Intel’s upcoming Lunar Lake MX processors will break new ground by utilizing TSMC’s N3B fabrication technology for their x86 cores, marking the first instance of Intel employing outsourced process node technology for its high-performance CPUs. The leaked information, originally posted by prominent hardware leaker @YuuKi_AnS and quickly removed, provides a glimpse into Intel’s strategic shift in the development of its flagship x86 cores.
The Lunar Lake MX platform represents Intel’s ambitious foray into uncharted territory, utilizing TSMC’s advanced 3nm-class N3B process technology for the compute tile. Traditionally, Intel has relied on its in-house fabrication processes, with the company’s official statements suggesting the use of its own 18A (1.8nm-class) process for Lunar Lake CPUs. This departure from the norm raises questions about Intel’s future strategies and its willingness to explore external partnerships for key components.
The leaked slides outline the Lunar Lake MX platform’s specifications, emphasizing a microarchitecture designed from the ground up to deliver breakthrough performance-per-watt efficiency. The processors are tailored for mobile devices, featuring up to eight general-purpose cores—comprising four high-performance Lion Cove cores and four energy-efficient Skymont cores. Additional features include a 12MB cache, up to eight Xe2 GPU clusters, and a cutting-edge six-tile NPU 4.0 AI accelerator. The platform accommodates a wide power range, supporting both 8W fanless and 17W–30W fanned designs.
One of the notable design choices is Intel’s decision to consolidate the CPU, GPU, and memory controller into the same tile while housing other components in a separate SoC tile. This innovative approach aims to optimize physical space and enhance overall efficiency. The Lunar Lake MX is set to come equipped with 16GB or 32GB of LPDDR5X-8533 memory-on-package, a move expected to further reduce the platform’s footprint and elevate performance.