Helical Raises $10M to Fix Pharma’s AI Reproducibility Problem

LONDON – April 14, 2026 – Helical, a startup building a virtual AI laboratory for the pharmaceutical industry, has secured $10 million in seed funding to tackle one of the most stubborn challenges in modern drug discovery: making computational science reproducible at scale. The round was led by the European venture firm redalpine, with significant participation from Google-backed Gradient Ventures, BoxGroup, and Frst.

In a striking vote of confidence from the core of the AI community, the round also includes angel investments from Aidan Gomez, CEO of Cohere and co-author of the seminal “Attention Is All You Need” paper that introduced the Transformer architecture, and Clement Delangue, CEO of the open-source AI hub Hugging Face. The investment will fuel the expansion of Helical’s platform, which is designed to bridge the growing chasm between powerful biological foundation models and the practical, day-to-day work of pharmaceutical R&D teams.

While AI has been heralded as the solution to pharma's daunting economics—where R&D spending exceeds $300 billion annually and over 90 percent of drugs fail in clinical trials—many initiatives stall. The problem is not the models themselves, but the fragmented workflow between a model’s prediction and a verifiable scientific decision. Helical aims to solve this by providing a unified application layer that turns theoretical AI outputs into dependable discovery systems.

The Bottleneck Beyond the Model

The pharmaceutical industry is not short on promising hypotheses, but it is constrained by the slow, expensive, and laborious process of physical experimentation. The emergence of large-scale biological foundation models—complex AI systems trained on vast biological datasets—promised a new era of in-silico discovery, where scientists could test ideas computationally before ever entering a wet lab.

However, the reality has been more complex. AI models can generate predictions, but they often operate as black boxes. Biologists and machine learning engineers frequently work in silos, using disparate tools and one-off analyses that are difficult to verify, reproduce, or build upon. This creates a critical bottleneck where promising computational leads fail to translate into confident biological decisions, stalling progress and wasting resources.

“The models alone don’t discover drugs. The system does,” said Rick Schneider, co-founder of Helical, in a statement. “Pharma teams need a system that turns foundation models into workflows scientists can run, validate, and defend. We built Helical to make in-silico science reproducible at pharma scale, so teams can go from hypothesis to decision in days instead of months.”

Helical’s platform provides this system through two interconnected interfaces: the Virtual Lab, designed for biologists and translational scientists to design and run computational experiments, and the Model Factory, built for data scientists and ML engineers to train, fine-tune, and deploy models. Both operate on the same underlying data and evidence, creating a shared, collaborative environment that closes the gap between computational theory and biological practice.

From School Friends to Pharma Innovators

Founded in early 2024, Helical was born from the combined vision of three school friends who approached the same problem from different, complementary disciplines. Rick Schneider brought tech and enterprise scaling experience from his time at Amazon and Celonis. Maxime Allard, a former data science lead at IBM, contributes deep expertise in reinforcement learning and robotics from his PhD research. The crucial link to the industry is Mathieu Klop, a practicing cardiologist and genomics researcher who understands the real-world needs of medical science.

This unique blend of expertise in enterprise tech, advanced AI, and clinical medicine allowed the trio to identify the missing piece in the pharma AI puzzle: not another model, but the operational infrastructure to make all models useful. Their vision attracted early support, including a pre-seed round in 2024, and has now culminated in a significant seed investment from strategic partners.

“We are at a unique point in time where biological foundation models and general language reasoning models are converging,” noted 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."

Early Validation and a Crowded Field

Despite its relative youth, Helical is already in production with several top-20 global pharmaceutical companies. A public collaboration with Pfizer on developing predictive blood-based safety biomarkers serves as a key proof point. According to the company, its platform has enabled teams to compress discovery timelines from years to weeks and has seen organic adoption as initial projects expand into adjacent therapeutic areas.

Helical enters a competitive and well-funded landscape. Giants like Google's Isomorphic Labs and well-established players like Recursion Pharmaceuticals and Exscientia are also leveraging AI for drug discovery. However, many competitors focus on developing their own proprietary models or drug pipelines. Helical differentiates itself by providing an enabling layer—an orchestration platform that allows pharma companies to better leverage their own data and a diverse ecosystem of models, whether open-source or internally developed.

This focus on infrastructure and reproducibility appears to be a key draw for its investors. The involvement of the CEOs of Cohere and Hugging Face is particularly telling. It signals a belief that the next phase of AI adoption in science will depend on robust, open, and accessible application layers that empower scientists, rather than on closed, monolithic systems.

With the new funding, Helical plans to deepen its deployments with existing clients, expand its footprint to other major pharma organizations, and continue enhancing its platform’s “compounding evidence layer,” which is designed to improve performance and reliability across different diseases. The company’s mission remains to equip every scientist with the ability to test hypotheses at the speed of computation, aiming to finally make in-silico discovery a reliable engine for R&D throughput.