stream can validate demand; the second determines whether the platform can create medicines with durable value.

The Series C therefore finances more than compute or laboratory capacity. It supports a team that has to connect model output, experimental data, drug-development judgment, and clinical operations without losing the lesson at each handoff. That is a less theatrical problem than generating a novel sequence, but it is where the category's commercial credibility will be earned.

What the Series C Signals

AI-native biotechnology is moving into a stage where model access alone cannot carry the story. Investors can compare which companies generate proprietary experimental data, which platforms have repeated pharmaceutical partnerships, and which internal candidates reach human studies. BigHat now sits inside that more demanding comparison set.

The $75M round gives BigHat room to deepen its data engine while BHB810 and BHB299 create new tests for it. Each program can make the platform smarter, but the platform's value will ultimately be judged by the quality of the medicines and evidence it produces. The next meaningful milestones will arrive through assays, manufacturing work, patient data, and the clinical decisions that follow.