Hopsworks · The feature store that predates the current wave and is now re-aiming itself at agents — a useful control on whether "AI infrastructure" is a new stack or the old data stack under new lighting.

A feature store, grown into what they now call an AI Lakehouse. The core job is the unglamorous one that breaks most production ML: making sure the numbers a model is scored on in production are computed the same way as the numbers it was trained on, and that you can rebuild a training set as it looked at a point in time. Around that sit an online store with sub-millisecond lookups (RonDB), an offline store, feature/training/inference pipelines, a model registry and deployment layer, and — since 5.0 — an agent-facing surface. Open-source core, deployable on AWS, GCP, Azure or your own Kubernetes.

Sector

Training & MLOps

Stage

Series A

On the roster since

15 Sept 2026