CLINICAL AI
1. A surgeon just got a live AI copilot in the operating theatre, and the results held up.
Prof. Hani Marcus and Mr. Danyal Khan, lead surgeons on the world's first AI-assisted brain tumour removal. Source: UCLH / BBC News, August 27, 2026
The signal: Neurosurgeons at University College London Hospitals completed the world's first AI-assisted live brain tumour removal on August 27, with an AI system reading real-time surgical video to flag hidden blood vessels and nerves the surgeon could not directly see.
The details:
Patient Rhys Hibbert, 48, had an 11mm pituitary tumour pressing on his optic nerves. Without surgery, he faced progressive blindness. The tumour sits millimetres from the carotid arteries and optic nerves at the base of the skull, making it one of the most risk-loaded approaches in neurosurgery.
The AI system, developed at the UCL Hawkes Institute and trained on hundreds of surgical videos, runs on an NVIDIA Clara IGX platform. It analysed the live endoscope feed and projected colour-coded overlays showing where critical structures were most likely hiding behind tissue, in real time. Lead surgeons: Prof. Hani Marcus and Mr. Danyal Khan at the National Hospital for Neurology and Neurosurgery.
The trial was funded by the National Institute for Health and Care Research (NIHR) and Google. A larger clinical trial is planned. The average UK neurosurgeon performs 10-20 of these procedures a year; this AI was trained on more surgeries than most surgeons see in a career.
Why it matters: Every autonomous clinical AI story so far has been about things AI does instead of a clinician. This one is about AI working in the room, in real time, as the surgeon's eyes into tissue they cannot directly see. That is a fundamentally different use case, and if the larger trial validates it, the 'AI in the OR' category gets a very clear proof point to build from.
CAPITAL & DEALS
3. $7.4B into digital health in six months. Here is where the money is actually going.
Source: Rock Health. H1 2026 mega deals made up 45% of digital health funding.
The signal: Rock Health's H1 2026 report confirms that digital health venture funding is back in growth mode: $7.4B raised across 244 deals, up $1B versus H1 2025, with AI rewriting where and how capital concentrates.
The details:
Nineteen companies pulled in 20 megadeals (rounds of $100M or more), representing 45% of all capital on just over 8% of deals. Biggest H1 names: WHOOP ($575M), Verily ($300M), Talkiatry ($210M), eMed ($200M), OpenEvidence ($250M), Forus ($160M), Aidoc ($150M). Mental health and weight management were the top two funded clinical areas.
Rock Health's analyst note: 'As AI changes what is easy to build, competitive advantage increasingly comes from qualities that are difficult to replicate.' Founder clinical domain expertise, proprietary workflow access, and white-glove service after the sale are now the durable edges. Pure technical differentiation is eroding as foundation models commoditise the build layer.
Digital health M&A ran at 115 acquisitions through H1, ahead of 2025's pace. No digital health IPOs yet in 2026, but Oura, WHOOP, and Virta Health are named as candidates. GLP-1 adjacents drew three megadeals; wellness peptides (Superpower, Protocole) attracted early seed as investors hunt for the post-GLP-1 body-composition layer.
Why it matters: The concentration of 45% of capital into a small number of megadeals is not a bull market returning equally. It is a maturation signal: late-stage companies that proved their model are getting funded at scale, while early-stage bets face a 'prove you are different in a way AI cannot commoditise' test. Series A HealthTech founders are raising into a fundamentally different market than the megadeal names above.