Expanding our enterprise inference capacity with IBM Cloud and NVIDIA

Together AI says it is working with IBM and NVIDIA to scale enterprise AI inference on IBM Cloud, starting with a large cluster of NVIDIA B300 GPUs connected by NVIDIA Spectrum-X Ethernet networking. The company calls it the first dedicated, large-scale inference cluster of its kind on IBM Cloud, and says it is the first customer using it.

The move fits Together’s bigger pitch: open models for enterprise use, not closed systems controlled by a few labs. The company says it already serves hundreds of trillions of tokens each month to more than a million developers, and that demand is still climbing. That scale is the real story here. Once usage gets that big, inference stops being a side problem and becomes the product.

Together argues that enterprises and AI-native companies keep choosing open models for two reasons: their data stays sovereign, and performance can land at a fraction of closed-model cost. The company says the economics improve as usage rises. That is the case it is making to buyers, not just to the open-source crowd.

The partnership also says something about where the market is going. Open-source AI, in Together’s telling, cannot just be clever and cheap; it has to run everywhere, fast, reliably, and with the guardrails businesses expect. This IBM Cloud cluster is one more attempt to prove that point at serious scale.