Converge Bio Scores $25M to Automate Drug Discovery—But Can Its AI Systems Beat Lab Scientists?

Betting $25 million on molecules: How Converge Bio is racing to automate drug discovery with AI—and why it might still fail to outpace the lab.

Converge Bio’s $25M Series A round, led by Bessemer Venture Partners, has fueled its mission to integrate AI into drug development. CEO Dov Gertz said:

"Our platform continues to expand across these stages, helping bring new drugs to market faster."

The company’s three AI systems—antibody design, protein yield optimization, and biomarker/target discovery—aim to streamline processes that traditionally take years. In one case study, their protein yield optimization system boosted output by 4–4.5X in a single iteration, a stark contrast to conventional methods.

However, Gertz emphasized the risks of molecular hallucinations, a unique challenge in scientific AI:

"The cost [of molecular hallucinations] is much higher" compared to text-based hallucinations.

To mitigate this, Converge employs a multi-model approach, combining LLMs, diffusion models, traditional ML, and statistical methods. Gertz added:

"We don’t rely on text-based models for core scientific understanding."

With 40 partnerships and 40 active programs, Converge’s team has grown from 9 to 34 employees in 14 months. Yet, it faces stiff competition from startups leveraging tools like AlphaFold for protein structure prediction.