The Consumer Electronics Show kicked off this week with the semiconductor industry's biggest players unveiling ambitious AI-focused roadmaps.
NVIDIA Unveils Rubin: The Six-Chip AI Supercomputer
NVIDIA CEO Jensen Huang opened CES with a sweeping two-hour presentation that positioned the company's new Rubin platform as the cornerstone of next-generation AI infrastructure. Named after pioneering American astronomer Vera Rubin, the platform represents what NVIDIA calls its first "extreme-codesigned" AI system, integrating six new chips into a unified architecture.
"Computing has been fundamentally reshaped as a result of accelerated computing, as a result of artificial intelligence," Huang declared on stage at the Fontainebleau Las Vegas.
The Vera Rubin NVL72 rack-scale system combines 72 Rubin GPUs with 36 Vera CPUs, delivering what NVIDIA claims is up to a 10x reduction in inference token cost compared to its Blackwell predecessor. Each Rubin GPU features 336 billion transistors, while each Vera CPU packs 88 custom Olympus cores with 227 billion transistors. The platform enters full production immediately, with Rubin-based products expected from partners in the second half of 2026.
NVIDIA Partners with Mercedes-Benz on AI-Defined Driving
Perhaps the most consumer-facing announcement from Huang's keynote came in autonomous driving. NVIDIA unveiled Alpamayo, a family of open-source reasoning models designed to help self-driving vehicles navigate complex scenarios by breaking down problems and explaining decisions—a capability Huang described as "the world's first thinking, reasoning autonomous vehicle AI."
The flagship Alpamayo 1 is a 10-billion-parameter model that approaches driving decisions more like a human would, reasoning through rare edge cases rather than simply reacting to data.
The Mercedes-Benz CLA will be the first production vehicle to ship with NVIDIA's complete autonomous driving stack, including Alpamayo. The vehicle was rated "the safest car in the world" by Euro NCAP and is expected to launch in the U.S. this quarter with Level 2+ automated driving capabilities.
"The ChatGPT moment for physical AI is here—when machines begin to understand, reason and act in the real world," Huang said. "Our vision is that someday, every single car, every single truck, will be autonomous."
Major Cloud Providers Commit to NVIDIA Rubin Deployment
The Rubin platform has already secured commitments from the world's leading cloud providers. Nvidia says that AWS, Microsoft, Google, and Oracle have all announced plans to deploy Vera Rubin-based systems in their data centers.
Microsoft CEO Satya Nadella said the company is "building the world's most powerful AI superfactories" and will deploy NVIDIA Vera Rubin NVL72 systems across next-generation Fairwater AI superfactory sites.
CoreWeave will integrate Rubin-based systems into its AI cloud platform beginning in the second half of 2026. CoreWeave CEO Mike Intrator called the Rubin platform "an important advancement for reasoning, agentic and large-scale inference workloads."
AI labs including OpenAI, Anthropic, Meta, and xAI have also signaled adoption. OpenAI CEO Sam Altman noted that "intelligence scales with compute," while Anthropic CEO Dario Amodei praised the "efficiency gains" that will enable "longer memory, better reasoning and more reliable outputs."
AMD Launches Helios: The "World's Best AI Rack"
Not to be outdone, AMD CEO Lisa Su delivered the CES opening keynote just hours later, unveiling the company's Helios AI server rack as a direct competitor to NVIDIA's offerings.
Su brought out a large Helios rack unit on stage and declared it the "world's best AI rack"—a pointed challenge to NVIDIA's dominance. The system matches NVIDIA's 72-GPU configuration with 72 of AMD's MI455X chips, signaling AMD's aggressive push into enterprise AI infrastructure.
Su emphasized the unprecedented scale of AI demand driving these developments, predicting that 5 billion people will use AI daily within five years. To meet that demand, she argued, technology companies will need to increase global computing capacity by 100 times in the coming years.
AMD Previews MI500 Series with 1,000x Performance Claims
AMD expanded on its data center ambitions by providing details on its upcoming MI500 series GPUs, which the company positions as a generational leap in AI processing capability.
AMD boldly claims the MI500 series will deliver up to a 1,000x increase in AI performance compared to its MI300X predecessors. While specific architectural details remain scarce, the claim signals AMD's confidence in competing at the highest levels of AI infrastructure.
The announcement comes as AMD's stock has risen 76 percent over the past 12 months—outpacing even NVIDIA's 30 percent gain—though NVIDIA's $4.5 trillion market cap still dwarfs AMD's $359 billion valuation.
AMD Ryzen AI 400 Brings Copilot+ to Desktops
On the consumer front, AMD introduced its Ryzen AI 400 series processors, marking the company's first Copilot+ certified desktop chips.
The new processors feature 60 TOPS XDNA 2 NPUs—up from 50-55 TOPS in the previous Ryzen AI 300 chips —placing them well above Microsoft's 40 TOPS minimum for Copilot+ certification. The top-end Ryzen AI 9 HX 475 offers up to 12 Zen 5 CPU cores with 5.2GHz max boost speeds.
AMD claims the new chips deliver up to 30 percent faster multi-tasking, 70 percent faster content creation, and 10 percent faster gaming compared to previous generations. The company also promises 70 percent better "unplugged connectivity" performance—a sign of improved battery efficiency.
AMD Enters Robotics with Generative Bionics Partnership
AMD showcased its expanding presence in physical AI by bringing Generative Bionics CEO Daniele Pucci on stage to unveil the GENE.01 humanoid robot for the first time.
Powered by AMD's CPUs and GPUs, the robot is designed to operate in industrial environments. AMD is an investor in Generative Bionics, reflecting the company's strategic bet on robotics as a growth market.
The partnership positions AMD alongside NVIDIA in the race to power the next generation of autonomous machines—a market both companies see as central to the AI revolution.
Intel Launches Panther Lake: First Chips on 18A Process
Intel brought urgency to its CES presentation, unveiling the Core Ultra Series 3 processors (codenamed Panther Lake) as proof that its ambitious turnaround strategy is bearing fruit.
The chips represent the first consumer platform built on Intel's 18A process technology—what the company calls "the most advanced semiconductor process ever developed and manufactured in the United States."
Intel CEO Lip-Bu Tan made a surprise appearance to declare that Intel is "delivering on its commitment to ship its first 18A products" and that the chips are now ramping in high-volume production. The announcement is significant given published reports last summer indicating yields were below 50 percent.
Systems using Panther Lake chips will begin shipping on January 27, with over 200 PC designs planned from partners.
Intel Claims 76% Gaming Performance Jump with Core Ultra Series 3
Intel positioned Panther Lake as a major leap in integrated graphics performance, targeting gamers and content creators who want powerful laptops without discrete GPUs.
The new processors deliver more than 50 percent higher multi-threaded performance compared to Lunar Lake while maintaining similar power efficiency. Intel claims the flagship Core Ultra X9 388H offers integrated graphics performance comparable to discrete RTX 4050 laptop GPUs.
The company claims a 76 percent improvement in gaming performance over the previous Core Ultra 9 285H, based on testing across 45 game titles at 1080p with XeSS upscaling. The new X-series chips feature 12 Xe3 graphics cores—triple the four cores found in standard configurations.
Intel also announced a dedicated handheld gaming platform built on Panther Lake, with more details expected later this year.
Intel Expands Panther Lake to Edge and Industrial Applications
For the first time, Intel is certifying its flagship mobile processors for embedded and industrial use cases—a move that expands the company's addressable market beyond traditional PCs.
Intel says it has tested and certified Panther Lake for edge applications including robotics, smart cities, automation, and healthcare, with requirements like extended temperature support and 24/7 reliability.
Edge systems powered by Core Ultra Series 3 will be available starting Q2 2026. The expansion reflects Intel's strategy to leverage its AI PC investments across multiple market segments as it competes with NVIDIA and AMD for the growing edge AI opportunity.
