Community Trust ScoreVerified
FLock.io just got a serious nod from the World Economic Forum. The WEF’s MINDS programme picked the company for its privacy-preserving AI work — specifically what it’s doing inside the UK’s National Health Service.
Two NHS trusts are already running on FLock.io’s federated learning platform: Moorfields Eye Hospital and University College London Hospitals. Moorfields is using it for eye disease detection. UCLH is focused on diabetes management, working with NHS researchers at UCL and clinical teams to build glucose monitoring alerts trained on data from over 400 patients. The models train locally. No raw data leaves the building. Partners across the UK, Europe, the US, and China can all contribute without ever handing over patient records — which, in healthcare, is pretty much the whole game.
Around 14,000 users are on the platform right now, including patients in diabetes management apps across the UK, Southeast Asia, and East Asia.
The £100 Million Savings Argument
FLock.io puts a number on what this could mean for the NHS: over £100 million saved annually through AI-driven diabetes prevention. The math they use is a 1% decrease in spending on the disease, which currently runs above £10 billion a year for the NHS. Whether that projection holds up in practice is unclear yet, but the scale of the NHS — a single-payer system with consistent data governance across trusts — makes it probably the best possible testing ground for federated learning at this size.
At Moorfields, the federated setup tackles a problem centralized AI can’t really handle: multi-site collaboration on sensitive imaging data. Each trust trains on its own image data. Encrypted updates get aggregated. No imaging data gets shared. The goal is to roll those models out across more NHS trusts, using the NHS’s structure as a kind of natural infrastructure for scalable federated learning. It’s a smart fit.
The Numbers Behind FLock.io’s Platform
FLock.io’s enterprise federated learning stack — which includes blockchain verification — claims some striking performance figures. Model accuracy up 37%. Costs down 44%. Deployment times cut by 63%. Training energy usage reduced by 80%. Those numbers come from the company, so take them with appropriate skepticism, but even directionally they’re the kind of figures that get procurement teams interested.
The platform is built to be production-ready. FLock.io says organizations can deploy custom AI models efficiently through its setup, which it calls AI Arena and FL Alliance. The pitch is lower total cost of ownership and faster deployment — two things hospitals and banks care about a lot, given how badly they tend to get burned by slow, over-budget enterprise software rollouts.
And banks are part of the pitch too. FLock.io positions federated learning as a fix for a specific problem in regulated industries: compliance risk pushes institutions toward generic, centralized AI models that don’t perform as well. Federated learning lets them train on their own data, locally, without the legal exposure that comes from pooling sensitive records somewhere central. Hospitals face the same pressure. It’s basically the same compliance headache, different data type.
Malaysia Pilot and International Expansion
Beyond the NHS, FLock.io is running a sovereign AI pilot with the government of Sarawak in Malaysia, focused on healthcare. That pilot is set to expand to hospitals in the US, Europe, and China. No firm timeline was given on when that expansion kicks off.
The international angle matters. Cross-border AI collaboration in healthcare is genuinely hard — data residency laws, different regulatory regimes, patient privacy rules that vary country by country. Federated learning is probably the cleanest technical answer to that problem right now, because the data never has to cross a border. Only the encrypted model updates do.
FLock.io’s approach also cuts down on two risks that tend to haunt centralized AI systems: data breaches and model poisoning attacks. When data stays local and only encrypted aggregations move through the network, the attack surface shrinks considerably.
The WEF recognition through MINDS puts FLock.io in front of a different audience than the usual crypto or AI startup circuit. It’s a signal the company is pushing hard into institutional and government markets — not just healthcare, but anywhere data sovereignty is non-negotiable.
FLock.io forecasts the Sarawak pilot will set a precedent for international healthcare AI collaboration, with rollouts planned across hospitals in the US, Europe, and China.
Frequently Asked Questions
Which NHS hospitals are using FLock.io’s technology?
Moorfields Eye Hospital uses FLock.io for eye disease detection, while University College London Hospitals uses it for diabetes management and glucose monitoring alerts trained on data from over 400 patients.
What savings does FLock.io project for the NHS?
FLock.io projects over £100 million in annual NHS savings through AI-driven diabetes prevention, based on a 1% reduction in the more than £10 billion the NHS currently spends managing the disease each year.
