The integration of artificial intelligence into drug discovery is accelerating as pharmaceutical and biotech firms seek to reduce lengthy research and development timelines and improve success rates against a backdrop of increasing costs. Over 200 startups are currently vying to embed AI directly into research processes, drawing heightened attention from investors. Converge Bio is the most recent company to capitalize on this trend, securing fresh funding as competition intensifies in the AI-powered drug discovery sector.

Based in Boston and Tel Aviv, the startup assists pharmaceutical and biotech companies in accelerating drug development by employing generative AI models trained on molecular data. It has now closed a $25 million oversubscribed Series A financing round led by Bessemer Venture Partners. The investment also included participation from TLV Partners and Vintage Investment Partners, with additional support from unnamed executives at Meta, OpenAI, and Wiz.

Operationally, Converge trains its generative models on DNA, RNA, and protein sequences and integrates them into existing industry workflows to expedite the drug development process. "Our platform continues to expand across these stages, helping bring new drugs to market faster," the company stated.

To date, Converge has launched several customer-facing AI systems. The startup has introduced three distinct platforms: one for antibody design, another for optimizing protein yield, and a third for biomarker and target discovery. CEO Gertz elaborated using the antibody design system as an example, explaining that it integrates three components. A generative model first creates novel antibody candidates. Predictive models then filter these based on molecular properties. Finally, a physics-based docking system simulates the three-dimensional interaction between the antibody and its target.

Gertz emphasized that the value resides in the complete, integrated system rather than any individual model. "Our customers don’t have to piece models together themselves. They get ready-to-use systems that plug directly into their workflows," he noted.

This new funding arrives approximately eighteen months after the company’s $5.5 million seed round in 2024. In that time, the two-year-old startup has scaled operations significantly. Converge has established 40 partnerships with pharmaceutical and biotech companies and is currently managing around 40 active programs on its platform. Its client base spans the U.S., Canada, Europe, and Israel, with expansion into Asia now underway.

The team has also grown rapidly, increasing from just nine employees in November 2024 to 34 today. The company has begun publishing public case studies demonstrating its impact. In one instance, Converge helped a partner increase protein yield by 4 to 4.5 times in a single computational cycle. In another, the platform generated antibodies exhibiting extremely high binding affinity, reaching the single-nanomolar range.

Interest in AI-driven drug discovery is surging industry-wide. Last year, Eli Lilly partnered with Nvidia to construct what they described as the pharmaceutical industry's most powerful supercomputer for discovery research. Furthermore, in October 2024, the developers behind Google DeepMind’s AlphaFold project were awarded the Nobel Prize in Chemistry for creating the AI system capable of predicting protein structures.

When asked about this momentum and its effect on Converge Bio, Gertz remarked that the company is operating within what he sees as the largest financial opportunity in the history of life sciences. He observed a sector-wide shift from traditional "trial-and-error" methods toward data-driven molecular design. "We feel the momentum deeply, especially in our inboxes. That skepticism has vanished remarkably quickly, thanks to successful case studies from companies like Converge and from academia," he added.

While large language models are attracting notice for their ability to analyze biological sequences and propose new molecules, challenges such as hallucinations and accuracy persist. "In text, hallucinations are usually easy to spot," Gertz said. "In molecules, validating a novel compound can take weeks, so the cost is much higher."

To address this, Converge combines generative models with predictive models to filter newly generated molecules, thereby reducing risk and improving outcomes for partners. "This filtration isn’t perfect, but it significantly reduces risk and delivers better outcomes for our customers," Gertz stated.

He also clarified the company's technical approach, expressing alignment with AI pioneer Yann LeCun. "We don’t rely on text-based models for core scientific understanding. To truly understand biology, models need to be trained on DNA, RNA, proteins, and small molecules," Gertz explained. Text-based LLMs are utilized solely as support tools, for instance to help clients navigate scientific literature related to generated molecules.

"They’re not our core technology," Gertz said. "We’re not tied to a single architecture. We use LLMs, diffusion models, traditional machine learning, and statistical methods when it makes sense."

Looking ahead, Gertz outlined an ambitious vision: "Our vision is that every life-science organization will use Converge Bio as its generative AI lab. Wet labs will always exist, but they’ll be paired with generative labs that create hypotheses and molecules computationally. We want to be that generative lab for the entire industry."