← Back to Home

BigHat raises $75 million as AI-designed cancer antibodies move into human trials

From computer design to laboratory measurements, BigHat aims to optimize multiple drug properties of antibodies together. The new funding will support the development of two cancer therapies, while the real test will be their safety and efficacy in humans.

By SURL BioNews

Designing an antibody that can recognize cancer cells is only the starting point for drug development. Whether the molecule can be reliably manufactured, function effectively, and retain a sufficient safety margin while attacking tumors also determines success or failure. U.S. biotech company BigHat Biosciences is combining artificial intelligence with automated experiments to try to address these challenges earlier. As its first drug candidate enters human trials, this design approach is also beginning to undergo clinical testing.

On September 24, 2026, BigHat announced the completion of a $75 million Series C financing round co-led by DFJ Growth and Premji Invest, bringing its total funding to $223 million. The funds will be used to expand its AI and experimental data generation platform and advance two cancer therapy candidates, BHB810 and BHB299, toward generating clinical data.

At the heart of the platform is a continuous cycle between design and experimental measurement. Co-lead investor Premji Invest said the platform can process more than 2,000 antibodies per week and measure more than a dozen properties for each molecule, with approximately one week from design to data. Its investment thesis is that models need to learn not only whether an antibody can bind to its target, but also whether the molecule can meet the multiple requirements needed to become a drug. However, these throughput and speed figures come from the investor's account and cannot be taken directly as evidence of a higher probability of clinical success.

BHB810, currently the most advanced candidate, is an antibody-drug conjugate (ADC) targeting CDH17. It uses an antibody to recognize a tumor-associated target and then delivers its drug payload toward cancer cells. In early September, the company announced that the first patient had been dosed in its Phase 1 trial. The trial is initially enrolling patients with advanced gastric cancer and gastroesophageal junction tumors to assess safety, tolerability, pharmacokinetics, and preliminary antitumor activity.

BigHat said BHB810 had shown activity in nearly 30 patient-derived and cell-derived tumor models, including gastric cancer models with low or heterogeneous CDH17 expression. This provides a preclinical basis for studies in humans, but tumor responses in models cannot yet be used to infer treatment effects in patients. The related announcements also provided no human efficacy or safety results that could be assessed.

The other candidate, BHB299, is a T-cell engager being developed for solid tumors expressing CEACAM6, intended to bring immune cells toward tumors. The company plans to begin human trials in 2027, with Premji Invest's account further specifying early 2027. This remains a development timeline, rather than a completed clinical milestone.

The financing gives BigHat more resources to sustain both its data platform and its own drug development. The most decisive evidence next will be whether the candidate molecules can achieve a dose range in humans that is both tolerable and has antitumor activity. Experimental cycles can shorten the wait between designs, but the evidence required for human trials and subsequent regulatory review must still be built step by step.

References

  1. The Business Journals
  2. BigHat Biosciences
  3. Premji Invest