Advanced Micro Devices CEO Lisa Su on Thursday revealed a new artificial intelligence server platform set for 2026, signaling a direct challenge to Nvidia’s dominance as OpenAI announced plans to adopt AMD’s latest chips.
The announcements were made at AMD’s “Advancing AI” developer conference in San Jose, California. Despite the high-profile reveals, AMD shares slipped roughly 2% following the event.
Su introduced the upcoming MI350 and MI400 series AI chips, positioning them as rivals to Nvidia’s Blackwell processors. The MI400 series, she said, will power a new server platform named “Helios,” scheduled for release next year.
As AI chipmakers increasingly pivot from selling individual processors to offering entire server systems integrated with proprietary networking components, AMD’s move into full-stack systems marks a strategic escalation in its competition with Nvidia.
During the keynote, AMD emphasized that many components of its Helios server—such as networking standards—would be made openly available to the broader industry, including competitors like Intel.
“The future of AI is not going to be built by any one company or in a closed ecosystem. It’s going to be shaped by open collaboration across the industry,” Su said, in a clear jab at Nvidia’s more closed approach. Nvidia’s NVLink technology, which connects its chips, has been proprietary—though the company recently began licensing it amid growing competition.
Su was joined onstage by OpenAI CEO Sam Altman, who confirmed that OpenAI is using AMD’s MI300X and MI450 chips.
“Our infrastructure ramp-up over the last year, and what we’re looking at over the next year, have just been a crazy, crazy thing to watch,” Altman said.
Executives from companies including Elon Musk’s xAI, Meta Platforms, and Oracle also appeared to discuss their adoption of AMD hardware. Meanwhile, cloud provider Crusoe told Reuters it plans to invest $400 million in AMD’s new chips.
AMD reiterated its intention to match Nvidia’s aggressive release cadence, with annual updates to its AI hardware lineup. While AMD has lagged behind Nvidia in capturing AI market share, it continues to bolster its software ecosystem and chip capabilities in an effort to close the gap.
To that end, AMD completed its acquisition of server builder ZT Systems in March and is expected to begin offering full AI systems akin to Nvidia’s rack-sized server products. Based in Santa Clara, AMD has also recently expanded its AI talent and made a string of acquisitions aimed at strengthening its design and software capabilities.
At the conference, Su noted that the company had acquired 25 companies in the past year that were related to the company’s AI plans.
Just last week, AMD brought on the team from chip startup Untether AI, and on Wednesday, it confirmed the hiring of several employees from generative AI startup Lamini, including the co-founder and CEO.
Still, AMD’s ROCm software stack continues to face an uphill battle against Nvidia’s CUDA, widely seen in the industry as a cornerstone of Nvidia’s sustained dominance.
During AMD’s earnings call in May, Su said that despite increasingly aggressive curbs on AI chip exports to China, AMD still expected strong double-digit growth from AI chips.
Pump Radar
Stock Alerts