The United Kingdom ranks among the top three nations globally for artificial intelligence research output and health technology patents. It also hosts the world's largest single-payer health system, generating longitudinal patient data of a scale and continuity that no private insurer or hospital network can replicate. By any rational measure, this should be fertile ground for a thriving domestic health technology industry. It is not.
What has emerged instead is a widening gap between invention and application. British firms are producing the research; American corporations are winning the contracts.
The NHS's patient data represents decades of unbroken clinical records across an entire national population. For AI developers building diagnostic tools, predictive models, or drug discovery platforms, access to that data is enormously valuable. But gaining lawful, structured access requires navigating a governance framework that demands substantial legal resource, information governance expertise, and, increasingly, the kind of sustained institutional relationships that only large organisations can maintain. The result is that companies with the capacity to clear those hurdles tend to be the same ones already embedded in the system: Microsoft, Oracle, Palantir.
Smaller British firms face a separate, slower problem. The NHS procurement environment has produced what industry observers now refer to as "pilot-itis": a cycle in which a startup secures a local clinical trial or an Integrated Care Board pilot, demonstrates clinical value, and then finds no clear route to national adoption. These pilots rarely convert into scalable contracts. Without scale, venture capital dries up. Without capital, the company cannot grow. The cycle repeats until the firm either folds or relocates.
Several have relocated. UK-founded medical AI companies have moved their headquarters to the United States or Germany in recent years, citing more accessible regulatory pathways, clearer reimbursement frameworks, and procurement processes that do not require years of localised piloting before a product reaches patients at any meaningful volume. The economic logic is straightforward. If the domestic market cannot absorb the product, the company goes where it can.
The competitive position of incumbent technology suppliers compounds this. NHS Trusts and Integrated Care Boards are under significant operational pressure, and when choosing digital systems, they tend towards suppliers they already know. Legacy vendors have responded by bundling AI functionality into existing enterprise contracts, offering capabilities at effectively zero marginal cost to the Trust. A British AI firm with a genuinely superior product still has to justify switching costs, procurement cycles, and clinical validation requirements that its bundled competitor has already bypassed. The playing field is not level; it is structurally tilted.
Policy discussions have begun to acknowledge this. Some health technology advisers and industry bodies have called for a reserved proportion of NHS digital procurement budgets to be directed towards domestic innovators, similar to small business set-aside provisions used in US federal contracting. Others have pointed to the Medicines and Healthcare products Regulatory Agency as a bottleneck, arguing that its approval timelines for AI software as a medical device remain too slow relative to what companies face in the United States under FDA breakthrough device designations.
Neither reform has materialised at any meaningful scale. The government has published strategies, convened taskforces, and expressed ambition. Execution has not followed.
The consequence of continued inaction is not abstract. The UK will have funded the basic science, trained the researchers, and provided the patient data that made the products possible. It will then license those products back from foreign owners at commercial rates, having derived no industrial or economic benefit from the underlying asset. This is not a theoretical risk. For several clinical AI applications already deployed in NHS settings, it is the current reality.
The question for government is whether the United Kingdom intends to build a health technology industry or simply to remain a well-resourced testing ground for one built elsewhere. Warm words about innovation will not settle it. Procurement reform, regulatory pace, and a willingness to back domestic firms before they leave will.
