Advanced Micro Devices (AMD) isn’t content with just playing second fiddle in the AI arms race.

Over the past year, we’ve seen the chipmaker going on the offensive, securing billion-dollar contracts, growing its data center footprint, and rolling out a string of high-performance accelerators targeting the heart of the AI infrastructure boom. 

It has been rewarded handsomely at the stock market, too.

AMD stock is up over 114% year to date, and is tracking almost 168% higher in the past six months. In the past month alone, the stock’s up more than 60%.

Behind AMD’s recent surge is a deeper shift potentially altering the balance of power in the hotly contested AI space.

At the center of that shift, CEO Lisa Su signals a vision that effectively challenges long-held assumptions over who’s setting the pace in AI, and what the next era of computing will look like.

AMD CEO Lisa Su says new MI300 and MI400 wins put AMD shoulder-to-shoulder with Nvidia.

Bloomberg/Getty Images

AMD CEO Lisa Su says AI’s big enough for more than one giant

In an interview with CNBC’s “Mad Money” host Jim Cramer, AMD CEO Lisa Su just dropped a clear message that the AI race isn’t one-sided. 

She effectively flipped the typical script that pits Nvidia as the only heavyweight. “The AI market is expanding, not shrinking,” she highlighted, emphasizing that more players essentially means more innovation. It seems AMD’s no longer just catching up, but also competing head-on.

More Nvidia:

  • IonQ CEO just threw a curveball at Nvidia
  • Why Nvidia’s Vera Rubin may unleash another AI wave
  • Nvidia just scored a massive AI win, but CEO Huang has regrets

That level of confidence comes with hard data.

Su said AMD’s MI300 chips are “seeing strong traction,” with a scale-up in production across clouds, enterprises, and national labs. Moreover, its MI400 series is set for 2025 deployment and is already powering two massive new Department of Energy supercomputers, one arriving early next year. 

We announced two new supercomputers… one will be deployed early next year, and the other with our next-generation MI400 series. That’s a sign of the confidence customers have in our roadmap.

AMD CEO Lisa Su

She also stressed the importance of the breadth of AI demand. 

Beyond the usual suspects in the cloud space, Su cited health care, energy, and enterprise inference as critical next-wave drivers. 

“The addressable market for AI and inference is actually increasing,” she highlights. “There’s a lot more compute that’s necessary… a big market is good for all of us.” 

Related: Top analyst issues surprise verdict on 40-year-old tech giant’s AI pivot

In short, AI’s far from being just a zero-sum game. 

Also, behind the optimism is some encouraging math: AI infrastructure spending is expected to soar past $200 billion by 2026, while AMD’s data center sales have risen by double-digits year-over-year over the past several quarters. 

Takeaways on AMD’s position in the AI race:

  • Lisa Su feels the AI race isn’t one-sided, arguing AMD is now competing head-on with Nvidia.
  • AMD’s MI300 and upcoming MI400 chips continue to gain traction, powering twoDOE supercomputers while expanding its potent footprint into cloud, health care, and enterprise AI markets.
  • AI spending is expected to skyrocket past $200 billion by 2026, and Su says AMD is “sharing the stage,” not chasing it.

AMD’s $1 billion supercomputer pact puts it squarely in the AI driver’s seat

Lisa Su also covered AMD’s powerful new $1 billion partnership with the Department of Energy, framing it as a “do it now” moment for U.S. computing power. 

The plan involves the development of a couple of AMD-accelerated AI supercomputers at Oak Ridge National Laboratory, a project Su hails as arguably an incredible speed run to national-scale AI.

Related: Cathie Wood pours $19.2 million on biotech stock, trims favorite

One of the machines, called Lux, serves as an AI factory cluster, while Discovery, a next-generation exascale system, is expected to follow later in the decade.

Hewlett Packard Enterprise will develop both, layered with Oracle Cloud Infrastructure, and integrated into Lux for cloud-scale training.

Details of AMD-Department of Energy deal:

  • Lux: powered by AMD Instinct MI355X GPUs, EPYC CPUs, along with Pensando networking, and is arriving early 2026.
  • Discovery: Led by the next-gen MI430 (MI400 series), and slated for late-decade deployment.
  • Investment: More than $1 billion in combined public-private funding.

The head-turning new move expands AMD’s DOE track record after Frontier and El Capitan, validating its accelerator roadmap in the process.

On top of that, Su linked the project to AMD’s quantum partnership with IBM, integrating classical and quantum computing for next-generation research.

Related: Palantir quietly makes massive AI war play