IBM and Together AI have teamed up to launch the first significant inference cluster on IBM Cloud, utilizing NVIDIA’s HGX B300 systems. Announced on August 11, 2026, in a $240 million multi-year agreement, this initiative is expected to enhance how enterprises manage AI workloads and is set to go live in the first quarter of 2027.
The new cluster will support Together AI in providing open-source model inference at scale, marking it as the first large-scale deployment dedicated to this purpose on IBM Cloud. Featuring NVIDIA Spectrum-X Ethernet networking, this infrastructure promises an impressive 30-fold increase in AI production efficiency compared to earlier versions, according to NVIDIA.
This strategic partnership aims to enable Together AI to enhance performance and reduce costs for businesses looking to scale their AI implementations effectively. Emphasizing the importance of open-source models, Together AI advocates for a development environment built on accessible and modular AI technology. The company recently secured $800 million in Series C funding, valuing it at $8.3 billion, allowing it to broaden its AI Native Cloud capabilities across inference, training, fine-tuning, and automated workflows. It has already experienced rapid growth, processing approximately 400 trillion tokens each month through its inference offerings.
Together AI's decision to collaborate with IBM and NVIDIA stems from their cutting-edge product strategies and their proven ability to provide the GPU capacity necessary for rapid AI scaling and cost efficiency. This partnership leverages IBM's strength in enterprise cloud solutions, aiming to facilitate Together AI's growth within the enterprise sector while democratizing access to open-source AI for developers worldwide.
"Businesses want the capabilities of advanced frontier models without the prohibitive costs associated with proprietary solutions. Achieving this requires a robust and scalable infrastructure," stated Vipul Ved Prakash, CEO of Together AI. "Partnering with IBM and NVIDIA equips us with that critical support, allowing us to deliver production-ready inference solutions more swiftly to a broader range of companies. This is a significant advancement in making open-source AI the preferred choice for enterprises."
Alan Peacock, General Manager of IBM Cloud, remarked, "Companies are racing to implement scalable agentic AI to realize tangible business benefits. IBM and NVIDIA are collaborating to provide cost-effective, scalable AI infrastructure that enables Together AI to drive innovation and shape the next generation of AI technology."
Dion Harris, Senior Director at NVIDIA, noted the vital role of AI factories in modern enterprises, describing them as fundamental infrastructures akin to utilities like electricity and telecommunications. The collaboration between IBM and Together AI, backed by NVIDIA's advanced resources, aims at delivering a high-performance computing environment that will empower organizations to deploy open-source AI solutions with the necessary efficiency and speed for real-time applications.
This development exemplifies the ongoing partnership between IBM and NVIDIA as they work to innovate and improve AI capabilities across various sectors. The combination of IBM Cloud's hybrid environment, powered by NVIDIA GPUs and integrated with Together AI's inference platform, provides businesses a steadfast foundation to design, launch, and scale AI systems. Recent announcements also highlight advancements in GPU-native data analytics and cloud infrastructure designed to support organizational AI initiatives effectively.
For more insight into the IBM and NVIDIA collaboration, further information can be found on IBM’s website.
IBM continues to be a front-runner in global hybrid cloud and AI solutions, aiding clients across more than 175 nations in transforming their data into actionable insights, optimizing processes, and gaining competitive advantages. With a focus on trust, inclusivity, and service, IBM remains committed to innovation in AI and cloud offerings.
Founded in 2022, Together AI positions itself as the AI Native Cloud, integrating cutting-edge open-source models with high-performance infrastructure. Serving more than a million developers and managing some of the most complex AI tasks in the industry, Together AI is enhancing the landscape of production-scale inference and related technologies for the next generation of AI companies.
