Acts of Evolution

BRIEFINGS · Cloud and AI compute providers

IBM and Together AI sign a $240 million deal for an inference cluster due in early 2027

IBM said on 11 August it has a $240 million multi-year deal to build an NVIDIA B300 inference cluster for Together AI, due in Q1 2027.

Event date: · Published

On 11 August 2026 IBM announced a multi-year $240 million agreement with Together AI. IBM said it is positioned to deploy a large cluster of NVIDIA HGX B300 systems on IBM Cloud, expected to be available in the first quarter of 2027, and that Together AI will use it to run inference for open-source models. IBM called it the first dedicated, large-scale inference cluster on IBM Cloud using those systems. IBM's release says Together AI recently raised $800 million at an $8.3 billion valuation and reports its inference service handles 400 trillion tokens a month. We file this because it is a priced capacity agreement for AI computing. How we read it: Together AI is securing outside capacity to meet demand for open-model inference. What would change this: a delay to the 2027 delivery. Why it matters to you: the companies serving AI models are locking in chips years ahead, which shapes how much AI service can be offered.

What this doesn't prove

IBM's press release. The $240 million figure and schedule are IBM's statements. Agreement length, GPU count and megawatt size are not given. The token-volume and funding figures are as reported in the IBM release.

Next fixture

  • Cluster delivery in Q1 2027
  • Further capacity agreements
  • Agreement term

Sources

Every figure above comes from these sources. Every briefing is checked against its sources before it publishes. Nothing here is a recommendation to buy or sell anything, and this is not investment advice. See the disclosures.

On the map

Together AI

Open the map