TileBio has created a self-learning AI platform that ingests millions of unlabelled whole-slide pathology images and extracts clinically relevant patterns without the need for manual annotation. By leveraging large-scale real-world data from NHS Greater Glasgow and Clyde, the system aims to generalise across cancer types, detect disease earlier, and uncover novel biomarkers useful for drug development. The company’s technology hinges on foundation-model-style representation learning, which promises faster, more accurate, and more interpretable results than traditional medical AI approaches that rely on expensive ground-truth labels. Near-term plans include large-scale model training, extensive clinical validation studies, and the recruitment of deep-learning engineers, data managers, regulatory specialists, and commercial leads. Longer term, TileBio intends to build one of the world’s largest pathology foundation models spanning more than two million whole-slide images and to deploy it population-wide for cancer triage. Financially, the company has raised £1.6 million in seed capital, providing runway to hit key technical and regulatory milestones. No revenue or user metrics have yet been disclosed.

Total raised

$2M

Funding rounds

1

Latest round

Seed

Latest activity

Mar 2026