Helical raises $10M to bridge the gap between foundation models and drug discovery decisions

Pharma artificial intelligence startup Helical Ltd. announced today that it has raised $10 million in new funding to expand its virtual AI lab platform, which turns biological foundation models into reproducible in-silico drug discovery workflows.

Founded in 2024, Helical was built around a specific gap its founders identified as biological foundation models have gained traction in pharma. Biological foundation models are large AI systems trained on vast datasets of genomic, proteomic and other molecular data that can predict how biological systems behave, much as large language models predict text.

The gap, as found by Helical’s founders, is one whereby, as teams were excited about the model layer, they nonetheless struggled to turn computational outputs into decisions scientists could trust, reproduce and defend. Experiments lived in one-off notebooks, and bench scientists and machine learning engineers worked in separate environments, meaning that results rarely transferred cleanly across programs.

Helical addresses that gap with an application layer that turns powerful models into systems scientists can run, trust and defend.

Pitched as “the virtual AI lab for pharma,” Helical’s platform offers two product surfaces built on shared data and shared models: a Virtual Lab designed for biologists and translational scientists and The Model Factory, which serves machine learning engineers and data scientists. By putting both functions in the same system, the company aims to close the gap between computational predictions and biological decision-making.

The three co-founders came to the problem from different directions. Rick Schneider built technology at Amazon.com Inc. and later helped German enterprise software company Celonis SE scale operations in France and Japan. Maxime Allard led data science teams at IBM Corp. before pursuing a doctorate focused on reinforcement learning and robotics. Mathieu Klop was a cardiologist and genomics researcher.

When bio foundation models emerged, the trio saw the chance to build the missing application layer that would let pharma teams move from model experimentation to reproducible, production discovery.

“The models alone don’t discover drugs. The system does,” said Schneider. “Pharma teams need a system that turns foundation models into workflows scientists can run, validate and defend.”

Helical is already in production with multiple top global pharma companies, including a public collaboration with Pfizer Inc. on predictive blood-based safety biomarkers. The company claims teams have compressed discovery timelines from years to weeks across deployments in target identification, biomarker discovery and therapeutic design.

The seed funding round was led by Redalpine Venture Partners AG. Gradient, BoxGroup Ventures, Frst Capital and individual investors including Cohere Inc. CEO Aidan Gomez, Hugging Face CEO Clement Delangue and professional soccer player Mario Goetze also participated.

“We are at a unique point in time where biological foundation models and general language reasoning models are converging” said Daniel Graf, general partner at Redalpine. “We backed Helical because we strongly believe they have what it takes to build the pharma AI orchestration platform that will drive this transition from siloed AI models to integrated virtual AI labs.”

Image: Helical