IBM Teams with Together AI on $240M Open-Source AI Inference Deal

IBM and Together AI have signed a multi-year agreement to deploy open-source AI inference on IBM Cloud, using NVIDIA AI infrastructure. The deal is worth $240 million and will enable Together AI to provide better performance and token economics for enterprises scaling their AI deployments.

The collaboration will utilize a large cluster of NVIDIA HGX B300 systems on IBM Cloud, which is expected to be available in Q1 2027. This deployment is the first dedicated, large-scale cluster built for inference on IBM Cloud using these systems.

Together AI's CEO, Vipul Ved Prakash, stated that enterprises want high-performance models without the closed-model price tag. The company has seen significant momentum for its inference product, which now serves 400 trillion tokens monthly.