AMD unveiled its next-generation AI strategy encompassing data centers, AI PCs, and embedded systems, presenting its vision of 'AI Everywhere, for Everyone.'

Dr. Lisa Su Lisa Su, CEO of AMD, emphasized AMD's end-to-end AI platform strategy in her CES 2026 keynote speech on the 5th, saying, "AI adoption is accelerating and the entire industry is entering the era of Yota-scale computing."

■ 'Helios' unveiled for Yota-scale computing

AMD has presented a blueprint for Yotta-scale computing as the future of AI infrastructure.

With the global computing capacity currently at around 100 zettaflops projected to expand to 10 yottaflops within the next five years, AMD has unveiled its core platform for this, the 'Helios' rack-scale system.

Helios delivers up to 3 AI exaflops of performance in a single rack, integrating MI455X accelerators, EPYC Venice CPUs, and Pensando Vulcano NICs to maximize bandwidth and efficiency for training ultra-large models.

■ Accelerate data center AI with the MI400 and MI500 series

AMD has unveiled its full lineup of Instinct MI400 series processors for the data center.

In particular, the MI440X GPU supports scalable learning, fine-tuning, and inference workloads based on an 8-GPU form factor, and enhances integration with existing infrastructure.

Additionally, MI430X is expected to be applied to global AI supercomputers such as 'Discovery' at Oak Ridge National Laboratory and 'Alice Recoque', France's first exascale system.

AMD also previewed the MI500 series, which is scheduled to be released in 2027.

Up to 1,000 compared to MI300XIt is being developed based on the CDNA 6 architecture, 2nm process, and HBM4E memory, aiming for improved performance.

■ Targeting the AI PC market… Ryzen AI 400 and Max+ series announced

In the AI PC field, the Ryzen AI 400 series and Ryzen AI PRO 400 series were unveiled.

Organically connects on-device AI and cloud AI through NPU 60 TOPS performance and full ROCm support.

Additionally, Ryzen AI Max+ 392·388 supports up to 128GB of integrated memory, enabling local execution of up to 128B parameter models.

The Ryzen AI Halo platform for developers has also been unveiled, enabling high-performance AI development environments to be implemented on small form factor PCs.

■ 'Physical AI' platform for embedded and robotics

AMD also announced the Ryzen AI Embedded P100 and X100 series for edge environments such as automotive, medical, and robotics.

Humanoid robot company Generative Bionics unveiled its AMD-based robot, GENE1.0, demonstrating the potential of embedded AI.

$150 million investment to expand AI education

AMD announced a $150 million global support program to expand access to AI education.

We are also participating in the 'Genesis Mission' in cooperation with the White House Office of Science and Technology Policy, contributing to the development of AI talent and strengthening national competitiveness.

Also, 'AMD AI Robo' conducted with Hack ClubMore than 15,000 students participated in the 'Tix Hackathon' and gained hands-on experience with AI.