Hyperbolic Launches Forge, an Infrastructure Layer for Reliable GPU Provisioning
June 10, 2026 | San Francisco, California
Hyperbolic Labs has introduced Forge, an infrastructure layer that turns fragmented GPU supply from providers around the world into reliable, ready-to-use compute. Announced on June 10, Forge manages the full machine lifecycle — provisioning, configuration, security hardening, monitoring, and post-run sanitization — so that customers get one consistent experience regardless of which data center supplies the underlying hardware.
The first product built on Forge is bare metal provisioning and instance management for on-demand GPU compute. Hyperbolic designed the layer to support more of its platform over time, including inference, managed Kubernetes, and Slurm. The company frames the core problem as ownership of the full lifecycle: rather than handing a customer a raw node from a third-party provider, Forge prepares, secures, and recycles each machine so that reliability does not depend on any single supplier’s automation, security controls, or recycling processes.
Hyperbolic is part of PVP’s portfolio of early-stage AI infrastructure companies. Forge extends the company’s open-access GPU marketplace from access alone toward dependable, production-grade compute, addressing the operational gap that often separates available hardware from infrastructure that teams can actually run serious workloads on.