In a converted warehouse near King's Cross, engineers at Isomorphic Labs spent the spring building on Google DeepMind's protein-folding breakthroughs, refining a system designed to compress years of drug discovery into months. Nine miles away, a maternity unit was still working through the recommendations of a safety review published the previous year, its staff stretched across wards that had not seen equivalent investment in a generation. Both are part of the same country's health story in 2026, though only one of them appears in the funding tables that Tracxn published this month, showing UK artificial intelligence companies raising $9.6bn in the first half of the year, a 360 per cent jump on the same period in 2025.

The headline number invites a simple reading, that Britain has become an AI powerhouse and that healthcare, along with everything else, stands to benefit. The detail complicates that story considerably. Five companies accounted for 84 per cent of the total raised, and London alone captured 98 per cent of it, up from 93 per cent in the second half of last year. Oxford, home to some of the country's most established biomedical research infrastructure, managed just over $100m. Cambridge, the other pillar of what used to be called the Golden Triangle, raised under $20m. For an NHS that depends on regional life sciences ecosystems to translate research into deployable clinical tools, that imbalance matters more than the topline figure.

Isomorphic Labs deserves particular attention here, because it is the rare AI company whose commercial proposition runs directly through the NHS's own priorities. Its $2.1bn round in May, backed by Thrive Capital, Alphabet and the UK's newly launched sovereign AI fund, was pitched explicitly at accelerating drug discovery, with partnerships already in place with Novartis, Eli Lilly and Johnson & Johnson. If the platform delivers on its claims, the eventual beneficiaries include NHS patients waiting for treatments that currently take a decade or more to reach clinic. But the distance between a well-funded London drug design engine and a therapy reaching an NHS formulary runs through the MHRA's approval capacity, through NICE's appraisal timelines, and through an NHS procurement system that has struggled for years to move at the pace venture capital now expects of it. Capital can arrive in a single funding round. Regulatory throughput and adoption infrastructure cannot be scaled the same way, and nothing in this quarter's figures suggests that gap has narrowed.

There is a second, quieter concern in the geography. Health innovation has historically depended on proximity between university research, teaching hospitals and industry, which is why Oxford and Cambridge built their reputations in the first place. If AI investment continues concentrating almost exclusively in London, driven by a handful of hyperscale bets like Nscale's infrastructure round and Wayve's autonomous vehicle programme, the risk is not that life sciences AI stops advancing but that its benefits arrive filtered through companies with little operational connection to the NHS trusts and regional health systems that would need to adopt their tools. A drug discovery breakthrough developed in King's Cross still has to survive contact with an NHS estate that in many regions cannot reliably fund basic digital infrastructure, let alone integrate AI-designed therapeutics into routine care.

None of this argues against the investment itself. British AI attracting global capital at this scale is, on balance, a genuine achievement, and Isomorphic Labs in particular represents exactly the kind of company that ministers cite when they talk about life sciences as a growth sector. But the figures published this week describe a financial story more than a health system one. Until MHRA capacity, NICE appraisal speed and NHS procurement keep pace with the money flowing into companies like Isomorphic Labs, the NHS risks watching from the sidelines as Britain becomes an AI investment success story that its own health service is among the last to feel.