IBM and Together AI Announce $240M Partnership for Nvidia-Powered AI Inference Cluster

IBM and Together AI have announced a multiyear, $240 million partnership to build a large-scale artificial intelligence inference cluster on IBM Cloud, marking a significant bet on the growing enterprise demand for cost-effective AI deployment. The collaboration, revealed on August 11, 2026, will utilize Nvidia's latest HGX B300 systems, built on the Blackwell architecture, and its SpectrumX Ethernet networking. The cluster is designed to provide inference services for open-source AI models and is scheduled to become operational in the first quarter of 2027.

The partnership positions IBM Cloud as a dedicated infrastructure hub for enterprises seeking to run AI workloads more affordably. Inference, the process of executing queries against a trained AI model, has emerged as a primary driver of computing demand. Together AI, a San Francisco-based startup valued at $8.3 billion as of July, operates a platform that enables companies to train and run workloads on open-source models such as DeepSeek, MiniMax, and Kimi. The company reports it currently processes approximately 400 trillion tokens per month.

"Enterprises want the performance of the best frontier models without the...