Taipei: NVIDIA said on Monday it plans to sell a technology to others to bundle chips together to speed up chip-to-chip communication needed to build and deploy artificial intelligence tools.
NVIDIA launched a new version of NVLink Tech on Monday called NVLink Fusion, which will be sold to other chip designers to help build powerful custom AI systems and link multiple chips together.
Marvell Technology and MediAtek plan to use NVLink technology called Fusion through customized chip work.
NVIDIA developed NVLink a few years ago, which is used to exchange large amounts of data between various chips, such as in the company’s GB200, combining two Blackwell Graphics Processing units with Grace processors.
NVIDIA CEO Jensen Huang announced the news at the Taipei Music Center at the Computex AI exhibition from May 20-23.
In addition to new technology production, Huang also announced plans to establish a Taiwan headquarters in the northern suburbs of Taipei.
His keynote talk discusses NVIDIA’s history of building AI chips and systems and the software the company built to support it.
Huang pointed out that his speech once spent 90% of the time on the company’s graphics chips, but that has changed.
Now, NVIDIA has become a source of video game graphics maker and a major producer of chips, which has powered AI Frenzy since Chatgpt launched in 2022.
Reuters previously reported that NVIDIA has been designing a CPU that will run Microsoft’s Windows operating system and use Arm Holdings’ technology.
Last year, Huang sparked “Jensanity” in Taiwan as the public and media were breathless, followed by the CEO being besieged by attendees at a trade show.
At the company’s annual developer meeting in March, Huang outlined how NVIDIA will position to address the shift from building large AI models to the computing needs of applications they run based on.
He has launched several new generations of AI chips, including the Blackwell Ultra, which will be available later this year.
The company’s Rubin chips will be followed by Feynman processors, which the processor is scheduled to arrive in 2028.
Nvidia has also launched a desktop version of its AI chip, called DGX Spark, with the goal of AI researchers. On Monday, Huang said the computer was fully produced and will be ready in “a few weeks.”
Computex will be the first major collection of computer and chip executives in Asia since U.S. President Donald Trump threatens to force companies to push companies to increase production, which will be the first major collection of computer and CHIP executives rally in Asia.
