IBM Corp. has signed a $240 million infrastructure deal with artificial intelligence startup Together AI Inc. to provide infrastructure for the company’s cloud platform. The partnership comes after Together AI raised $800 million in funding from a consortium that included Nvidia Corp., whose chips IBM will use to run Together AI’s workloads.
The San Francisco-based startup operates a cloud platform optimized for machine learning, which processes about 400 trillion tokens per month for customers. Developers can use the platform to train custom AI models, fine-tune existing ones, and perform inference.
The infrastructure deal will see IBM deploy a “large cluster” of HGX B300 systems for Together AI. The hardware will become available through IBM Cloud in the first half of 2027. The HGX B300 is a motherboard equipped with eight Blackwell Ultra chips, which can provide 15 petaflops of performance when processing data in the NVFP4 format.
Related: Hackers defend small water utilities online
The Blackwell Ultra comprises two chiplets linked together by a custom interconnect, hosting 160 computing modules called streaming multiprocessors. Each streaming multiprocessor comprises more than 130 cores and contains memory optimized to store intermediate results.
The HGX B300 uses eight ConnectX-8 SuperNIC chips to move data between its Blackwell Ultra accelerators. A BlueField-3 data processing unit coordinates traffic to and from external systems. According to the filing, the machines include at least two terabytes of memory and a pair of central processing units with 56 cores apiece.
Together AI will use IBM’s HGX B300 cluster to host workloads for customers of its public cloud, particularly running inference workloads that use open-source AI models. The AI startup offers five different inference services as part of its platform, including one optimized for media generation models and another that provides serverless infrastructure.
In the middle of this development, the partnership compares to other collaborations in the field, where companies are increasingly looking to optimize their cloud infrastructure for AI workloads. This trend is driven by the growing demand for AI-powered services and the need for scalable and efficient infrastructure to support them.
Related: Texas data center raises emission concerns
Together AI Chief Executive Vipul Ved Prakash said, “This cluster lets us bring production-grade inference to more companies, faster, and it’s a big step in our push to make open-source AI the obvious choice for enterprises.”
The partnership between IBM and Together AI is expected to benefit both companies, with IBM providing the necessary infrastructure to support Together AI’s growing customer base. The deal also highlights the growing importance of AI-optimized cloud infrastructure and the need for companies to invest in scalable and efficient solutions.
They are well-positioned to take advantage of this trend. With the deployment of the HGX B300 cluster, Together AI will be able to offer its customers faster and more efficient AI services, which is expected to drive growth and adoption in the industry.