Harell Data

About Harell Data

Harell Data is developing an intelligent infrastructure platform aimed at advancing scientific discovery by addressing critical bottlenecks in AI research. The company focuses on the physical sciences, such as drug discovery and material science, where valuable training data often remains siloed, unorganized, and inaccessible. This fragmentation hinders the progress of large-scale AI models.

The venture was founded by Harlan Robins, who previously co-founded and served as Chief Scientific Officer at Adaptive Biotechnologies. His experience decoding the human immune system highlighted a significant gap in the market: entities that generate high-value, proprietary research data often lack a secure and profitable method to commercialize these assets for AI model training. In February 2026, Harell Data secured $15 million in a Seed funding round from investors including Fuse and Cercano Management to build out its platform. The company has an 11-person team split between Bellevue, WA, and Palo Alto, CA.

Harell Data provides an end-to-end ecosystem that allows data owners to securely share proprietary datasets with AI modelers. The platform brings high-performance compute and an orchestration layer directly to the data, a model that prevents the raw data from ever leaving the secure environment. This integrated approach includes on-demand GPU compute, fine-tuning infrastructure, inference deployment pods, and lineage tracking. The business model enables data creators to earn a direct share of the compute revenue generated when their datasets are used for training models, creating a new monetization path beyond speculative, long-term royalties. Early partners providing datasets include A-Alpha Bio and Adaptive Biotechnologies.