BigHat Raises $75M as BHB810 Enters Phase 1

San Mateo-based BigHat Biosciences announced a completed $75 million Series C on September 24, 2026, bringing its disclosed funding to $223 million. The round, co-led by DFJ Growth and Premji Invest, is intended to advance AI-designed cancer therapies and the experimental platform used to design them. One candidate, BHB810, has reached Phase 1 human testing; the second named lead program, BHB299, remains preclinical.

Axios’s September 24 report independently confirmed the financing and identified having a candidate in the clinic as the notable milestone. That milestone has a narrow meaning: a patient has received an investigational treatment, but patient safety and efficacy results have yet to establish whether it works. The new capital supports development at two different stages, rather than two therapies already being tested in people.

What the completed round will fund

The financing covers both BigHat’s therapeutic pipeline and its system for generating biological data. Its stated purpose is to move BHB810 and BHB299 toward clinical readouts while expanding the automated experiments that inform protein design. No allocation for either candidate or the platform was disclosed with the round, so the full raise cannot be treated as a budget for a particular trial.

Those uses of capital carry different kinds of risk. Laboratory measurements can help researchers choose and refine a molecule before human testing, while a clinical study examines how that molecule behaves in patients. Funding both activities may give BigHat more opportunities to test its designs, but completing the financing does not establish that either candidate is safe, effective or commercially viable. The relevant distinction now is which parts of the development work have happened and which results are still pending.

BHB810 has entered Phase 1

BigHat’s September 1 first-patient notice reports that BHB810, a CDH17-directed antibody-drug conjugate designed with its platform, had been given to a patient in a Phase 1 trial. The investigational therapy is being studied for advanced gastric and other gastrointestinal cancers. It is BigHat’s first internally developed program to reach human testing.

The dose-escalation study is intended to assess safety, tolerability, how the drug moves through the body and preliminary antitumor activity. Initial enrollment focuses on advanced gastric and gastroesophageal junction tumors. Dosing the first participant establishes that the program crossed into clinical development; it supplies none of the study’s answers on its own.

Before the trial, experiments in tumor models produced findings that BigHat described as encouraging for tumor clearance and safety. Those were preclinical observations, with different subjects and conditions from a human study. The measurable achievement is that a molecule engineered through BigHat’s design process reached a patient. Whether its proposed properties produce a tolerable treatment and useful activity in people remains an open clinical question.

BHB299 is still being prepared for human testing

Premji Invest’s account of the round describes BHB299 as a CEACAM6-targeting T-cell engager for solid tumors that is completing preclinical development, with human trials planned for early 2027. Premji Invest co-led the financing and has a financial interest in the program’s progress. The proposed trial timing is therefore a development plan, not a completed clinical milestone.

BHB299 also tests a different molecular approach from BHB810. An antibody-drug conjugate carries a drug payload toward a target; a T-cell engager is designed to bring immune cells into contact with tumor cells. Both require evidence about selectivity and safety that cannot be supplied by the other candidate’s progress. BHB810’s entry into a trial does not mean BHB299 has completed the work needed to enter one.

Platform speed and clinical proof measure different things

BigHat’s platform description claims a one-week turnaround from design to experimental data, more than 2,000 molecules produced and characterized per week, and more than 20 assays spanning developability, binding affinity and function. These are company-reported measures of laboratory capacity. They indicate the stated pace and range of experiments, not the rate at which an investigational therapy succeeds in patients.

The platform joins computational design with automated production and testing of proteins. Measurements from one design cycle can inform the next, including measurements of properties beyond whether a molecule binds its intended target. That mechanism helps explain why the financing supports the experimental system alongside the drug candidates: the same process can generate data for future designs as current programs advance. Its value as a drug-development advantage, however, depends on whether selected molecules perform well beyond the laboratory.

The boundary is clearest in BHB810. Reaching Phase 1 shows that BigHat could move an AI-designed candidate into a human study, an observable development milestone. It does not show that faster design produced a safer or more effective medicine than another method would have produced. Such comparisons would need appropriate clinical evidence, while the immediate question is how BHB810 performs in its own trial.

The next evidence will come from patients

The financing is complete and BHB810 has begun human testing. BHB299 remains in preclinical development, and the platform’s throughput figures remain company claims about experiments. The next substantive clinical evidence will be BHB810’s safety and preliminary activity data; BHB299 must first reach its planned human study before it can generate comparable patient evidence.

For now, the Series C funds a transition that BigHat has begun but has not finished. Its first AI-designed candidate has reached patients, while the results needed to judge that treatment—and the broader clinical value of the system that designed it—are still ahead.