Moore Threads Teases Lushan GPU With Record Video Memory Capacity

Chinese chipmaker Moore Threads has teased an upcoming graphics card codenamed "Lushan" that aims to provide the largest video memory capacity in consumer GPU history. The company is scheduled to hold a full unveiling of the hardware on December 29 at its MUSA Developer Conference 2026.

According to a social media post from Moore Threads MTT S80 product manager AIR_FORCE, the Lushan GPU is designed for both AAA PC gaming and local artificial intelligence inference. While the specific VRAM count remains unspecified in the teaser, the hardware is expected to target a higher framebuffer than the RTX 5090 to accommodate high-resolution gaming and large language models.

The company reported that the Lushan architecture includes significant upgrades over previous MTT S80 and MTT S90 desktop offerings. Moore Threads stated that "AI computing performance is improved by 64 times, and it is equipped with the new UNITE rendering architecture, a new generation of self-developed hardware ray tracing engine, and supporting AI generative rendering technology."

Industry analysts noted that increased memory capacity is critical for running large language models like Qwen3.8-27B, as even 8-bit quantized versions of such models require between 27 and 30GB of video framebuffer. The Lushan's ability to handle these workloads will depend on both memory bandwidth and total memory count to avoid performance bottlenecks during prompt processing.

Moore Threads has previously demonstrated competitive gaming performance, with its MTT S90 outperforming NVIDIA’s RTX 4060 in certain tests due to optimized drivers. The company also reported that driver optimizations allowed the MTT S80 to double its graphics performance within one year of its release.

FI Journal
Privacy Overview

This website uses cookies so that we can provide you with the best user experience possible. Cookie information is stored in your browser and performs functions such as recognising you when you return to our website and helping our team to understand which sections of the website you find most interesting and useful.