IBM and Together AI Forge Groundbreaking Collaboration to Expand Open-Source AI Capabilities on IBM Cloud

A New Era in AI Infrastructure

In a groundbreaking move, IBM has formed a strategic collaboration with Together AI to enhance the capabilities of open-source AI applications via NVIDIA's infrastructure on the IBM Cloud. This partnership, formalized in a multi-year agreement worth $240 million, aims to leverage the cutting-edge NVIDIA HGX B300 systems, marking the creation of the first extensively dedicated cluster designed for AI inference on IBM Cloud.

This innovative infrastructure is expected to be operational by the first quarter of 2027 and will allow enterprises to efficiently scale their AI operations, providing robust support for fast-paced business demands. According to sources from NVIDIA, this new architecture functions to provide thirty times more output in terms of AI factory capabilities compared to previous generations.

Together AI is at the forefront of promoting open-source AI, and this collaboration reinforces their mission. The very foundation of Together AI's philosophy lies in the belief that developers should have accessible tools to construct with open, modular stacks. Although originally founded in 2022, the company has rapidly advanced, recently finalizing an $800 million Series C funding round, bringing its valuation to an impressive $8.3 billion. The proceeds will aid in expanding its AI Native Cloud, which encompasses a range of services from model inference to training, fine-tuning, and automated agent workflows. This significant investment reflects their growing traction in delivering AI capabilities, now serving over 400 trillion tokens each month.

The decision to partner with IBM and NVIDIA does not come lightly; it is rooted in the impressive innovation trajectories of both IBM and NVIDIA, alongside their ability to fulfill the surging demand for GPU capacity. Such a rapid scale is vital for companies aiming to leverage AI while keeping operational costs in check. The strength of IBM’s enterprise-grade cloud services coupled with NVIDIA's GPU capabilities aims to make open-source AI a staple in the toolkit of developers and enterprises.

Vipul Ved Prakash, the CEO of Together AI, stated, "Enterprises want top-tier models without the excessive costs associated with proprietary solutions, and that requires underlying infrastructure that is both efficient and reliable at scale." This collaboration offers a pathway to achieving those needs, enabling production-level inference that businesses can utilize more readily.

On the flip side, Alan Peacock, the General Manager of IBM Cloud, also emphasized the urgency of enterprises to engage with agentic AI to foster real business outcomes. He remarked, "IBM and NVIDIA are creating scalable and economical infrastructure shaped for enterprise requirements, which will propel Together AI's innovation drive for the next generation of AI infrastructure."

In essence, the introduction of AI factories—comparable now to electrical and telecommunications infrastructures—is rapidly transforming how enterprises can convert data into actionable intelligence. Dion Harris, NVIDIA’s Senior Director of HPC and AI Infrastructure Solutions, noted that with NVIDIA's recent technology integrated into IBM Cloud's framework, they will provide an accelerated platform to deploy open-source AI.

This partnership illustrates a continuing commitment between IBM and NVIDIA to not only uplift AI capabilities through infrastructure but also ensure that AI remains accessible for organizations of all sizes. The combination of hybrid ecosystems powered by AI, cloud solutions, and high-performance computing tools serves as a sturdy foundation for advancing AI technologies. Together, with substantial recent strides in GPU-driven analytics and consulting services, IBM and NVIDIA are set on revolutionizing AI applications across various sectors.

For those intrigued by the possibilities this partnership might unfold, further details on the collaboration can be discovered at IBM's official website.

In conclusion, the future of AI infrastructure is promising as it steadily evolves with partnerships like this, making foundational technologies more accessible and efficient for enterprises worldwide.