Chelveston AI factory to tackle complex healthcare and logistics in 2GW UK supercomputing push
AI Pathfinder will build an AI data centre in Chelveston from December, set to handle complex healthcare and logistics work. It marks the first phase of 2GW UK supercomputing.

New AI data centre to take complex healthcare work
18 September 2025, 12:21 BST
AI Pathfinder will build an AI data centre in Chelveston, Northamptonshire, with construction starting in December. The site is planned to handle complex workloads across healthcare and logistics and is the first step toward 2GW of national supercomputing capacity.
This follows a UK government £31bn pact with several US tech firms to speed AI research and advances in drug discovery. For clinical leaders, it signals better access to compute for imaging, genomics, and decision support.
AI Pathfinder says it controls multiple sites across the UK and has experience delivering complex projects. The government also confirmed a £900m investment by Prologis to expand Daventry International Rail Freight Terminal-strengthening rail freight, reducing road reliance, and supporting jobs that underpin medicines and device supply chains.
What this could enable in healthcare
- Medical imaging: model training and inference at scale for radiology and pathology.
- Genomics and multi-omics pipelines for research and precision medicine.
- Clinical NLP on EHRs for cohorting, safety surveillance, and coding support.
- Hospital flow and logistics optimisation across theatres, beds, and pharmacy.
- Model testing for drug discovery and trial simulation in secure environments.
What leaders should do now
- Prioritise 3-5 high-impact, compute-intensive use cases with clear clinical and operational outcomes.
- Tighten data governance: DPIAs, PHI minimisation, role-based access, and audit logging.
- Plan integration: FHIR/HL7, traceability, and human oversight at decision points.
- Vet providers on security (ISO 27001, NHS DSPT), data residency, incident response, and uptime SLAs.
- Define validation paths: bias testing, clinical safety (DCB 0129/0160), and MHRA routes for SaMD.
- Upskill staff in prompt design, data science, and MLOps; establish an internal playbook for approvals and rollout.
For policy and regulatory context, see the UK approach to AI regulation and MHRA's program on Software and AI as a Medical Device: UK AI regulation and MHRA SaMD change programme.
Timeline and scale
Construction is due to begin in December near Rushden. The company presents this as the first phase of delivering 2GW of supercomputing capacity across the UK.
If your team needs structured upskilling for safe, effective AI projects, explore AI courses by job role.