Surprisingly Useful AI Article Enhancements
CES Keynote Lays Out Path To Yotta-Scale Computing And Local Ai For PCs
AMD has used its CES 2026 keynote to outline an ambitious plan to scale Ai infrastructure from data centres to classrooms and PCs, revealing new chips, rack-scale platforms and a $150 million commitment to support education initiatives.
AMD Chair and Chief Executive, Lisa Su, said the company’s deep product stack and partner collaborations were laying the groundwork for what she described as the “era of yotta-scale computing.”
She needed to inform the tech-savvy room that a Yottabyte was bytes.
“Ai adoption is accelerating. We are entering the era of yotta-scale computing, driven by unprecedented growth in both training and inference,” Su said.
“AMD is building the compute foundation for this next phase of Ai through end-to-end technology leadership, open platforms and deep co-innovation with partners across the ecosystem.”
‘Helios’ Rack Platform Targets Trillion-Parameter Models
At the heart of AMD’s keynote was a first look at its ‘Helios’ rack-scale reference platform, designed to deliver up to 3 Ai exaflops of performance in a single rack.
Targeting trillion-parameter training workloads, Helios combines AMD Instinct MI455X accelerators, AMD EPYC “Venice” CPUs and AMD Pensando “Vulcano” NICs, unified via the open-source ROCm software stack.
AMD said the platform was designed for maximum bandwidth and energy efficiency and would underpin the transition from today’s 100 zettaflops of compute to 10 yottaflops within five years.
Alongside Helios, AMD formally launched its full MI400 Series of accelerators, including the Instinct MI440X, aimed at on-premises enterprise Ai deployments. The eight-GPU form factor is designed for scalable training, fine-tuning and inference and is intended to slot into existing data centre infrastructure.
The company also previewed the MI500 Series, expected to arrive in 2027. AMD said MI500 chips would deliver up to 1,000 times the Ai performance of 2023’s MI300X, leveraging 2nm process nodes, HBM4E memory and a new CDNA 6 architecture.
Ryzen Ai Push For Consumer And Developer Markets
AMD expanded its Ryzen Ai portfolio with a new 400 Series and Ryzen Ai Pro 400 Series, targeting premium Ai PCs. The chips feature a 60 TOPS NPU and full ROCm support for scaling workloads between cloud and device. Systems are expected to ship from January, with OEM availability in Q1.
Two new chips — Ryzen Ai Max+ 392 and Max+ 388 — support up to 128-billion-parameter models with 128GB of unified memory, aiming to boost local inference, content creation and gaming on thin-and-light notebooks and small form factor desktops.
For developers, AMD introduced the Ryzen Ai Halo platform — a compact desktop aimed at building Ai models with a high tokens-per-second-per-dollar ratio. Halo is expected in Q2.
Embedded Ai Targets Edge And Robotics
AMD has also entered the embedded Ai market with the Ryzen Ai Embedded P100 and X100 Series. These chips are designed for automotive displays, healthcare systems and autonomous robotics, offering power-efficient Ai compute in constrained environments.
The company said the chips support use cases including humanoid robotics and physical Ai deployments at the edge.
Related: Best Business Laptops for work & school
Related: Best Gaming Laptops
Related: Best Portable Laptop
White House Endorsement And Education Pledge
Su was joined on stage by White House Office of Science and Technology Policy Director, Michael Kratsios, who spoke about AMD’s role in the Genesis Mission — a US government initiative aimed at securing national leadership in Ai.
Kratsios pointed to new AMD-powered systems at Oak Ridge National Laboratory, including the Discovery and Lux supercomputers, as evidence of AMD’s growing influence in national-scale infrastructure.
He also encouraged private sector efforts to expand Ai education. AMD responded by pledging $150 million toward classroom and community initiatives to make Ai more accessible to students.
Student Hackathon And Industry Participation
The keynote closed with recognition of over 15,000 student participants in the AMD Ai Robotics Hackathon, delivered in partnership with Hack Club.
AMD also highlighted collaborations with partners including OpenAI, Luma Ai, Liquid Ai, Blue Origin, Illumina, Absci, Generative Bionics and AstraZeneca, many of which are using AMD hardware to train and deploy large-scale models.
“Together, we’re showing what’s possible when the industry comes together to bring Ai everywhere, for everyone,” Su said.
Last Updated on January 7, 2026 by Nick Ross
