IBM, Together AI ink $240 million deal for Nvidia-powered AI inference cluster
By Anhata Rooprai
Aug 11 (Reuters) - IBM and startup Together AI have signed a $240 million multi-year agreement to build a large-scale artificial intelligence cluster on IBM Cloud using Nvidia systems, the companies said on Tuesday.
The cluster will provide inference for open-source AI models, which have gained traction as businesses seek to rein in AI costs and weigh concerns about cybersecurity incidents involving models from Anthropic, OpenAI and Meta.
Inference, the process of running trained AI models to generate responses, has become one of the largest drivers of demand for computing capacity, prompting cloud providers and chipmakers to spend billions of dollars expanding AI infrastructure.
The cluster on IBM Cloud will use Nvidia's HGX B300 systems, which link the chipmaker's newer Blackwell processors, and its Spectrum-X Ethernet networking gear. Nvidia has said the Blackwell chips are optimized for AI inference.
The initial cluster will feature about 2,000 Nvidia Blackwell 300 chips and will be located in the U.S., Together AI's chief revenue officer Kai Mak told Reuters in an interview.
"We think this will be sold out at least two to three months ahead of time," Mak added. "We'll have full offtake well before it's ready for service."
San Francisco-based Together AI's platform lets companies train and run AI workloads on open models such as DeepSeek, MiniMax and Kimi at lower costs than closed systems. It was last valued at $8.3 billion in July.
(Reporting by Anhata Rooprai in Bengaluru; Editing by Tasim Zahid)