Nvidia to sell tech to enhance communication between AI chips

US tech company, Nvidia announced Monday it will begin selling a new chip interconnect technology to external designers.

The move is aimed at boosting communication between AI chips, enabling faster, large-scale system performance.

The new offering, called NVLink Fusion, builds on Nvidia’s existing NVLink system.

It allows multiple chips to share data more efficiently—critical for complex AI workloads.

“Nvidia launched a new version of its NVLink tech called NVLink Fusion on Monday that it will sell to other chip designers to help build powerful custom AI systems with multiple chips linked together,” Reuters reported.

MediaTek and Marvell Technology are among the first companies planning to adopt NVLink Fusion for their custom chip efforts.

Nvidia originally developed NVLink years ago. It’s now used in the company’s advanced GB200 chip, which combines two Blackwell GPUs with a Grace CPU to enable massive data throughput.

The announcement was made by Nvidia CEO Jensen Huang during a keynote at the Taipei Music Center, ahead of the Computex AI exhibition, which runs May 20–23.

Huang’s keynote traced Nvidia’s evolution. He noted that past presentations focused “90%” on graphics chips—but that focus has shifted dramatically.

“Now, Nvidia has grown beyond its roots as a video game graphics chip maker into the dominant producer of chips that have powered the AI frenzy that has gripped the tech industry since ChatGPT’s launch in 2022,” the article notes.

Nvidia is also working on CPUs that run Microsoft’s Windows OS using Arm Holdings technology, a development Reuters previously reported.

Last year at Computex, Huang generated major buzz in Taiwan, sparking what the media dubbed “Jensanity.” Crowds of fans and media followed his every move at the show.

At Nvidia’s developer conference in March, Huang explained the company’s shift from training AI models to enabling real-world AI applications.

“He unveiled several new generations of AI chips, including the Blackwell Ultra, which will be available later this year,” the article reports. “The company’s Rubin chips will be followed by Feynman processors, which are set to arrive in 2028.”

Nvidia also launched a new desktop AI workstation called DGX Spark. Huang said it’s in full production and should be ready in just a few weeks.

Computex 2025 is expected to host more than 1,400 exhibitors. It’s the first major gathering of chip industry leaders in Asia since former U.S. President Donald Trump’s tariff threats aimed at shifting production back to the U.S.

(Reuters)

ALSO READ TOP STORIES FROM NIGERIAN TRIBUNE