The IBM $240M Together AI deal will create a large-scale artificial intelligence cluster on IBM Cloud, using Nvidia technology to support the growing demand for enterprise AI inference.
IBM and Together AI have signed a multi-year agreement worth $240 million to develop the infrastructure. The project marks IBM’s first dedicated large-scale inference cluster built with NVIDIA HGX B300 systems and NVIDIA Spectrum-X Ethernet networking.
IBM Targets Large-Scale AI Inference
Inference is the stage where trained AI models process user requests and generate answers, predictions or actions. As companies deploy AI agents for complex and real-time tasks, demand for reliable inference infrastructure has increased.
Read More: AI Startup Discovered Materials Raises $9 Million to Speed Up Chip Material Discovery
Under the agreement, IBM will build the cluster on IBM Cloud. Together AI will use this infrastructure to support its inference platform, which currently processes 400 trillion tokens per month. Tokens are the basic units that AI models use to read and generate text and other information.
Read More: AvenuesAI Q1 Profit Jumps 45% YoY To Rs 85 Cr, Revenue More Than Doubles
The scale of Together AI’s platform highlights the growing need for high-performance computing. AI companies require powerful processors, fast networking and flexible cloud capacity to serve customers without compromising speed or efficiency.
Together AI Focuses On Open-Source Models
Together AI co-founder and CEO Vipul Ved Prakash said the company’s work with IBM and NVIDIA supports its goal of making advanced open-source AI broadly accessible. He pointed to better infrastructure and improved inference economics as important factors in expanding access to these models.
Open-source AI models allow companies and developers to build applications with more control over deployment and customisation. However, operating these models at scale can be expensive. The IBM and Together AI partnership is designed to address that challenge by combining cloud infrastructure with Nvidia’s specialised AI systems.
For enterprises, the focus is not only on developing AI models. Companies also need to run them consistently across business operations. This includes customer service, software development, data analysis and agentic AI applications that can manage multi-step tasks.
IBM And Nvidia Expand Broader Partnership
The new deployment adds to IBM and NVIDIA’s wider collaboration. Their joint initiatives cover hybrid cloud computing, GPU-native data analytics, unstructured data processing and consulting services.
IBM brings its cloud infrastructure, enterprise technology expertise and consulting capabilities to the partnership. NVIDIA provides the HGX B300 systems and Spectrum-X networking technology used for the new cluster. Together AI contributes its inference platform and focus on open-source AI models.
The agreement also reflects IBM’s effort to strengthen its position in the rapidly expanding AI infrastructure market. Large enterprises are increasingly looking for cloud systems that can support demanding AI workloads while offering predictable performance and cost control.
Enterprise AI Demand Sets The Next Stage
The agreement came as enterprise demand accelerated for agentic AI applications. These systems are designed to handle complex tasks, respond to changing information and operate with greater independence than conventional software tools.
Such applications can require continuous inference and fast responses. This makes infrastructure an important part of the overall AI strategy. A capable model alone is not enough if businesses cannot operate it efficiently at scale.
IBM stock rose after the deal was announced, signalling a positive market response. The deployment now moves the partnership from announcements towards infrastructure execution.
In the future, the IBM and Together AI collaboration could help more enterprises access open-source AI models through cloud-based systems. Its progress will depend on the delivery of the Nvidia-powered cluster, the performance of Together AI’s platform and the ability to control inference costs as usage grows.