During its Q3 2025 earnings call, Intel unveiled exciting details about its forthcoming CPU families, including Panther Lake, Nova Lake, and Coral Rapids, along with insights into their foundry plans involving 18A and 14A process technologies.
Intel Reveals 18A As Foundation for Future Client and Server Products
Intel’s CEO, Lip-Bu Tan, emphasized the company’s strategic direction regarding its client, server, and foundry businesses. His address underscored the developments previously outlined, with several significant updates.
First on the agenda was the much-anticipated Panther Lake CPU. Tan reiterated that the initial CPU WeU will debut by the end of this year, with more variants expected in the first half of 2026. Notably, the official unveiling is set for CES 2026, where Intel will showcase the Core Ultra Series 3 family. The premier WeU is designed as a higher-end option, while subsequent models are expected to be more affordable as the 18A technology matures.
Tan also observed the fiercely competitive nature of the high-end desktop segment, with Intel making progress despite the challenges. The imminent release of the Core Ultra 200S Plus, which will refresh the Arrow Lake CPU family on the LGA 1851 socket, will soon give way to the Nova Lake-S platform on the LGA 1954 socket in the second half of 2026. The Nova Lake architecture aims to deliver impressive specifications, potentially offering up to 52 cores along with new Xe3P Arc integrated GPUs and additional architectural enhancements.
Additionally, the team will spearhead the establishment of a new basic and design service business to provide tailored silicon for a diverse clientele. This effort will not only broaden the reach of our core x86 intellectual property but also leverage our design expertise to create solutions ranging from general-purpose computing to fixed-function applications. Our plans are firmly on track to introduce the first Panther Lake WeU by year-end, with further variants set to debut in the subsequent half of the coming year.
These advancements will solidify our position in the notebook sector for both consumer and enterprise markets, offering cost-optimized solutions that span our entire PC range, from entry-level options to premium Core Ultra products. Despite the competitive landscape in high-end desktops, we’ve seen consistent increases in Arrow Lake shipments, and our next-gen Nova Lake product is poised to enhance our offerings significantly, particularly as we continue to innovate in the PC-thinning market.
Lip-Bu Tan – Intel CEO (Q3 2025 Earnings Call)
On the server front, demand for the Granite Rapids (Xeon 6 P-Core) CPUs remains robust, extending to all major hyperscalers. Although the Client Computing Group (CCG) segment may see a decline in the next quarter, strong interest in server processors is expected to yield benefits for the Data Center and AI (DCAI) group.
Looking ahead, Intel’s Xeon family will expand with the introduction of the Xeon 6+ “Clearwater Forest”chips around mid-2026, leveraging the 18A process technology. Furthermore, the Diamond Rapids P-Core series, also crafted under the 18A technology, will be part of the Xeon 7 lineup, accompanied by the new Coral Rapids WeUs, which are set to reintegrate Simultaneous Multithreading (SMT) into the Xeon architecture.
Our position at the forefront of AI technologies remains strong, with consistent demand for Granite Rapids across all major hyperscaler clients.
The feedback from hyper-scalers regarding the Diamond Rapids remains positive, and we are advancing the Coral Rapids project, which will include SMT for enhanced performance. We are currently in the defining phase of this product and are focused on fulfilling our roadmap in the upcoming fiscal year.
Lip-Bu Tan – Intel CEO (Q3 2025 Earnings Call)
Shifting focus to Intel’s foundry operations, the company has made noteworthy advancements on the 18A process, which is now entering high-volume manufacturing (HVM) at the Fab 52 facility in Arizona. Intel is actively working towards achieving the 18A-P Process Design Kit (PDK) milestone, which is a performance-optimized variant promising a 10% increase in efficiency. Tan reaffirmed that the 18A technology will serve as the foundation for at least three generations of both client and server products. He also shared insights regarding the ongoing developments concerning the 14A process and Embedded Multi-Die Interconnect Bridge (EMIB) technology.
We are on a steady path with the Intel 18A process technology and are set to launch Panther Lake into the market this year. Yields are improving at a consistent rate, and high-volume manufacturing at our Fab 52 facility is fully operational.
Moreover, we are advancing our Intel 18A-P efforts and hitting significant PDK milestones. Our 18A process will support at least three future generations of client and server developments.
As for 14A, our team continues to refine technology definitions, transistor architecture, and foundational IP. Engagement with prospective external customers is ongoing, and we are encouraged by their feedback, guiding our strategic decisions.
Additionally, our advanced packaging initiatives are progressing, especially regarding EMIB technologies, where we see genuine differentiation. The growth of AI infrastructure has driven unprecedented demand for wafer capacity, creating significant opportunities for Intel’s foundry services. We are uniquely positioned to meet this surge as we maintain a disciplined investment strategy focused on scalability and capability.
We are committed to evolving our foundry services into a long-term venture characterized by trust and adaptability, ensuring we meet the distinct requirements of various customers in terms of power, performance, yield, cost, and timely delivery.
This mindset shift is crucial as we steer Intel foundry towards long-term success, concentrating on opportunities across existing and new markets while leveraging the increasing demand for computing power in an AI-driven landscape.
Lip-Bu Tan – Intel CEO (Q3 2025 Earnings Call)
Lastly, regarding Intel’s GPU developments, both the Arc series and other GPUs were mentioned, along with the recent announcement of the Crescent Island, the first Xe3P accelerator tailored for AI inference workloads. Intel aims to establish a multi-generational roadmap, focusing on GPUs equipped with superior memory and bandwidth capabilities.
In the near future, we plan to enhance AI capabilities across our Xeon processors, AI PCs, ARC GPUs, and through our open software ecosystem. Future iterations of inference-optimized GPUs will be released annually, featuring advancements in memory and bandwidth tailored to meet enterprise requirements.
Lip-Bu Tan – Intel CEO (Q3 2025 Earnings Call)
As a final note, this update builds upon Intel’s previous announcements over the last several months. With the imminent release of the Panther Lake utilizing the 18A process, industry observers are keen to see its real-world performance and the innovations it may bring forth in upcoming client and server products.
Related Articles:
AMD Announces Radeon AI PRO R9700 Pricing at $1299 and Reveals Launch Date
20:20
China Introduces UBIOS: The First Independent Firmware Standard Set to Replace UEFI
19:52
Top Budget Gaming CPUs Ranked for 2025
19:00
Leave a Reply