Intel says its next-generation 14A process is advancing quickly and already resonating with external customers, setting up a strong follow-on to 18A while the company expands its collaboration with NVIDIA and navigates ongoing supply constraints across the PC market.
According to Intel VP John Pitzer at a recent industry conference, 14A is in the definition phase and looks stronger at this stage than 18A did. The difference, he emphasized, is early, direct engagement with external customers—something Intel didn’t do until later in 18A’s development. That shift is feeding more mature, industry-standard PDKs and helping optimize 14A for a wider range of designs from day one.
14A also doubles down on two marquee technologies: second-generation Gate-All-Around transistors and second-generation backside power delivery. Because 18A was Intel’s first transition from FinFET to GAA and its first use of backside power, 14A benefits from those hard-won learnings. Internally, Intel says 14A’s performance and yield metrics are tracking meaningfully ahead of where 18A was at a similar point. Meanwhile, the first 18A products are slated to arrive with Panther Lake later this year, with broader ramp expected in early 2026.
Intel also outlined how its partnership with NVIDIA will unfold across data center and client PCs.
On the data center side, Intel will supply a custom Xeon CPU designed to interface with NVIDIA’s NVLink Fusion interconnect. NVIDIA will integrate that CPU into its systems and handle go-to-market at the system level. The arrangement allows Intel’s x86 silicon to tap into NVIDIA’s high-bandwidth fabric, similar to how other platforms are adopting NVLink Fusion.
On the client side, the companies are teaming up on a new class of x86 SoCs that incorporate an NVIDIA RTX graphics tile. Intel will integrate the RTX tile alongside its CPU as part of a multi-tile design, initially targeting premium notebook PCs with plans to expand to broader price tiers over time. In this model, the graphics tile itself will be provided and sold by NVIDIA, while Intel delivers the SoC integration to OEMs. Both companies will continue to pursue their own CPU and GPU roadmaps independently, with these “halo” designs opening a fresh segment for high-performance laptops.
Beyond product plans, Intel addressed a tight supply environment that’s pressuring older nodes. The company expects improved cost structure as new wafers begin coming out of Arizona early next year. In the near term, Intel is prioritizing the mid to high end of the PC market, raising prices on select 10nm and 7nm parts like Raptor Lake while lowering prices on current-gen platforms such as Arrow Lake and Lunar Lake to keep the broader PC stack supplied. That mix helps bridge the market until Panther Lake on 18A arrives in premium systems in the first half of 2026. Intel has already enacted notable price reductions on Arrow Lake in some regions to accelerate adoption.
What it means for buyers and builders:
– Expect faster time-to-value from Intel’s 14A thanks to earlier external feedback, second-gen GAA, and refined backside power delivery.
– High-end laptops could see a leap in on-device graphics performance via RTX tiles integrated directly into Intel SoCs.
– Data centers will gain new CPU options purpose-built to ride NVIDIA’s NVLink Fusion fabric.
– Short-term pricing and availability may fluctuate as Intel de-emphasizes low-end PCs, raises prices on select legacy parts, and cuts prices on Arrow Lake and Lunar Lake to cover more of the market.
With 18A-based Panther Lake queued for premium systems and 14A pushing ahead with healthier yields and industry-aligned PDKs, Intel is positioning its CPU and process roadmap to compete across data center and client while opening the door to novel GPU-accelerated PCs through its work with NVIDIA.
