Inside Fujitsu & Nvidia’s Growing AI Infrastructure Alliance

The infrastructure requirements for running AI at scale have pushed technology companies to forge partnerships that combine different types of computing hardware. 

Fujitsu, the Japanese information technology services and equipment company, has now expanded its collaboration with Nvidia, to co-develop infrastructure that integrates central processing units with graphics processing units for enterprise AI deployment.

The partnership will see Fujitsu combine its Monaka CPU series with Nvidia GPUs through a connection technology called NVLink Fusion, which links the two types of processors to reduce the time it takes for data to move between them. 

The companies plan to target sectors including healthcare, manufacturing and robotics with platforms designed to run AI agents, which are software systems that can perform tasks autonomously.

The partnership: Explained

“Fujitsu’s strategic collaboration with Nvidia will accelerate AI-driven business transformation in enterprise and government sectors,” says Takahito Tokita, Representative Director and CEO of Fujitsu.

“By combining the cutting-edge technologies of both companies, we will develop and provide full-stack AI infrastructure, starting with sectors such as manufacturing where Japan is a global leader.”

The collaboration builds on work Fujitsu has done in supercomputing and enterprise systems in Japan. 

The company operates data centres and provides cloud infrastructure services to businesses and government organisations. 

By integrating CPUs and GPUs at the data centre level, Fujitsu aims to offer clients a platform that can handle the computational demands of training large AI models – while also running inference workloads, which involve using trained models to generate predictions or responses.

Fujitsu Monaka CPUs connecting with Nvidia GPUs via NVLink

NVLink Fusion is a technical approach to solving bandwidth constraints that emerge when CPUs and GPUs need to share data during AI workloads. 

Traditional connections between the two processor types can create bottlenecks that slow down performance. 

By using NVLink, which Nvidia developed as a high-speed interconnect, the combined system aims to increase throughput and support larger-scale model training.

Jensen Huang, Founder and CEO of Nvidia, frames the partnership in terms of infrastructure requirements for AI deployment: “The AI industrial revolution has begun and we must build the infrastructure to power it – in Japan and across the globe,” he says.

“Fujitsu is a true pioneer in computing and Japan’s trusted leader in supercomputing, quantum research and enterprise systems. Together, Nvidia and Fujitsu are connecting and extending our ecosystems to forge a powerful partnership for the era of AI.”

The companies are working on three areas. 

The first involves co-developing a platform for AI agents built on Fujitsu Kozuchi, the company’s AI platform and Nvidia Dynamo, which orchestrates AI workloads. 

These agents will be customised for specific industries using Nvidia NeMo, a framework for building language models and Fujitsu’s Takane AI model. 

The agents will then be packaged as Nvidia NIM microservices, a format designed to make it easier for customers to deploy AI applications.

The second area focuses on the hardware integration itself, combining Monaka CPUs with Nvidia GPUs to create what the companies describe as silicon-level optimisation. 

This involves ensuring that Fujitsu’s ARM-based processor architecture works efficiently with Nvidia’s CUDA programming environment, which developers use to write software that runs on Nvidia GPUs.

The partner ecosystem to support deployment across Japanese industries

The third area involves building a partner ecosystem to drive adoption.

Fujitsu and Nvidia plan to work with other companies to develop use cases for industrial automation and robotics, starting with Japanese industries before expanding to other markets. 

Fujitsu will use its existing data centre and cloud infrastructure to support the rollout.

Takahito adds: “To further support the expanding needs of AI infrastructure, Fujitsu and Nvidia will expand this partnership in the areas of high-performance computing and quantum.”