Red Hat AI: What’s new and what’s next, from production inference to autonomous agents

Red Hat Summit 2025 unveiled our enterprise AI vision, including distributed inference with llm-d, an expanded model ecosystem, and Llama Stack and Model Context Protocol (MCP) for agentic AI, culminating in Red Hat AI 3. One year later, the focus shifts to enterprise-scale production and integrated intelligent systems, showcasing customer success in finance, telecommunications, and industrial automation. Join us for a comprehensive update and roadmap, diving deep into: - Distributed inferencing at scale, with llm-d for cost-effective, low-latency inference. - Enterprise Model-as-a-Service (MaaS) for self-service deployment and governance - Production retrieval-augmented generation (RAG) for reliable, scalable, auditable pipelines. - Continual alignment for model customization beyond fine-tuning. - Accuracy with inference time scaling (ITS) to boost model accuracy without retraining. - The evolution of Llama Stack for multi-agent collaboration and advanced tool use. - Our comprehensive approach to security, governance, and trust with security-focused AI/ML lifecycle, guardrails, and sovereign AI. This session highlights what's new and what's next with Red Hat AI, and how it provides an integrated, open foundation for AI applications with real-world results.