Intel NPU5 in Panther Lake

Panther Lake's NPU 5: area-lean, FP8-savvy, and built to be everywhere

October 9, 2025

David Harold

Intel’s fifth-gen NPU shrinks the design while getting faster where it counts. NPU 5 consolidates the compute into fewer but “fatter” neural compute engines (NCEs), doubles MAC density per NCE, adds native FP8, fixes channel padding, and moves more work off the SHAVE DSPs via programmable activations and a reworked data-conversion pipeline. From NPU 4 to NPU 5: make the MACs do more Intel’s redesign goal was area efficiency, not headline unit count. Where NPU 4 used six NCEs (Lunar Lake), NPU 5 pares that to three larger NCEs—each with twice the MACs—so you get more math, less overhead per

...

Enjoy full access with a TechWatch subscription!

TechWatch is the front line of JPR information gathering service, comprising current stories of interest to the graphics industry spanning the core areas of graphics hardware and software, workstations, gaming, and design.

A subscription to TechWatch includes 4 hours of consulting time to be used over the course of the subscription.

Already a subscriber? Login below

Intel NPU5 in Panther Lake

Enjoy full access with a TechWatch subscription!

This content is restricted

Login