IBM and Together AI have inked a multiyear agreement worth $240 million, with IBM set to roll out a massive NVIDIA-based AI computing cluster on IBM Cloud. The cluster, set to go live in the first quarter of 2027, will utilize NVIDIA HGX B300 systems connected via NVIDIA Spectrum-X Ethernet networking. It will cater to Together AI’s inference workloads for open-source AI models.
The initial deployment will feature approximately 2,000 NVIDIA Blackwell 300 GPUs and will be situated in the US, as per Together AI chief revenue officer Kai Mak. The capacity is anticipated to be fully booked two to three months before its launch. This deployment marks IBM Cloud’s first dedicated large-scale inference cluster built around HGX B300 systems, known for delivering significantly higher AI performance than previous generations.
Inference, where trained models process requests and produce outputs, has emerged as a key driver of demand for computing power, leading cloud providers and chipmakers to expand AI infrastructure. Together AI specializes in offering infrastructure and software for various AI workloads, including inference, training, fine-tuning, and agent-based tasks. The company recently secured $800 million in funding and has raised commitments for over 500 MW of compute capacity to support its growing infrastructure needs.
The IBM-NVIDIA collaboration aims to deliver scalable, cost-effective AI infrastructure to accelerate innovation in the realm of AI. The joint effort seeks to make open-source AI the go-to choice for enterprises. This partnership underscores the importance of robust, high-performance infrastructure to support the adoption of cutting-edge AI models.
The deployment of the NVIDIA-powered cluster on IBM Cloud is a testament to the evolving landscape of AI infrastructure. As more companies look to leverage AI technologies for business growth, partnerships like the one between IBM and NVIDIA become instrumental in driving real-world outcomes through AI at scale.
