China’s NVIDIA Rival? Moore Threads Reveals Lushan and Huashan GPUs

Chinese chipmaker Moore Threads has revealed two next-generation GPUs: Lushan for gaming and content creation, and Huashan for artificial intelligence workloads.

Both chips are based on the new Flower Harbor, also known as Huagang, GPU architecture. According to reports, Moore Threads claims a major leap over its previous MTT S80 and MTT S90 graphics cards: up to 15x higher gaming performance, 50x better ray tracing, 64x faster AI performance, and up to 4x more memory. The company also says the new architecture supports modern APIs, including DirectX 12 Ultimate, which was one of the biggest weaknesses of its earlier consumer GPUs.

The AI-focused Huashan accelerator is even more ambitious. It reportedly uses a dual-chiplet design with HBM memory and is positioned against NVIDIA’s Hopper and Blackwell-class accelerators. Moore Threads also promotes MTLink, a high-speed interconnect designed to scale AI clusters to extremely large GPU counts.

The first Lushan gaming cards and Huashan AI accelerators are expected in 2026. The big question is whether Moore Threads can turn impressive claims into real-world performance, stable drivers and a strong software ecosystem.

This is Xenon Lab News — AI, GPUs, semiconductors and future computing in one place.

#XenonLab #XenonLabNews #MooreThreads #LushanGPU #HuashanGPU #ChineseGPU #NVIDIA #AIChips #GamingGPU #GPU #GraphicsCard #RayTracing #DirectX12Ultimate #AIHardware #Semiconductors #TechNews #HardwareNews #FutureTech #Blackwell #Hopper

Видео China’s NVIDIA Rival? Moore Threads Reveals Lushan and Huashan GPUs канала Xenon Lab | Science & Technology