IBM and AI accelerator company Groq announced a strategic technological and commercial partnership on October 20, 2025. The goal is to provide enterprise customers with ultra-fast and cost-effective inference (execution) for AI agents. As part of the agreement, GroqCloud, powered by its unique LPU (Language Processing Unit) chips, will be integrated with IBMs watsonx Orchestrate platform. This will allow AI agents built in watsonx to perform tasks with minimal latency. The solution supports popular open-source libraries like vLLM and is designed to solve a major enterprise challenge: the high cost and low speed of AI when scaling from pilot projects to full production.
IBM and Groq Announce Partnership to Accelerate AI Inference in the Enterprise
