News about open source

The open-source llm-d project has released version 1.0, a hardened core for distributed AI inference designed to help organizations serve and scale large language models more efficiently. New features include multi-tenant quality-of-service controls, zero-downtime operations with high-availability routing, and session-aware routing for agentic AI workloads. Founded in 2025 by Red Hat, IBM, Google Cloud, CoreWeave and NVIDIA and now governed by the Cloud Native Computing Foundation, the project has more than 500 contributors from over 25 organizations. Recent demonstrations showed cost reductions of five to 10 times for context-heavy agentic workloads and three times greater output throughput, while IBM Research reported 11.5 times lower inference cost than an all-proprietary configuration.