Federated learning platform that trains models at the edge without moving raw data

Scaleout was founded in 2018 as a spin-out from Uppsala University in Sweden, with roots in research on large-scale distributed systems and scientific computing. Its founding group includes Andreas Hellander, Salman Toor, Daniel Zakrisson, Ebba Kraemer, Jens Frid, Morgan Ekmefjord, and Ola Spjuth, and the team of roughly twenty includes several PhDs and academic staff. The company raised about $1.5M in late 2023 from Navigare Ventures, Almi Invest, and Uppsala University Invest, and was named to Sweden's 33-listan innovation ranking. Its open-source FEDn framework, described in published work on scalable federated learning, is the technical foundation the commercial platform is built on.

The commercial positioning is sovereign edge AI, aimed at data that is legally, physically, or operationally immovable: medical images, vehicle sensor feeds, and defence sensor data. Rather than shipping a finished model to a device and letting it drift, Scaleout keeps the training loop running in the field, so nodes adapt to local conditions and propagate what they learn to the fleet as encrypted weight updates within the same operational window. A node without an uplink keeps inferring, capturing, and training, then synchronizes on reconnect. The company's Tactical CV Network applies this to computer vision for counter-drone and ISR work, and Scaleout has been selected as a NATO DIANA innovator and demonstrated autonomous drones with BAE Systems in the Arctic.