LAS VEGAS, NEVADA – JANUARY 04: AMD Chair and CEO Dr. Lisa Su delivers a keynote tackle at CES 2023 at The Venetian Las Vegas on January 04, 2023 in Las Vegas, Nevada. CES, the world’s largest annual shopper know-how commerce present, runs from January 5-8 and options about 3,100 exhibitors displaying off their newest services and products to greater than 100,000 attendees. David Becker/Getty Photographs/AFP (Picture by David Becker / GETTY IMAGES NORTH AMERICA / Getty Photographs by way of AFP)
- AMD is buying Nod.ai to enhance its open-source AI capabilities.
- The acquisition is a part of the corporate’s “AI development technique.”
- The acquisition is anticipated to be accomplished inside the present quarter.
The rivalry between Superior Micro Units, Inc (AMD) and Nvidia Corp dates again a long time — even earlier than the previous purchased ATI for its graphics division. As we speak, they’re nonetheless battling it out as the highest two graphic cardmakers for gaming PCs, however that’s not the place the rivalry ends.
Each corporations’ major focus is now shifting to AI, and the competitors has by no means been extra fierce. Though Nvidia dominates the market at this time with its AI chips, AMD is changing into a robust challenger because it does extra to accentuate the rivalry.
Nvidia holds a dominant place available in the market for high-performance chips used within the improvement of ChatGPT and the AI companies which have surged in recognition throughout the tech {industry} in 2023. That surge in demand following ChatGPT’s frenzy has propelled Nvidia’s market capitalization past US$1 trillion. However the surge was so sturdy and so sudden that the necessity for Nvidia’s AI chips has led to a provide scarcity, which the corporate has acknowledged and is actively addressing.
AMD is utilizing the scarcity to actively slender the AI hole on Nvidia, saying its most up-to-date transfer this week. AMD is buying open-source AI software program firm Nod.ai – a transfer made to broaden the corporate’s open AI software program capabilities. The US-based chip big mentioned the settlement strongly aligns with its AI development technique centered on an open software program ecosystem that lowers the boundaries of entry for patrons by developer instruments, libraries, and fashions.
“The addition of Nod.ai will convey an skilled group that has developed an industry-leading software program know-how that accelerates the deployment of AI options optimized for AMD Intuition™ knowledge middle accelerators, Ryzen™ AI processors, EPYC™ processors, Versal™ SoCs and Radeon™ GPUs to AMD,” the corporate defined.
AMD improves as a prospect with the acquisition of Nod.ai. Supply: X
Nod.AI, or Nod Labs, builds open-source applied sciences “for future AI programs,” in line with the startup. It primarily focuses on reinforcement studying, a sort of system that “learns” by way of trial and error. The acquisition is predicted to shut this quarter, an AMD spokesperson informed CNBC.
“At Nod.ai, we’re a group of engineers targeted on problem-solving — shortly – and shifting at tempo in an {industry} of fixed change to develop options for the following set of issues,” mentioned Nod.ai’s co-founder and CEO Anush Elangovan. “By becoming a member of forces with AMD, we’ll convey this experience to a broader vary of shoppers on a worldwide scale.”
DANA POINT, CALIFORNIA – SEPTEMBER 26: Dr. Lisa Su, Chair and CMO, AMD speaks onstage throughout Vox Media’s 2023 Code Convention at The Ritz-Carlton, Laguna Niguel on September 26, 2023 in Dana Level, California. Jerod Harris/Getty Photographs for Vox Media/AFP (Picture by Jerod Harris / GETTY IMAGES NORTH AMERICA / Getty Photographs by way of AFP)
Including to Elangovan’s context, Nod.ai delivers optimized AI options to high hyperscalers, enterprises, and startups. The compiler-based automation software program capabilities of Nod.ai’s SHARK software program cut back the necessity for handbook optimization and the time required to deploy extremely performing AI fashions to run throughout a broad portfolio of the info middle, edge and consumer platforms powered by AMD CDNA™, XDNA™, RDNA™ and “Zen” architectures.
AMD believes that buying Nod.ai will considerably improve its means to supply AI prospects with open software program to deploy extremely performing AI fashions tuned for AMD {hardware} shortly. Vamsi Boppana, senior vp of the Synthetic Intelligence Group at AMD, mentioned, “The addition of the proficient Nod.ai group accelerates our means to advance open-source compiler know-how and allow moveable, high-performance AI options throughout the AMD product portfolio.”
Boppana sees a bonus within the sheer proven fact that Nod.ai’s applied sciences are already extensively deployed within the cloud, on the edge, and throughout a broad vary of endpoint gadgets at this time. The information of the acquisition got here on the heels of the corporate’s buy of French startup Mipsology to strengthen its AI inference software program capabilities.
On August 24, the California-based chip big unveiled the acquisition, saying that Mipsology, “a frontrunner in AI software program and long-standing AMD accomplice,” will assist the corporate “speed up our buyer engagements and broaden our AI software program improvement capabilities.” Boppana, in a press releasewrote, “The group will assist develop our full AI software program stack, increasing our open ecosystem of software program instruments, libraries, and fashions to pave the way in which for streamlined deployment of AI fashions operating on AMD {hardware}.”
Each acquisitions are a part of AMD’s bigger technique to problem Nvidia’s dominance within the AI computing area with a broad portfolio of what AMD chair and CEO Lisa Su has known as “management GPUs, CPUs, and adaptive computing options for AI inferencing and coaching.”
In truth, since final 12 months, Su and different executives have articulated a complete AI technique that addresses what she mentioned is a “multibillion-dollar development alternative throughout cloud, edge, and an more and more numerous variety of clever endpoints.”
