Skip to content

NewsGPU

Moore Threads Teases 'Lushan' Graphics Card With Record VRAM for Gaming and Local AI

Moore Threads has teased Lushan, a consumer graphics card it says will have the largest VRAM capacity of any consumer card, aimed at AAA gaming and local AI. No capacity, price or UK availability is confirmed; full details are due at its MUSA Developer Conference on 29 December 2026.

By , Editorial team

2 min read

Moore Threads Teases 'Lushan' Graphics Card With Record VRAM for Gaming and Local AI

Chinese GPU maker Moore Threads has teased a new consumer graphics card, codenamed Lushan, and the headline claim is memory. According to reports from Wccftech and 3DM, the company is promising the largest video memory (VRAM) capacity of any consumer graphics card so far, with the card pitched at both AAA gaming and running large language models locally. Forge PC hasn't seen or tested the card, and no capacity, price or release date has been confirmed.

What Moore Threads has said

  • A teaser, not a launch. Wccftech reports that the full unveiling is scheduled for 29 December 2026 at Moore Threads' MUSA Developer Conference (MDC 2026).

  • Record VRAM, but no number. Wccftech says the teaser follows an 18 November social media post from the product manager of Moore Threads' MTT S80 card, and that the VRAM amount hasn't been specified. 3DM reports that the product manager said on Bilibili that the card would "significantly refresh the upper limit of VRAM capacity" for consumer cards. Whether that means 32GB, 48GB or more is speculation for now.

  • Generational claims. 3DM reports claims of 16 times the geometry processing, 50 times the ray tracing and 64 times the AI computing performance of the previous generation. Wccftech quotes the 64-times AI claim and the translated teaser's mention of a new UNITE rendering architecture, a new self-developed hardware ray tracing engine and "AI generative rendering" technology.

These are the company's own claims, made against its own older cards, so they say nothing yet about how Lushan would compare with Nvidia or AMD.

Why VRAM matters for local AI

For local AI, the amount of VRAM decides which models fit on the graphics card. Wccftech gives the example that 8-bit quantised versions of some 27-billion-parameter models need around 27GB to 30GB of VRAM, more than most consumer cards have. It also points out that capacity isn't the whole story: memory bandwidth affects how quickly a model generates text. If you're sizing a card today, our guide to how much VRAM you need and our best GPUs for local LLMs are the places to start.

What to watch for on 29 December

  • The actual capacity and memory bandwidth. A large frame buffer helps only if the card can feed it.

  • Drivers and software support. For gaming, that means game compatibility; for local AI, support in the apps people actually use to run models.

  • UK availability. Neither report mentions UK or European availability.

Our verdict

Lushan is a teaser with one bold promise, record consumer VRAM, and no confirmed specs, price or UK availability. It's worth watching if you run local AI models, but don't change your plans until Moore Threads reveals the details on 29 December and independent testing follows.

The bottom line

Lushan is a teaser with one bold promise, record consumer VRAM, and no confirmed specs, price or UK availability. It's worth watching if you run local AI models, but don't change your plans until Moore Threads reveals the details on 29 December and independent testing follows.

Sources

  1. 01Wccftech, China's Moore Threads Has Teased An Upcoming Graphics Card Codenamed "Lushan"
  2. 023DM International, Moore Threads Officially Announces "Lushan" with the Largest VRAM Capacity in Consumer Graphics Card History

About the author

Forge PC Staff

Editorial team

The shared byline for articles researched, written and edited by the Forge PC editorial team. How we source, check and correct our work is set out in our editorial policy.

More from Forge PC Staff

Read next

How to Turn Thinking On or Off in Ollama (DeepSeek-R1, Qwen 3, gpt-oss)

How-To

How to Turn Thinking On or Off in Ollama (DeepSeek-R1, Qwen 3, gpt-oss)

Ollama lets you control reasoning models' thinking: --think and --think=false on the command line, /set think and /set nothink in chats, --hidethinking for scripts and a think field in the API. Check /api/show first, as models such as gpt-oss use low, medium and high levels instead.

Forge PC Staff2 min

How to Use Embedding Models in Ollama for Local Semantic Search and RAG

How-To

How to Use Embedding Models in Ollama for Local Semantic Search and RAG

Embedding models turn text into vectors for semantic search and RAG. Pull one such as embeddinggemma in Ollama, generate vectors with ollama run or the /api/embed endpoint, store them, then embed questions with the same model and compare using cosine similarity, all locally.

Forge PC Staff3 min

How to Turn On Variable Refresh Rate (VRR) in Windows 11

How-To

How to Turn On Variable Refresh Rate (VRR) in Windows 11

Windows 11's Variable refresh rate switch enables VRR for DirectX 11 full-screen games that don't support it natively, and Optimizations for windowed games helps windowed and borderless DX10/11 games use VRR. Turn on FreeSync or G-SYNC first; the VRR switch only appears with suitable drivers and a VRR monitor.

Forge PC Staff2 min