Red Hat announced the release of Red Hat AI 3 on October 17, 2025, a new version of its enterprise platform for AI development and deployment. A key innovation is the support for distributed inference through the llm-d module. This technology allows a single large language model to be "split" into parts and executed in parallel across multiple servers or GPUs. This approach solves the problem of running giant models that do not fit into the memory of a single accelerator and makes the inference process more scalable and cost-effective. The platform also lays the groundwork for creating agentic systems by offering a unified API for model interaction. Red Hat AI 3 is aimed at companies moving from AI experimentation to full-scale production use, who need a reliable and manageable solution.
Red Hat Introduces Red Hat AI 3 Platform with Distributed Inference Support
