Four Intel GPUs host Meta’s Muse Spark

High-capacity memory brings large models into air-cooled servers.

November 9, 2026

Jon Peddie

Meta’s Muse Spark Open combines 816 billion parameters with a sparse architecture that activates 42 billion per token. Intel has announced day-zero support on its upcoming GPU D-series and current Xeon processors. Four D-series cards provide 640 GB of aggregate memory in an air-cooled PCIe configuration. Low-precision quantization makes the model practical to host, while open-source serving tools support deployment. The configuration offers enterprise developers a promising route to private agentic inference, subject to independent performance and availability validation. Meta released Muse Spark Open as open weights on October 15, 2026, and Intel announced support for the model on the

...

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

Four Intel GPUs host Meta’s Muse Spark

Enjoy full access with a TechWatch subscription!

This content is restricted

Login