What You Should Know
– Converge Bio has announced a $25M Series A funding round led by Bessemer Venture Partners, bringing its total raised to $30M.
– Moving beyond the “AI promise” gap, the company provides end-to-end AI systems that plug directly into existing drug development workflows, already serving over a dozen pharmaceutical and biotech customers to optimize target discovery and protein manufacturing.
The Systemic Pivot: Bridging the Gap Between Promise and Reality
While the AI drug discovery sector has seen massive momentum since the recent Nobel Prize in Chemistry for AlphaFold developers, a gap remains between benchmark performance and real-world clinical utility. Converge Bio aims to close this gap by shifting the focus from individual models to integrated AI systems.
- No-Code Biologist Interface: Biologists use the platform to generate actionable outputs—such as novel targets or optimized antibodies—without needing to write code or build infrastructure.
- Experimental Validation: The platform’s models are trained on large-scale datasets obtained from high-throughput screening and rigorously curated public data, ensuring outputs are grounded in biological reality.
- Data Ownership: Customers can create private, fine-tuned instances of models using their own proprietary data while retaining full ownership.
Proven Results: 40 Programs and 7x Yield Increases
The platform is already delivering measurable commercial results across oncology, neurodegenerative, and autoimmune diseases. In the past year, Converge completed over 40 programs, achieving significant milestones:
- Antibody Binding: Discovery of novel antibodies with strong, single-digit nanomolar binding affinities.
- Manufacturing Efficiency: Consistent improvements in protein manufacturing yields by 4 to 7x.
- Patient Response: Identification of novel molecular biomarkers to optimize how patients respond to specific therapies.
Despite the momentum, there’s still a significant gap between AI promise and reality in drug discovery. “The conversation needs to shift from models to AI systems,” said Dov Gertz, CEO and co-founder of Converge Bio. “Unlike ChatGPT, you can’t simply prompt a model and get useful results. There’s a long road from a model that performs well on benchmarks to an AI system a biologist can actually use. It requires high-quality data, the right architectures for the domain, and a tight experimental validation loop. That’s why the vast majority of drug development is still done the old way: through trial and error, taking years and costing hundreds of millions of dollars.”
