Serverless GPU/CPU runtime targeted at AI workloads. Closed a 2026 H1 round to scale GPU pool and Sandboxes product.
Open source ↗