Flower Labs · Federated learning *is* the direction's core question made into a product — what it costs to send your data somewhere else — and they are the open-source default for it, sitting on four separate documentation roots with no single door in.

Train a model across many devices or organisations without the training data ever leaving them — federated learning, as an open-source Python framework (Flower) plus a hosted grid. Their own homepage now sells something wider than that: "a full-stack AI neolab … building open-source frontier models, agents, and infrastructure" (og:description, read 2026-09-15). The product surface behind that phrase is four named things — Flower SuperGrid (the framework), Flower Hub (app registry, with app verification), Flower Model, Flower Agent — plus FlowerBench (an agent benchmark), an LLM leaderboard, and two of their own models, Lizzy-7B ("a UK-built open frontier LLM for sovereign AI") and Endeavor 1.0.

Sector

Training & MLOps

Stage

Series A

On the roster since

15 Sept 2026