Skip to main content

TileBio

Category: AI in Healthcare

A University of Glasgow spin-out developing self-learning AI foundation models for digital pathology that learn the visual language of tissue directly from unlabelled medical images to improve cancer diagnosis and biomarker discovery. TileBio was founded in 2025. The company is led by Christopher Walsh. Based in Glasgow, Scotland, United Kingdom. Team size: 1-10. Total funding raised: $2.0M. Latest round: Seed. Key investors include Twin Path Ventures, Scottish Enterprise, GU Holdings Ltd (University of Glasgow), Cancer Research UK (via Cancer Tech Accelerator).

Founded
2025
Headquarters
Glasgow, Scotland, United Kingdom
Team size
1-10
Total funding
$2.0M

Value proposition

AI platform that learns the language of diseased tissues directly from millions of unlabelled medical pathology images, bypassing the industry's biggest bottleneck of expensive, biased human-annotated labels, while generalising across all cancer types and subtypes.

Products and solutions

Self-learning AI pathology foundation model that converts histopathology images into an AI-learned visual language of tissue, enabling automated interpretation for clinical diagnostics, biomarker discovery, and drug development. Uses proprietary self-supervised learning on unlabelled whole-slide images to produce interpretable histomorphological phenotype clusters (HPCs).

Unique value

Self-supervised AI that learns a universal visual language of tissue from unlabelled pathology images at scale, combined with exclusive access to millions of NHS clinical whole-slide images for training one of the largest pathology foundation models globally.

Target customer

Clinical pathology departments, pharmaceutical companies for drug development and biomarker discovery, and life sciences researchers

Industries served

Healthcare (pathology, oncology diagnostics), Pharmaceutical (drug discovery, biomarker development), Biotechnology (precision medicine)

Technology advantage

Proprietary self-supervised learning method (Barlow Twins backbone) for histopathology that learns stain-robust, scale-robust representations without manual annotations; converts tissue morphology into structured "documents" enabling LLM-style interpretation; exclusive data partnership with NHS West of Scotland Innovation Hub and NHS Greater Glasgow and Clyde for population-scale training data.

How they differentiate

Self-supervised learning approach that requires zero human-annotated labels; ability to generalise across all cancer types and subtypes; access to large-scale real-world NHS clinical data (2M+ whole slide images from NHS Greater Glasgow and Clyde); proprietary visual language model that treats tissue morphology as structured documents; interpretable phenotype clusters (HPCs) with biological and clinical relevance.

Main competitors

Paige AI (US), PathAI (US), Ibex Medical Analytics (Israel), Mindpeak (Germany), Proscia (US)

Key partnerships

NHS Greater Glasgow and Clyde (data partnership for training on 2M+ whole slide images), NHS West of Scotland Innovation Hub, Terrain Life Science (strategic collaboration on AI-powered precision medicine, Sep 2025), University of Glasgow (spin-out relationship, MedTech Innovation Fund, Infinity G accelerator)

Major milestones

2024-07: Accepted into Phase II of Cancer Tech Accelerator (Cancer Research UK), 2025-09: Strategic collaboration announced with Terrain Life Science, 2025-11: Published self-supervised AI research on lung adenocarcinoma in bioRxiv, 2026-02: Raised £1.6M seed round led by Twin Path Ventures, 2026-03: Official public launch as University of Glasgow spin-out

Market positioning

Early-stage academic spin-out competing in the rapidly growing AI digital pathology market, differentiated by self-supervised learning approach and exclusive access to large-scale NHS clinical data. Positioned as a foundation model builder for pathology rather than a narrow disease-specific AI tool.

Geographic focus

UK (primary), with ambitions for global expansion

About Christopher Walsh

Post-Doctoral Researcher at Cancer Research UK Scotland Institute; PhD in deep learning and cancer science from University of Glasgow (2024); BSc (1st Class) in Computer Science with AI from Heriot-Watt University

Latest news about TileBio

More AI in Healthcare companies

Official website: