In This Article:
By Stephen Nellis and Max A. Cherney
SAN FRANCISCO/TAIPEI (Reuters) -Nvidia announced a new software platform on Monday that will create a marketplace for cloud-based artificial intelligence chips.
Nvidia's graphics processors, or GPUs, dominate the market for training AI models and a raft of new cloud players called "neoclouds" such as CoreWeave and Nebius Group have emerged to specialize in renting out Nvidia's chips to software developers.
The Santa Clara, California-based company announced the new tool called Lepton that lets cloud computing companies sell GPU capacity in one spot.
In addition to CoreWeave and Nebius, other firms joining the Lepton platform are Crusoe, Firmus, Foxconn, GMI Cloud, Lambda, Nscale, SoftBank Corp and Yotta Data Services.
Nvidia cloud vice president Alexis Bjorlin said that despite demand surging for chips at startups and large companies, the process of finding available chips has been "very manual".
"It's almost like everyone's calling everyone for what compute capacity is available," Bjorlin told Reuters in an interview. "We're just trying to make it seamless, because it enables the ecosystem to grow and develop, and it enables all of the clouds - the global clouds and the new cloud providers - access to Nvidia's entire developer ecosystem."
Absent so far from Nvidia's Lepton partner list are major cloud providers such as Microsoft, Amazon Web Services or Alphabet's Google. Bjorlin said the system is designed for them to be able to sell their capacity on the marketplace if they choose.
Lepton will eventually allow developers to search for Nvidia chips located in specific countries to meet data storage requirements, Bjorlin said. It will also allow companies that already own some Nvidia chips to more seamlessly search for more to rent.
“It’s a good move for them,” said Mario Morales, group vice president at research firm IDC. “Nvidia is close to about 5 million developers. So they want to figure out a way where they can give them (developers) more access to the technology.”
Nvidia did not disclose what the business model will be for the new software platform or say whether it would charge commissions or fees to either developers or clouds.
However, Bjorlin said developers will "still retain their own relationship to the underlying compute providers, so they've contracted directly with them."
(Reporting by Stephen Nellis in San Francisco and Max A. Cherney in Taipei, additional reporting by Wen-Yee Lee in Taipei; Editing by Sam Holmes and Kate Mayberry)