The Investor
11 Aug 2026, 16:12
IBM Expands AI Infrastructure Push With $240 Million Together AI Agreement
IBM (NYSE: IBM) announced a multi-year $240 million agreement with Together AI to deploy a large-scale artificial intelligence inference cluster on IBM Cloud, further expanding the company’s exposure to growing enterprise AI infrastructure demand.
Under the agreement, IBM plans to deploy NVIDIA HGX B300 systems combined with NVIDIA Spectrum-X Ethernet networking. The cluster, expected to become available in the first quarter of 2027, will be used by Together AI to provide production-scale inference for open-source AI models.
The deployment will be IBM Cloud’s first dedicated large-scale inference cluster based on HGX B300 systems. NVIDIA says the architecture can deliver as much as 30 times greater AI factory output compared with previous generations.
Together AI has been scaling rapidly as demand for open-source AI models grows. The company says its inference platform currently processes around 400 trillion tokens per month and recently raised $800 million at an $8.3 billion valuation.
For IBM, the agreement strengthens its position as an infrastructure provider for increasingly compute-intensive AI workloads. It also deepens IBM’s existing relationship with NVIDIA, spanning GPUs, networking, cloud infrastructure and enterprise AI software.
The $240 million multi-year commitment provides IBM with another significant AI infrastructure customer while demonstrating demand for its GPU-based cloud capacity. The companies expect the platform to help enterprises run open-source AI models with improved performance and lower inference costs.
IBM (NYSE: IBM) announced a multi-year $240 million agreement with Together AI to deploy a large-scale artificial intelligence inference cluster on IBM Cloud, further expanding the company’s exposure to growing enterprise AI infrastructure demand.
Under the agreement, IBM plans to deploy NVIDIA HGX B300 systems combined with NVIDIA Spectrum-X Ethernet networking. The cluster, expected to become available in the first quarter of 2027, will be used by Together AI to provide production-scale inference for open-source AI models.
The deployment will be IBM Cloud’s first dedicated large-scale inference cluster based on HGX B300 systems. NVIDIA says the architecture can deliver as much as 30 times greater AI factory output compared with previous generations.
Together AI has been scaling rapidly as demand for open-source AI models grows. The company says its inference platform currently processes around 400 trillion tokens per month and recently raised $800 million at an $8.3 billion valuation.
For IBM, the agreement strengthens its position as an infrastructure provider for increasingly compute-intensive AI workloads. It also deepens IBM’s existing relationship with NVIDIA, spanning GPUs, networking, cloud infrastructure and enterprise AI software.
The $240 million multi-year commitment provides IBM with another significant AI infrastructure customer while demonstrating demand for its GPU-based cloud capacity. The companies expect the platform to help enterprises run open-source AI models with improved performance and lower inference costs.
