Intel Debuts Core Ultra Series 3 Panther Lake AI PCs

At the Consumer Electronics Show on Jan. 5, Intel unveiled its Core Ultra Series 3 processors, code‑named Panther Lake, announcing what it described as the first broadly available AI‑PC platform produced on its 18A process node. The company presented the lineup as mobile first, blending CPU, integrated Arc‑class graphics and on‑chip neural processing units to accelerate AI tasks locally on consumer laptops and a wider range of edge devices.

Intel said Series 3 marks the first client silicon manufactured on Intel 18A, a node reported as a 2‑nanometer class process, and that production is underway in the United States. The operational status of 18A represents a test of Intel’s revised process roadmap and, if sustained, could revive plans to offer third‑party manufacturing capacity at its plants. Company officials also framed the launch as a platform play: beyond ultraportable notebooks, they intend Series 3 to appear in mini‑PCs, embedded systems and industrial hardware.

The Series 3 family comprises 14 SKUs. Intel outlined a slate that ranges from the high‑end Core Ultra X9 388H, a 16‑core part that can reach up to 5.1 GHz, down to models such as the Core Ultra 5 322, an 8‑core chip clocked to 4.4 GHz. All SKUs include low‑power neural processing units; Intel reported those NPUs can deliver up to 50 TOPS, enabling background and always‑on AI features without fully engaging the main GPU or CPU.

Graphics are integrated and described as Arc‑class designs, with press materials making reference to Xe3 Arc technology. Intel also presented a performance comparison in which a mobile Arc B390 GPU variant was claimed to be 73 percent faster than AMD’s Radeon 890M, a claim the company promoted as part of the Series 3 performance story. Intel emphasized that its design intent assigns heavier AI workloads to the GPU while using NPUs to handle simpler or lower‑power inference tasks, a balance intended to improve both performance and battery life.

On the adoption front, Intel said more than 200 partner designs will ship with Series 3, and the company asserted that the chips have been certified for a range of embedded and industrial uses including robotics, healthcare devices and other edge computing applications. Intel said edge systems based on Series 3 are expected to begin shipping in the second quarter of 2026.

Availability for consumer systems was presented as immediate: preorders were scheduled to begin Jan. 6, 2026, with initial global system availability set for Jan. 27 and additional designs rolling out through the first half of the year. Analysts and device makers will watch whether the combination of new process technology, integrated AI acceleration and broad partner support can deliver meaningful battery life and performance gains, and whether the operational 18A fabs can underpin longer‑term shifts in Intel’s manufacturing and foundry ambitions.