TL;DR
Converge Bio, a Boston- and Tel Aviv–based startup using generative AI trained on molecular sequences, closed a $25 million Series A led by Bessemer Venture Partners. The company says its integrated systems — for antibody design, protein yield optimization and target discovery — are already running about 40 customer programs across multiple regions.
What happened
Converge Bio announced an oversubscribed $25 million Series A financing led by Bessemer Venture Partners, with participation from TLV Partners, Vintage Investment Partners and unnamed executives from Meta, OpenAI and Wiz. The two-year-old company, which operates out of Boston and Tel Aviv, develops generative models trained on DNA, RNA and protein sequences and packages those models into customer-facing systems intended to accelerate stages of the drug-development lifecycle. Converge has released three discrete product systems — for antibody design, protein yield optimization and biomarker/target discovery — and describes the antibody workflow as a combination of generative models, predictive filters and a physics-based docking simulator. Since its $5.5 million seed round in 2024, the startup says it has grown to 34 employees, signed roughly 40 partnerships and is running about 40 programs across the U.S., Canada, Europe and Israel while beginning to expand into Asia.
Why it matters
- The round underscores continued investor interest in AI applications for drug discovery amid a crowded startup landscape.
- Converge’s packaged systems aim to reduce the integration burden for pharma partners by delivering ready-to-use models rather than isolated algorithms.
- Combining generative models with predictive filters and physics-based docking is intended to mitigate risks tied to false positives and costly experimental validation.
- Progress from companies like Converge could shorten parts of the R&D timeline if computational outputs translate reliably into wet-lab results.
Key facts
- Series A raise: $25 million, described as oversubscribed.
- Lead investor: Bessemer Venture Partners; other backers include TLV Partners and Vintage Investment Partners.
- Additional funding came from unidentified executives at Meta, OpenAI and Wiz.
- Headquarters and bases: Boston and Tel Aviv.
- Core approach: trains generative models on DNA, RNA and protein sequences and integrates them into pharma workflows.
- Three customer-facing systems launched: antibody design, protein yield optimization and biomarker/target discovery.
- Antibody design workflow combines generative models, predictive filtering and physics-based docking simulations.
- Earlier financing: $5.5 million seed round in 2024.
- Commercial traction: about 40 partnerships and roughly 40 active programs; customers in the U.S., Canada, Europe and Israel; beginning expansion into Asia.
- Team growth: expanded to 34 employees from nine in November 2024.
What to watch next
- Outcomes from the ~40 active programs — whether computational designs produce reproducible wet-lab results and downstream leads.
- The company’s stated expansion into Asia and how that affects partnerships and regulatory interactions.
- Further publishing of case studies or peer-reviewed validation of claimed gains in yield and binding affinity.
- not confirmed in the source: specific plans for the new funds or timelines for commercial milestones.
Quick glossary
- Generative model: A machine learning model that creates new data samples — in this context, proposed molecular or sequence designs — based on patterns learned from training data.
- Predictive model: A model that estimates properties or outcomes (for example, stability or binding affinity) of candidate molecules to help prioritize which to test experimentally.
- Docking (physics-based): A computational technique that simulates the three-dimensional interactions between molecules, often used to assess how well an antibody or small molecule might bind a target.
- Biomarker: A measurable biological indicator (such as a molecule or gene expression pattern) used to detect or monitor a disease process or response to a therapy.
Reader FAQ
How much did Converge Bio raise and who led the round?
Converge raised $25 million in a Series A led by Bessemer Venture Partners.
What does Converge Bio’s technology do?
The company trains generative models on DNA, RNA and protein sequences and bundles those models into systems for antibody design, protein yield optimization and biomarker/target discovery.
How many customers or programs does the company have?
Converge says it has signed about 40 partnerships and is running roughly 40 programs across several regions.
How will the new funding be used?
not confirmed in the source
Artificial intelligence is moving quickly into drug discovery as pharmaceutical and biotech companies look for ways to cut years off R&D timelines and increase the chances of success amid rising…
Sources
- Converge Bio raises $25M, backed by Bessemer and execs from Meta, OpenAI, Wiz
- AI / Machine Learning | July Round-Up 2025
- Generative AI – Major Product Launches and Partnerships
- List of Artificial Intelligence Companies in France – 39
Related posts
- A deep dive on agent sandboxes — Codex, macOS Seatbelt, Linux seccomp and policy
- Meta-backed Hupo pivots from mental wellness to AI sales coaching, raises $10M
- Over 100 VC-backed tech startups reached unicorn status in 2025