Helical raises $10M to turn biological foundation models into systems for pharma

Helical has raised $10M in seed funding to turn biological foundation models from an interesting research technology into working systems for pharma companies. That sends an important signal to the market: investors aren't simply betting on yet another AI platform, but on a team that is already deploying its products at major customers. In an industry where mistakes are expensive, actual production use often counts for more than bold promises or demos.

The startup is based in London, but its roots are in Luxembourg: three childhood friends from Luxembourg founded the company. The fund redalpine led the $10M round, and the angel investors include executives from Cohere and Hugging Face. That lineup matters in its own right. It shows that the project has the confidence not only of traditional venture investors but also of people directly involved in building modern AI models and the infrastructure around them.

The key part of Helical's story isn't the size of the round but the maturity of the product. The company says it is already running in production with several pharma groups from the global top 20, and it has a public collaboration with Pfizer. That is a rare level of validation for a young AI startup. Pharma companies don't adopt new tools on impulse: they have long validation cycles and strict requirements for data quality, reproducibility, and integration into existing research workflows. If a product has made it into real use, it has cleared at least part of the hardest stretch of the road.

The goal of "turning foundation models into systems" is telling, too. A base model can perform well in lab settings and on test datasets, but businesses need concrete tools rather than abstract capabilities: a clear interface, reliable pipelines, quality control, compatibility with internal data, and transparent decision logic. In biotech and pharma, the gap between model and product is especially wide, because the work involves complex biology, expensive experiments, and drug development cycles that can stretch over many years.

The market generally uses the term bio foundation models for large models trained on biological sequences, molecular structures, or related scientific datasets. They have long been discussed as the next big layer of AI after language models, because they can help uncover connections that are hard to spot by hand. But the approach has limits: pretty predictions on slides aren't enough for pharma, which needs tools it can test on internal data and check against lab results. That's why investors increasingly look past the "magic" model itself and focus on whether a team can build it into a real research pipeline.

That makes Helical's bet look pragmatic. Rather than selling the market only the promise of "AI for drug discovery," the company, judging by its positioning, is building a layer that makes these models usable in research teams' everyday work. For pharma companies, that could mean faster analysis of biological data, easier pattern discovery, and more structured work with hypotheses. For the AI market, it's another sign of a shift from the race for the biggest models to a race for systems that actually fit into industry workflows. This applied layer is usually the biggest bottleneck to scaling.

The takeaway is simple: by the standards of the AI boom, Helical's round is relatively small, but the company makes up for it with the quality of its signals: early deployments at major pharma companies, a public tie-up with Pfizer, and a strong investor lineup. Today, new value in industry AI is being created in exactly these transitions from research hypothesis to operational tool. If the company can prove that its approach consistently delivers value in real R&D workflows, interest in applied bio-AI systems in pharma will only grow over the next few years.