The UK has the science. Now it needs the engine room.

In 2024, a Nobel Prize was awarded for work done a few Tube stops from King’s Cross: using AI to predict protein structures. It was a very British breakthrough, born of the UK’s extraordinary depth in computational biology. And it tells us something important. The question is no longer whether AI can accelerate scientific discovery. It is whether the UK will build what is needed to turn that scientific strength into companies, treatments and better care at home. 

I started my career in orthopaedics at Addenbrooke’s Hospital in Cambridge, working on biomarkers for early disease onset. Back then, we just called it machine learning, and the ƒlimiting factor was always the same: time. A professor would spend an entire career on one protein interaction, then hand the work to the next generation. Today, with modern AI infrastructure, researchers can see those ideas through within their own lifetimes. Scientists I speak to describe compressing years of sequential experiments into weeks of parallel ones. 

That acceleration is not theoretical. In London, Basecamp Research trained Eden, a 28-billion-parameter biological foundation model, on our infrastructure in about seven weeks, then used it to design DNA insertions for genetic disease targets and antimicrobial peptides that worked in disease models. A few years ago, that scale of training was the preserve of a handful of technology giants. Now a London startup can do it. This is what levelling the playing field looks like in practice: a ten-person team with the right infrastructure can run science that once required the resources of big pharma. 

So what stands between the UK’s research strength and broader commercial and societal impact? Three things, in my view. 

First, compute. The UK produces world-class models, methods and spinouts, but ambition is still rationed by access to accelerated computing. Drug discovery models that could train in months instead of years or decades only do so if the GPUs, the software stack and the engineering support are actually available to the teams that need them. This is why we are investing in UK infrastructure alongside our data centres across Europe and the US. Think of it like the road network: researchers are building remarkable vehicles, and someone you trust has to build the roads, bridges and service stations so they can step on the accelerator. Infrastructure should be invisible; the impact should be visible. 

Second, data. The NHS holds one of the richest longitudinal health datasets in the world, and the UK’s biobanks and genomics programmes are globally envied. Yet valuable insights remain trapped between institutions, departments and systems. Breaking down those silos, with privacy, sovereignty and clinician oversight designed in from the start, would do more for UK life sciences than almost any other single intervention. Compliance and trust are not obstacles to speed. Speed without trust is meaningless, especially in healthcare. The infrastructure layer can carry much of that burden: traceable, auditable, reproducible workflows, with researchers and institutions owning their data and IP. 

Third, translation. The UK is strong at discovery and less strong at scaling it. Part of the answer is regulatory: clear, risk-based pathways and sandboxes that let startups validate AI-driven approaches without armies of lawyers. Part of it is cultural. I meet brilliant academic teams still working with the technology of ten years ago, and my advice is always the same: start now. The teams that treat AI as a colleague, one you train on your specific problem, will define the next decade. The ones that wait will be left behind. It is the same logic I apply to clinical practice: AI is not going to replace the radiologist, but AI will replace the radiologist who doesn’t use AI. 

None of this is about believing the hype. It is about looking at the results. AI-guided design reaching wet-lab validation. Clinical documentation time cut so nurses can monitor more patients. Foundation models trained in London informing therapies for diseases where every patient carries a different mutation. Companies like Prima Mente are combining neuroscience and AI to study brain health, protect against neurological diseases, and develop early diagnostic on dementia. These are not demos; they are working systems, and many of them are British. 

The prize for getting this right is not only economic, though the economics are compelling. It is a health system where diagnosis comes earlier, where rare diseases are not ignored because development costs are too high, and where treatments are designed for the individual rather than the average. My ambition for the UK is simple: that the country which predicted the protein fold also builds the engine room where the next generation of medicines is designed, tested and brought to patients. 

The science is here. The talent is here. If we match them with accessible infrastructure, connected data and the confidence to move, the UK will not just participate in the AI era of discovery. It will lead it.