← Back to Home

BigHat raises $75 million to bring AI-designed cancer antibodies into human testing

Between designing an antibody and proving that a drug works lie the challenges of laboratory experiments and clinical trials. BigHat’s new funding will support two cancer drug candidates and expand the biological experimentation platform that enables its AI to keep learning.

By SURL BioNews

AI can propose new antibody designs, but predictions alone cannot determine whether they can safely treat cancer. BigHat Biosciences, a biotechnology company based in San Mateo, California, is taking that question into human trials: the first patient has been dosed in a Phase 1 trial of its first drug candidate, while a new funding round will support further clinical development and expansion of its experimental platform.

BigHat announced on September 24 that it had completed a $75 million Series C financing round, co-led by DFJ Growth and Premji Invest. The company said it has raised a total of $223 million. The funds will be used to expand its automated biological data platform and advance BHB810 and BHB299 toward clinical results.

BHB810, the first to enter human trials, is an antibody-drug conjugate (ADC) targeting CDH17, being developed for gastric cancer and other advanced gastrointestinal tumors. This type of drug combines a target-recognizing antibody with a drug payload, with the aim of delivering anticancer activity to tumor cells. For a candidate molecule, binding to its target is only the starting point; entry into cells, payload release, and effects on healthy tissue all influence treatment efficacy and risk.

The other drug candidate, BHB299, uses a T-cell engager design intended to direct immune cells to attack solid tumors expressing CEACAM6. It remains in preclinical development. BigHat expects to begin human trials in 2027; co-lead investor Premji Invest described the timing as early 2027, which is more specific than the company’s announcement but remains a development plan.

Supporting these two drugs is a platform that connects molecular design with experimental measurements. According to Premji Invest, BigHat processes more than 2,000 antibodies each week and measures more than ten properties of each molecule, taking about one week from design to obtaining experimental data. The approach is to have AI learn about antibody binding ability, stability, and other properties needed for development simultaneously, then use experimental results to refine the next round of designs.

The value of this cycle lies in capturing biological behavior that models alone struggle to account for. Premji Invest also noted that the platform retains data on failed molecules, such as antibodies that readily aggregate or lose function after changes in conformation. However, these throughput and speed figures come from the investor’s account; the materials released for this financing round did not provide a complete dataset or comparative studies sufficient to independently assess the platform’s performance.

With funding secured and the first patient dosed, BigHat has an opportunity to translate its design capabilities into clinical evidence, but the announcements so far have not disclosed human safety or efficacy results for BHB810. The next key question is whether the drug candidates can establish acceptable doses and safety ranges in humans and gradually demonstrate antitumor effects. Experimental cycles can be accelerated, but whether a drug proves viable still has to be answered by clinical data.

References

  1. bizjournals.com
  2. BigHat Biosciences
  3. Premji Invest