Artificial intelligence is rapidly advancing into the field of drug discovery. Pharmaceutical and biotech companies are adopting these tools to shorten research and development timelines and improve success rates, all while managing rising costs. More than 200 startups are now competing to integrate AI directly into research workflows, drawing significant interest from investors.
Converge Bio is the latest company to capitalize on this shift, securing new capital as competition intensifies. The Boston- and Tel Aviv–based startup helps pharma and biotech companies develop drugs faster using generative AI trained on molecular data. It has raised a $25 million oversubscribed Series A round led by Bessemer Venture Partners. TLV Partners and Vintage Investment Partners also participated, along with additional backing from unidentified executives at Meta, OpenAI, and Wiz.
In practice, Converge trains generative models on DNA, RNA, and protein sequences and then integrates them into existing drug development workflows to accelerate the process. The drug-development lifecycle has defined stages, from target identification and discovery to manufacturing and clinical trials. Within each stage, there are experiments the platform can support. The company’s platform continues to expand across these stages with the goal of helping bring new drugs to market faster.
So far, Converge has introduced three discrete customer-facing AI systems: one for antibody design, one for protein yield optimization, and one for biomarker and target discovery. The antibody design system serves as an example. It is not a single model but an integrated platform. First, a generative model creates novel antibodies. Next, predictive models filter those antibodies based on their molecular properties. Finally, a docking system uses physics-based models to simulate the three-dimensional interactions between the antibody and its target. The value lies in the complete system, providing customers with ready-to-use tools that plug directly into their workflows without needing to assemble models themselves.
This new funding arrives about a year and a half after the company raised a $5.5 million seed round in 2024. Since its founding two years ago, the startup has scaled quickly. Converge has signed 40 partnerships with pharmaceutical and biotech companies and is currently running about 40 programs on its platform. It works with customers across the U.S., Canada, Europe, and Israel and is now expanding into Asia.
The team has grown rapidly to 34 employees from just nine in November 2024. The company has also begun publishing public case studies. In one instance, Converge helped a partner increase protein yield by four to four and a half times in a single computational iteration. In another, the platform generated antibodies with extremely high binding affinity, reaching the single-nanomolar range.
AI-driven drug discovery is experiencing a surge of interest industry-wide. Last year, Eli Lilly partnered with Nvidia to build what they called the pharma industry’s most powerful supercomputer for drug discovery. In October 2024, the developers behind Google DeepMind’s AlphaFold project won a Nobel Prize in Chemistry for creating the AI system that predicts protein structures.
Regarding this momentum, Converge Bio’s CEO notes the company is witnessing the largest financial opportunity in the history of life sciences, with the industry shifting from trial-and-error approaches to data-driven molecular design. The skepticism that existed when the company was founded has vanished remarkably quickly, thanks to successful case studies from companies and academia.
While large language models are gaining attention for their ability to analyze biological sequences and suggest new molecules, challenges like hallucinations and accuracy remain. In text, hallucinations are usually easy to spot, but validating a novel molecular compound can take weeks, making the cost of error much higher. To tackle this, Converge pairs generative models with predictive ones to filter new molecules, reducing risk and improving outcomes for its partners. This filtration is not perfect, but it significantly lowers risk.
When asked about experts who remain skeptical of using LLMs for such purposes, the CEO expressed agreement that text-based models are not sufficient for core scientific understanding. To truly understand biology, models must be trained on DNA, RNA, proteins, and small molecules. At Converge, text-based LLMs are used only as support tools, for example to help customers navigate scientific literature. They are not the core technology. The company uses a variety of methods including LLMs, diffusion models, traditional machine learning, and statistical techniques as appropriate.
The company’s vision is for every life-science organization to use Converge Bio as its generative AI lab. Wet labs will always exist, but they will be paired with generative labs that create hypotheses and molecules computationally. Converge aims to be that generative lab for the entire industry.