Lisa Su, CEO of AMD, holds an AMD Ryzen AI Halo Developer Platform during an AMD keynote address at CES 2026, on Jan. 5, 2026. (Photo courtesy of REUTERS)
AMD, considered NVIDIA’s main rival, has launched a comprehensive market offensive by unveiling a wide range of new products spanning from AI datacenter supercomputer systems to personal PC AI and industrial embedded AI processors.
AMD CEO Lisa Su showcased the next-generation AI datacenter rack ‘Helios’ during the ‘CES 2026’ keynote speech held at the Venetian Hotel in Las Vegas, Nevada on Jan. 5 (local time).
Helios consists of 72 units of the next-generation graphics processing unit (GPU) ‘Instinct MI455’, unveiled for the first time that day, and 18 units of the datacenter central processing unit (CPU) ‘Venice’.
The ‘MI455’ chip, built on the latest 2-nanometer process technology, features 432GB (gigabytes) of 4th generation high bandwidth memory (HBM), delivering approximately 10 times better AI inference performance compared to its predecessor.
CEO Su introduced Helios on stage, saying “This is the world’s best AI rack” and describing it as not simply a server rack but a “monster,” citing the exponential increase in global AI computing demand as the reason such products are needed. AI users have exceeded 1 billion within just a few years since ChatGPT’s launch, with projections reaching 5 billion users in the future.
She explained that global computing capacity was 1 zettaflops in 2022. FLOPS refers to the number of complex numerical calculations with decimal points performed per second. One zetta represents 10 to the 21st power, which is 1 billion times larger than one tera.
CEO Su noted that global computing capacity had increased 100-fold to 100 zettaflops by last year, three years later, and projected it would increase another 100-fold within five years, requiring 10 yottaflops (yotta representing 10 to the 24th power). She explained that the “monster” Helios is a product designed to prepare for this ‘yotta’ era.
AMD also introduced the ‘Ryzen AI 400’ and ‘Ryzen AI Pro 400’ series processors for personal AI PCs. The company also unveiled ‘Ryzen AI Halo’, a small box-sized AI developer mini PC platform. These products are scheduled for sequential release in the first and second quarters of this year.
The company also presented the ‘Ryzen AI Embedded’ P100 and X100 series for automotive and industrial equipment. These will be installed in vehicle infotainment systems, autonomous driving, and industrial automation robots. In the gaming sector, AMD announced plans to release the ‘Ryzen7 9850X3D’ processor in the first quarter of this year.
During the keynote speech, Su showcased the company’s ecosystem by inviting OpenAI co-founder Greg Brockman, World Labs CEO Fei-Fei Li, and Blue Origin Senior Vice President John Couluris to the stage for discussions. She also invited White House Office of Science and Technology Policy Director Michael Kratsios to emphasize AMD’s participation in the U.S. government’s AI initiative ‘Genesis Mission’.
