Nvidia launches NVLink Fusion to boost chip-to-chip AI performance; partners include Marvell, MediaTek, Fujitsu, and more. Taiwan HQ also announced.
Nvidia said on Monday that it plans to sell a technology that will tie chips together to speed up the chip-to-chip communication needed to build and deploy artificial intelligence tools.
The company launched a new version of its NVLink tech called NVLink Fusion that it will sell to other chip designers to help build powerful custom AI systems with multiple chips linked together.
CEO Jensen Huang announced NVLink Fusion at the Taipei Music Center, the site of the Computex AI exhibition, which runs from 20 May to 23.
Marvell Technology and MediaTek plan to adopt the NVLink tech called Fusion for their custom chip efforts, Nvidia said. Other partners include Alchip, Fujitsu and Qualcomm.
Nvidia’s chips, however, face a cloudy outlook in China. In an interview with Stratechery’s Ben Thompson, Huang said the company “walked away from $15 billion of sales” in China after the U.S. put limits on shipments of its H20 chip.
Also Read: Explained: Reinforcement Learning
The company said last month it would take $5.5 billion in charges related to these curbs.
Nvidia’s NVLink is used to exchange massive amounts of data between various chips such as in its GB200, which combines two Blackwell graphics processing units with a Grace processor.
In addition to the announcement on new tech production, Huang disclosed the company’s plan to build a Taiwan headquarters in the northern suburbs of Taipei.
