Intel Nova Lake CPUs May Bring 12-Core Xe3P Graphics and 18A-P Entry-Level Chips
Intel’s next-generation Nova Lake processors are shaping up to be one of the company’s most important CPU launches in years, especially for users who care about integrated graphics performance. New details suggest that several Nova Lake SKUs will feature upgraded Xe3P integrated graphics with up to 12 GPU cores, while some entry-level chips may be built using Intel’s own 18A-P process technology.
The biggest highlight is the reported use of Intel’s Xe3P GPU architecture, which is expected to deliver a stronger integrated graphics experience than current Xe designs. Intel has already made major progress with its Xe graphics roadmap, with Xe2 showing strong results in mobile chips and discrete Arc graphics. Xe3 is also expected to play a key role in upcoming laptops and handheld gaming devices. Nova Lake appears set to push that momentum further with Xe3P, a more advanced version designed for higher performance and improved features.
According to recent hardware leaks, Intel’s Nova Lake-H mobile lineup will include multiple models with a 12-core Xe3P iGPU configuration. These chips are expected to replace the 12-core Xe3 graphics options found in the Panther Lake generation. Meanwhile, Nova Lake-H models with smaller 4-core and 2-core graphics configurations are expected to continue using the standard Xe3 architecture rather than Xe3P.
This split suggests Intel may be positioning Xe3P as a premium integrated GPU solution for higher-end Nova Lake chips, while keeping regular Xe3 graphics for mainstream and lower-power designs. For laptop buyers, that could mean a clearer performance gap between entry-level and high-end Nova Lake systems, especially in lightweight gaming, content creation, AI workloads, and media acceleration.
Intel is also reportedly preparing a Nova Lake desktop processor with the same 12-core Xe3P graphics configuration. This chip is said to use an 8-core CPU design made entirely of efficiency cores and may target edge computing markets. Even so, its desktop form factor makes it especially interesting, because a powerful integrated GPU on a socketed platform could give Intel a stronger answer to AMD’s Ryzen APUs.
AMD has traditionally held a strong position in integrated graphics thanks to its Radeon-based APUs, including models with up to 8 RDNA compute units. However, early comparisons between Intel Xe3 graphics and AMD RDNA 3.5 suggest Intel may have a significant advantage in some mainstream segments. In certain benchmarks, Intel’s integrated GPU performance has reportedly approached nearly double that of competing Ryzen parts. If Xe3P improves further, Nova Lake could become a serious option for users who want solid graphics performance without buying a dedicated graphics card.
Beyond standard laptop and desktop processors, Intel is also believed to be developing more GPU-heavy designs for future platforms. These may include larger “Halo” products and custom silicon featuring NVIDIA RTX tiles. If these projects reach the market, Intel’s CPU lineup could become much more diverse, covering everything from low-power notebooks to premium hybrid chips with much stronger graphics capabilities.
Another major detail involves manufacturing. Entry-level Nova Lake chips with a 4+0 core configuration are reportedly planned for Intel’s internal 18A-P process. The rest of the mainstream and high-end Nova Lake lineup is expected to use TSMC’s N2P node.
This is important because Intel’s 18A process is already tied to Panther Lake’s compute tile, while 18A-P is viewed as an enhanced version intended to improve performance, efficiency, and competitiveness. Although 18A-P has been discussed heavily for external foundry customers, Intel may also use it in its own products. If these entry-level Nova Lake chips arrive as expected, they could become an early public showcase for Intel’s next-generation manufacturing technology.
For the higher-end Nova Lake-S desktop platform, current rumors point to Intel Core Ultra 400 branding, a Coyote Cove performance-core architecture, and Arctic Wolf efficiency and low-power efficiency cores. The top desktop configuration is expected to feature up to 52 cores in total, including up to 16 performance cores, 32 efficiency cores, and 4 low-power efficiency cores. Cache capacity could also rise sharply, with reports suggesting up to 160 MB to 320 MB of combined L2 and L3 cache, plus a large bLLC cache ranging from 144 MB to 288 MB.
Nova Lake-S is also expected to support DDR5 memory speeds of up to 8000 MT/s with CUDIMM support, up to 36 PCIe 5.0 lanes, up to 16 PCIe 4.0 lanes, and a new LGA 1954 socket. Power targets may vary widely, with PL1 ratings around 125W to 175W and very high maximum power limits for enthusiast-class configurations.
On the AMD side, the rival 2027 platform is expected to be Olympic Ridge, potentially launching as part of a future Ryzen 10000 family. These chips are rumored to use Zen 6 cores on TSMC’s N2P process, with up to 24 cores and 48 threads, support for the existing AM5 socket, DDR5 memory with CUDIMM support, and a 125W-plus TDP range. This points to a major desktop CPU battle in 2027, with Intel focusing heavily on hybrid core counts and advanced cache designs, while AMD continues refining its high-performance Zen architecture.
If the latest details prove accurate, Intel Nova Lake could be a major turning point for integrated graphics and process strategy. The combination of 12-core Xe3P iGPUs, TSMC N2P for higher-end chips, and Intel 18A-P for entry-level models suggests a broad, flexible lineup aimed at laptops, desktops, edge systems, and graphics-focused computing.
Nova Lake is still some time away, with a launch expected around 2027, so specifications may change before release. Still, the early picture is clear: Intel is preparing a more aggressive CPU and GPU strategy, and Nova Lake may be one of the first platforms to fully show what that new direction can deliver.
