AMD unveiled new AI processors for corporate data centers at CES in Las Vegas on Monday, targeting Nvidia's market dominance with specialized chips for on-premise deployment.
CEO Lisa Su showed the MI440X, an enterprise version of AMD's MI400 series designed for smaller corporate data centers. The chip fits existing infrastructure not specifically built for AI clusters, allowing companies to keep data in-house rather than using cloud services.
The company also demonstrated its MI455 AI processors for data center server racks, which power systems sold to firms including ChatGPT maker OpenAI. AMD's Helios system based on the MI455X and new Venice CPU design will go on sale later this year.
OpenAI President Greg Brockman joined Su on stage, emphasizing chip advancements are critical for OpenAI's computing needs. AMD signed a deal with OpenAI in October that will add billions to annual revenue, with first MI400 series deployments rolling out this year.
Su previewed the MI500 series launching in 2027, claiming it offers 1,000 times the performance of older MI300 processors from 2023. The announcement follows Nvidia CEO Jensen Huang's CES presentation of the Vera Rubin AI platform, which enters full production this year.
AMD has created a multibillion-dollar AI chip business over the past two years but trails Nvidia's tens of billions in quarterly AI revenue. The Santa Clara-based company is one of Nvidia's strongest rivals in AI hardware but has struggled to match its market success.
"We don't have nearly enough compute for what we could possibly do," Su said during her keynote. "The rate and pace of AI innovation has been incredible over the last few years. We are just getting started."
The MI440X targets corporate customers seeking local hardware deployment, while the MI455 serves larger data center operations. Both chips represent AMD's push to capture enterprise AI spending currently dominated by Nvidia.
AMD also launched Ryzen AI 400 Series processors for AI PCs alongside Ryzen AI Max+ chips for local inference and gaming. The announcements came as Intel unveiled Panther Lake chips for laptops at CES, intensifying competition across AI hardware segments.
Industry analysts note AMD's OpenAI partnership provides validation but unlikely to immediately challenge Nvidia's market position. Nvidia continues selling every AI chip it manufactures amid strong demand for generative AI infrastructure.
The CES presentations highlight escalating competition in AI hardware as companies race to supply computing power for large language models and other AI applications. AMD's corporate-focused strategy aims to capture enterprise customers while Nvidia dominates hyperscale data centers.
