Skip to content

Taking frontier compute to the data with federated learning

20 November 20266 min read
Guest Insights
Taking frontier compute to the data with federated learning

FLock.io’s Head of EMEA, Tiffany Wang, offers a perspective on frontier compute from a company that relies on compute to train its models. Although the UK is home to a large share of the world’s data centres, the vast majority are owned by US hyperscalers. For nations and enterprises to truly lead in frontier AI, next-generation hardware alone won’t suffice if we’re still running models trained in Silicon Valley.

Without change, the UK will remain far from owning its own models, and the country’s most valuable data will stay trapped inside NHS trusts, local councils and other regulated institutions.

By combining frontier compute with federated learning, however, we can put that data to work without moving the raw data itself, keeping it secure while collaboratively building the future of sovereign AI.

Watch the video below to hear Tiffany explain this in her own words.

techUK FLock frontier compute

techUK