BigHat Biosciences has raised $75 million in Series C financing as the clinical-stage biotechnology company expands an AI-native platform that combines protein design models with autonomous laboratory experimentation. The company says the funding will support its data-generation infrastructure and advance AI-designed therapeutics, including BHB810, which entered Phase 1 clinical testing in September 2026.
The next stage of AI development is increasingly moving beyond digital workflows and into systems that can generate and validate data in the physical world.
BigHat Biosciences is building around that model in drug discovery.
The San Mateo-based biotechnology company announced a $75 million Series C financing, bringing its total funding raised to $223 million, according to the company. DFJ Growth and Premji Invest co-led the round, with participation from new and existing strategic and financial investors.
The financing is intended to expand BigHat’s AI and experimental platform while supporting its pipeline of AI-designed protein therapeutics.
The company’s technology combines computational protein design with high-throughput automated experimentation. BigHat describes the resulting system as a continuous learning loop in which AI generates molecular designs, experiments produce biological data and those results feed back into subsequent design cycles.
Building AI infrastructure around biological data
One of the more important aspects of BigHat’s platform is that it is not solely a model-development system.
The company has built infrastructure to generate the biological data needed to train, evaluate and improve AI models for protein engineering.
Its Milliner platform connects computational design with automated wet-lab experimentation. BigHat says the system can design and evaluate thousands of antibodies each week while simultaneously optimizing multiple properties relevant to therapeutic development.
That creates a different type of AI development cycle from conventional software.
In a typical machine-learning workflow, engineers collect an existing dataset, train a model and evaluate its predictions. In protein design, the quality and relevance of experimental data can become a limiting factor because predictions ultimately have to be validated against physical molecules.
BigHat’s approach is to integrate those two stages.
Its platform generates designs, synthesizes and tests molecules, captures experimental results and uses those results to inform subsequent computational work. The company says its laboratory infrastructure is directly connected to its data systems to reduce manual handoffs and maintain the relationship between experimental results and individual designs.
From AI models to autonomous experimentation
That architecture also explains BigHat’s emphasis on agentic therapeutic design.
The company says it has established partnerships with frontier AI companies to deploy state-of-the-art models for agentic therapeutic design. Rather than using AI simply to predict a molecular property, the broader objective is to connect computational reasoning with experimental execution.
This represents an emerging category of AI infrastructure in which autonomous systems interact with physical-world tools.
In BigHat’s case, the environment is a high-throughput biological laboratory. The system can coordinate computational models, molecular design and automated experiments, creating a feedback loop between AI predictions and physical measurements.
The potential advantage is not simply faster inference. A continuously operating experimental system can generate new proprietary data that can subsequently improve the models used for future designs.
Clinical validation adds another layer
BigHat’s technology is now being tested against the requirements of therapeutic development.
The company announced in September that the first patient had been dosed in a Phase 1 study of BHB810, a CDH17-directed antibody-drug conjugate for gastric cancer and other advanced gastrointestinal tumors. BigHat describes BHB810 as the first clinical program originating from its AI discovery platform.
Its second development candidate, BHB299, is an avidity-driven T-cell engager targeting CEACAM6-expressing solid tumors. BigHat says it expects the program to enter human trials in 2027.
The distinction between computational performance and clinical performance remains important. A model can generate promising molecular candidates without guaranteeing that those candidates will succeed in human trials.
BigHat’s move into clinical development therefore provides a longer-term test of whether its integrated AI-and-experimentation approach can translate computational design into viable therapeutic candidates.
Data becomes a competitive AI asset
The financing also highlights a broader issue in AI-driven scientific discovery: proprietary data can be as important as the models themselves.
Foundation models and protein language models can provide powerful starting points for molecular design, but organizations still need high-quality biological data to evaluate and improve those systems.
BigHat’s strategy is to control both sides of that equation.
The company develops AI systems for protein engineering while operating an experimental platform capable of producing new biological measurements. This creates a potential compounding loop in which each experimental campaign contributes additional data to future design cycles.
That model resembles the broader trend toward AI systems that can generate their own domain-specific training and evaluation data rather than relying exclusively on static public datasets.
Pharmaceutical partnerships extend the platform
BigHat has also been using its platform through partnerships with pharmaceutical companies.
The company lists collaborations involving Amgen, Merck, Johnson & Johnson, AbbVie and Eli Lilly. In August 2026, BigHat said it had completed three project collaborations with Merck using its machine-learning and high-speed wet-lab capabilities to design antibody candidates across multiple therapeutic areas and protein-engineering challenges.
Those partnerships provide an additional application for the platform: using AI and automated experimentation as infrastructure that can address specific protein-engineering problems for external organizations.
The financing gives BigHat additional capital to expand both its internal pipeline and this platform model.
AI-native drug discovery moves toward closed-loop systems
BigHat’s latest funding illustrates how AI in biotechnology is evolving from predictive modeling toward closed-loop scientific systems.
In this model, AI does not end when it produces a prediction. The prediction triggers an experiment, the experiment produces new data, and that data becomes an input for the next computational cycle.
For AI infrastructure, that creates requirements extending beyond model training. Automation, laboratory robotics, data integrity, experiment orchestration and multimodal learning become part of the overall system.
BigHat’s challenge now is to demonstrate that this integrated architecture can consistently produce therapeutics that progress through increasingly demanding stages of development.
The $75 million financing gives the company additional resources to pursue that objective while expanding the data and AI infrastructure behind its approach.
Market Landscape
AI-driven drug discovery is moving from standalone prediction models toward integrated platforms that combine machine learning, protein-language models, automated experimentation and biological data generation.
BigHat’s model represents one version of this shift: computational design and laboratory experimentation operate as a connected learning system. Other AI-biotech companies are pursuing related approaches across protein engineering, small-molecule discovery, molecular simulation and experimental automation.
The strategic significance lies in controlling the full learning loop. Better models can generate better candidates, while faster and higher-quality experiments can provide more useful data for subsequent model improvement.
BigHat’s clinical-stage programs provide an opportunity to evaluate whether that infrastructure can translate into therapeutic development outcomes rather than remaining primarily a computational research platform.
Top Insights
- BigHat raised $75 million in Series C financing to expand its AI-driven protein design platform and advance AI-designed therapeutics.
- Its Milliner platform connects computational molecular design with automated experimentation to create a continuous AI-to-biology learning loop.
- BHB810 entered Phase 1 testing in September 2026, providing an early clinical milestone for BigHat’s AI discovery platform.
- The company’s strategy emphasizes proprietary biological data generation alongside frontier AI models for protein engineering.
- Pharmaceutical collaborations with Merck, Lilly, Amgen, J&J and AbbVie extend the platform beyond BigHat’s internal therapeutic pipeline.
Power Tomorrow’s Intelligence — Build It with TechEdgeAI
