On the evening of January 5, 2026, Intel took the stage in Las Vegas to launch Core Ultra Series 3, codenamed Panther Lake: the first AI PC platform whose compute tile is built on Intel’s 18A process. Jim Johnson, senior vice president and general manager of the Client Computing Group, delivered the keynote, and CEO Lip-Bu Tan made a surprise appearance to stress that Intel had kept its commitment to ship its first 18A products before the end of 2025.
The stakes for Intel are bigger than one processor generation. 18A is the bet the foundry business is riding on, and the launch comes with hard numbers attached: global availability on January 27, more than 200 laptop designs from partners, and Johnson’s claim that this will be “the most broadly adopted AI PC platform Intel has ever shipped.”
18A in Volume: The Process Is the Story
Panther Lake is Intel’s first client chip with its compute tile fabricated on 18A, and the node is already in high-volume production. Johnson called 18A “the most advanced process in the world,” pointing to RibbonFET gate-all-around transistors and PowerVia backside power delivery; Intel’s own press kit positions it as the most advanced process node in the United States. For buyers and product teams, the subtext matters more than the slogan: Intel needs this product to succeed commercially to prove that 18A is a credible foundry option for external customers, not just an internal milestone.
Architecture: Cougar Cove, Darkmont, and Arc B390
The design blends Lunar Lake’s efficiency with Arrow Lake’s scalability. Performance cores are Cougar Cove and efficiency cores are Darkmont, with Intel claiming roughly 50% more performance than Lunar Lake at similar power. The integrated Arc B390 GPU carried much of the show: Intel billed it as delivering discrete-class gaming performance — 2x a competitor in select titles and 73% faster on average — and it is the first integrated graphics to support multi-frame generation. EA demoed Battlefield 6 running at up to 145 fps using XeSS with 4x frame generation. The GPU now sits on its own chiplet, and 4K streaming draws a third of the power of the previous generation.
The AI Pitch Shifts to GPUs and the Edge
Unlike rivals who put the NPU front and center, Tom’s Hardware noted that Intel leaned heavily on the GPU in this launch. The concrete AI claims: a 70-billion-parameter model running on-device with a 32K context window, a new hybrid local-plus-cloud “AI Super Builder” branding, and edge and embedded product lines folded into Series 3 under the banner of “agentic actioning.” There was even a robotics footnote: Oversonic Robotics’ RoBee humanoid, built in Italy, is being tested with Series 3 edge processors to do person and gesture recognition on-device, without touching the cloud. It is a different entry point into the same physical-AI era Jensen Huang declared from the same CES stage a day earlier — Intel is selling the client-side compute that physical AI runs on.
January 27: Ecosystem by the Numbers
Preorders opened the day after the keynote, with worldwide availability on January 27, 2026. More than 200 partner laptop designs are coming, spanning consumer, commercial, and edge use cases. Tan’s framing was blunt: Series 3 is “the next evolution of the PC.” Given where Intel has been these past few years, the numbers themselves are the message — first 18A product in volume, 200-plus designs committed, a firm on-sale date — evidence that the platform cadence has not slipped.
What It Means for Developers and Product Teams
Three practical effects. First, the envelope for local models on x86 laptops is widening: if the 70B-at-32K claim survives independent verification, hybrid architectures that infer locally and fall back to the cloud can be planned on mainstream hardware. Second, folding edge into the same platform means consumer and industrial workloads can share a toolchain, cutting the cost of cross-device deployment. Third, the AI PC is becoming a baseline rather than a differentiator: with 200-plus designs shipping NPU-plus-Arc-B390-class silicon, product decisions can assume on-device inference exists. The standing caveat: numbers like “2x” and “73% faster” come from the vendor, so wait for independent reviews before betting on them.
Sources
- Intel at CES 2026 Press Kit — Intel
- Intel CES 2026 keynote live blog — Tom’s Hardware
- Intel at CES 2026 live updates — Engadget
AI-assisted summary compiled from the sources above, reviewed by a human before publishing.