IBM said it signed a multiyear US$240 million agreement with Together AI, a San Francisco-based AI cloud startup, on Aug. 11 to scale open-source AI inference on IBM Cloud.

The cluster is expected to be available in the first quarter of 2027.

Under the agreement, IBM could deploy a large cluster of NVIDIA HGX B300 systems and NVIDIA Spectrum-X Ethernet networking on IBM Cloud. Together AI plans to use the cluster for open-source model inference.

Together AI operates an AI cloud for inference, training, fine-tuning, and agentic workflows, and said it supports more than 200 open-source models.

IBM already offers other NVIDIA graphics processing unit instances on its cloud, but said this would be its first dedicated large-scale inference cluster on IBM Cloud using HGX B300 systems.

Together AI recently raised a US$800 million series C round at a US$8.3 billion valuation, the companies said.

🔗 Source: IBM

Recent IBM developments