Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
Jensen Huang, co -founder and chief executive director of Nvidia Corp., speaks during a computational conference in Taipei, Taiwan, Monday, May 19, 2025.
Bloomberg | Bloomberg | Gets the image
Nvidia Jensen CEO Jensen Juan made alerts on Monday and revealed new products aimed at preserving the company at the Center for Development and Computing Artificial Intelligence.
One of the most notable ads was its new Nvlink Fusion program, which will allow customers and partners to use non-Nvidia Central processing and graphic units along with NVIDIA and its Nvlink products.
So far, Nvlink has been closed for chips made by Nvidia. Nvlink is a technology developed by Nvidia to connect and exchange data between its graphic processors and processors.
“NV Link Fusion is what you can build AI Semi-Custom Infrastructure, not just half-legged,” Juan said in Taiwan’s Computex 2025, the largest electronics conference in Asia.
According to Huang, Nvlink Fusion allows AI infrastructure to combine NVIDIA processors with different processors and integrated schemes (ASICS). “In any case, you have the benefit of using NV links and NV links.”
Nvidia announced on Monday What are Chipmaking partners for NVLink already include Mediator. Marvell. Alchep. ASTRA Labs. Consistent and Cadence. Under Nvlink Fusion, Nvidia customers like Fujitsu and Qualcomm Technologies He will also be able to connect his own Nvidia graphic processors at AI data processing center, he added.
According to Ray Wang, a semiconductor analyst and technology based in Washington, Nvlink presents Nvidia’s storage plans for ASIC -based processing centers that are traditionally seen as NVIDIA competitors.
While NVIDIA takes a dominant position in the graphic processors used for overall learning AI, many competitors see place to expand in chips designed for more specific applications. Some of the NVIDIA’s largest competitors in AI Computing – which are also one of its largest customers – include cloud vendors such as Google, Microsoft and Amazon, all of them build their own processors.
NVLINK FUSION “consolidates Nvidia as the Center for the AI AI-Navate Factory If these systems are not completely built with NVIDIA chips,” Van said, noting that it opens for NVIDIA for customer service, which is not completely based on Nvidia, but looking to integrate some of its GPU.
“When widely accepted, NVLINK FUSION can expand the NVIDIA industry trail, contributing to a deeper cooperation with the CPU customer developers and ASIC designers in the construction of the AI future infrastructure,” Van said.
However, NVLink Fusion risk reduce the demand for NVIDIA processor, allowing Nvidia customers to use alternatives, reports Rolf Bulk, NEW Street Research Studies Analyst.
However, “at the system level additional flexibility increases the competitiveness of NVIDIA-based solutions based on GPU compared to alternative new architecture, helping NVIDIA keep its position at the AI computing center,” he said.
The competitors of Nvidia Broadcom, AMD and Intel are still missing from the NVLink Fusion ecosystem.
Other updates
Juan opened his main speech with the next -generation Grace Blackwell Systems update for the AI load. He said the “GB300” company, which will be released in the third quarter of this year, offers a higher total system performance, he said.
On Monday, Nvidia also announced the new NVIDIA DGX Cloud Lepton, AI platform with a computing market, which, according to Nvidia, connect the world developers with tens of thousands of graphic processes with the global cloud suppliers.
“DGX Cloud Lepton helps solve the critical problem providing reliable, high -performance GPU resources by combining clouds Press -Resolve.
In his speech, Juan also announced the plans of the new office in Taiwan, where the AI Supercomputer project will also be created with the Taiwanese Foxconn, officially known as Hon Hai Technology Group, the world’s largest manufacturer.
“We are pleased to cooperate with Foxconn and Taiwan to help build AI Taiwan infrastructure, as well as support TSMC and other leading companies to promote innovation in the II and robotics,” Juan said.
