Fujitsu, talent, AI hiring staffing solutions

Fujitsu and NVIDIA have extended an alliance to build next-generation IT platforms that are designed for agentic applications.

The two companies plan to build a platform that integrates FUJITSU-MONAKA class CPUs based on Arm processors that are due out in 2027 and graphical processor units (GPUs) and a set of networking ASICs, dubbed NVIDIA NVLink Fusion, to optimally run AI agents.

Additionally, Fujitsu will integrate its AI framework, dubbed Fujitsu Kozuchi, and its Takane AI model with the NVIDIA Dynamo platform for running AI inference engines and the NVIDIA NeMo observability framework and the NVIDIA NIM microservices framework.

The goal is to integrate at the silicon level to optimally run AI agents that have been developed for specific vertical industries, says Fujitsu CEO Takahito Tokita. Those AI models and agents will then self-evolve as AI advances continue to be made, he adds. “We will realize a high-speed AI computing platform by jointly developing infrastructure from the silicon level,” says Tokita.

Initially, Fujitsu plans to make this platform available in Japan before making it more widely available across the various vertical industry segments that Fujitsu currently serves around the world.

Fujitsu is once again showing the world what will come next, says NVIDIA CEO Jenson Huang. “We are entering into a new industrial era,” he says. “The next industrial revolution, artificial intelligence, is the most powerful technology force of our time. Every application will be transformed by AI.”

It’s not clear to what degree IT organizations will need to acquire new infrastructure to optimally run AI agents but The Futurum Group projects AI agents will drive $6 trillion in economic value by 2028. Most existing IT infrastructure platforms were, however, designed to run legacy applications that typically did not require high-speed processors to analyze massive amounts of data. As IT teams deploy next-generation platforms, the issue then becomes determining whether to rely on cloud services to access the infrastructure required or to acquire and deploy next-generation servers that will be deployed in on-premises IT environments.

Regardless of the approach, the cost of running thousands, possibly even millions, of AI agents is going to be substantial. Each organization will need to carefully evaluate the total cost of building and deploying AI agents before committing to deploy them. In the meantime, there is little doubt that most organizations will be aggressively experimenting with AI agents across a wide range of use cases involving everything from customer service to research and development.

The challenge, as always, will be the need to gain access to IT infrastructure in advance of any return on investment (ROI) that might be generated. Hopefully, the more AI agents that are built, the simpler they will become to build and deploy, which should help justify any IT infrastructure investments that are made.

In the meantime, however, organizations may need to come to terms with the simple fact that the total amount of budget dollars that will need to be allocated to IT is about to become substantially higher.

TECHSTRONG TV

Click full-screen to enable volume control
Watch latest episodes and shows

Tech Field Day Events

TECHSTRONG AI PODCAST

SHARE THIS STORY