Moore Threads

[PUBLIC]

HQ: Beijing, China · Est. 2020 · 1 model registered

Recent press

As of Sep 2026, DEPLOY verifies Moore Threads: 1 model, 1 active incident. 4 sources back the record.

Ownership

PUBLIC

Capital / Funding

Undisclosed

Headquarters

Beijing, China

IP / Patents

Undisclosed

Platform Models

1

Audited Incidents

0

Overview

Moore Threads is a Beijing-based domestic Chinese GPU designer that pairs its own silicon (the MTT S-series) with a full-stack CUDA-alternative software layer (MUSA SDK) and a training-oriented cluster reference architecture called KUAE. Founded in October 2020 by former Nvidia China vice president Zhang Jianzhong, the company sits in the small group of firms trying to replace Nvidia inside China now that H100/H200/B-series parts are export-controlled. It was added to the U.S. Entity List in October 2023, which severed its access to leading-edge TSMC nodes and forced it onto older process technology.\n\nUnlike Western neoclouds such as CoreWeave or Lambda, Moore Threads is not primarily a GPU-rental operator; it is a chip vendor whose cloud presence takes the form of KUAE reference-design AI compute centers built with Chinese telco and enterprise partners, plus a MUSA cloud service for developers. Its go-to-market is domestic-only: Chinese model labs, state-owned enterprises, universities, and provincial "intelligent computing" projects that need to run large-model training on non-U.S. silicon.\n\nAfter years of private funding from ByteDance, Tencent, Sequoia Capital China, GGV, and Shenzhen Capital Group, Moore Threads listed on the Shanghai Stock Exchange's STAR Market in December 2025, raising more than one billion dollars in what became one of the year's most-watched Chinese semiconductor IPOs. The listing crystallized its position as the highest-profile publicly traded pure-play Chinese GPU maker, with a business model that mixes silicon sales, KUAE cluster deployments, and a smaller cloud/consumer product line.

Audited Posture AssessmentPeer Percentile: 99.3%

Safety profile within acceptable range for category. Well-capitalized with $1.6B raised and recent funding activity. Limited public visibility and press coverage relative to peers.

  • PUBLIC

Key Facts

CEOZhang Jianzhong (James Zhang)
HeadquartersBeijing, China
FoundedOctober 2020
StockSSE: 688795 (IPO Dec 5, 2025)
Co-founderZhang Yubo
ArchitectureMUSA (Moore Threads Unified System Architecture)
Key ProductsMTT S60, MTT S70, MTT S80 (desktop GPU); MTT S2000 (server GPU)
US Entity ListAdded October 2023 by US Dept of Commerce Bureau of Industry and Security
IPO Proceeds~$1.1 billion (Dec 2025)
Revenue FY2024438 million yuan (~$61M USD), 252% YoY growth
Revenue H1 2025~702 million yuan (~$97M USD)
Notable InvestorsSequoia Capital China, GGV Capital, Shenzhen Capital Group, ByteDance, Tencent, High-Flyer/DeepSeek (Liang Wenfeng)
Prior CompanyCEO Zhang Jianzhong was Nvidia global VP & GM China (2006-2020)
HQ + foundedBeijing, China; October 2020
Founder and CEOZhang Jianzhong, former Nvidia China vice president
Public statusListed on Shanghai Stock Exchange STAR Market, December 2025; shares rose 468% on debut
IPO proceedsRaised over $1 billion in December 2025 STAR Market listing
Silicon lineupMTT S60 and S80 consumer/workstation GPUs (12nm); MTT S2000, S3000, S4000 data-center parts; MTT S5000 unified training and inference card announced 2025
Cluster reference designKUAE (Kua'e) large-model pretraining platform, positioned as the customer-deployable AI compute center built on MTT S4000/S5000 GPUs; not a hyperscale on-demand cloud
Notable pre-IPO investorsByteDance, Tencent, Sequoia Capital China, GGV Capital, Shenzhen Capital Group
U.S. export statusAdded to U.S. Commerce Department Entity List October 2023, cutting access to leading-edge TSMC production
PositioningDomestic-China GPU vendor plus KUAE cluster integrator; training-focused with a growing inference story via S5000; not a global GPU-rental neocloud in the CoreWeave/Lambda sense
Software stackMUSA SDK as the CUDA-alternative programming layer, with cloud-native GPU virtualization and AI training/inference suites

Safety & Regulatory Posture

0 entries
Audited Incident & Recall FootprintLatest event: Oct 2023

1 incident on record (1 serious). Most recent: Oct 2023.

serious: 1

Leadership & Founders

2 entries

Current Program Leadership & Executives

Zhang Jianzhongsecondary-verified

CEO & Founder, Founder & CEO

Zhang Yubosecondary-verified

Co-founder

Partnerships & Commercial Deployments

0 entries
Total Tracked0
Operating0
Announced0
Concluded0

FAQ

12 entries
What is Moore Threads?

Moore Threads is a Beijing-based domestic Chinese GPU designer that pairs its own silicon (the MTT S-series) with a full-stack CUDA-alternative software layer (MUSA SDK) and a training-oriented cluster reference architecture called KUAE. Founded in October 2020 by former Nvidia China vice president Zhang Jianzhong, the company sits in the small group of firms trying to replace Nvidia inside China now that H100/H200/B-series parts are export-controlled. It was added to the U.S. Entity List in October 2023, which severed its access to leading-edge TSMC nodes and forced it onto older process technology.\n\nUnlike Western neoclouds such as CoreWeave or Lambda, Moore Threads is not primarily a GPU-rental operator; it is a chip vendor whose cloud presence takes the form of KUAE reference-design AI compute centers built with Chinese telco and enterprise partners, plus a MUSA cloud service for developers. Its go-to-market is domestic-only: Chinese model labs, state-owned enterprises, universities, and provincial "intelligent computing" projects that need to run large-model training on non-U.S. silicon.\n\nAfter years of private funding from ByteDance, Tencent, Sequoia Capital China, GGV, and Shenzhen Capital Group, Moore Threads listed on the Shanghai Stock Exchange's STAR Market in December 2025, raising more than one billion dollars in what became one of the year's most-watched Chinese semiconductor IPOs. The listing crystallized its position as the highest-profile publicly traded pure-play Chinese GPU maker, with a business model that mixes silicon sales, KUAE cluster deployments, and a smaller cloud/consumer product line.

What does Moore Threads make?

Moore Threads has 1 robot model on the DEPLOY registry: MTT S80 (Moore Threads builds both physical robots and the AI that runs them).

Is Moore Threads publicly traded?

Yes. Moore Threads is publicly traded on public markets, per the DEPLOY registry.

Who competes with Moore Threads?

On the DEPLOY registry, peer companies to Moore Threads building in the same form factors include Roborock, Unitree Robotics, DJI, DLR Institute of Robotics and Mechatronics.

Who is the CEO of Moore Threads?

Zhang Jianzhong is the ceo of Moore Threads, per the DEPLOY registry.

How can I invest in Moore Threads?

Moore Threads is publicly traded (public markets), so its shares can be bought through any broker.

Where is Moore Threads headquartered?

Moore Threads is headquartered in Beijing, China.

Who owns Moore Threads?

Moore Threads is publicly held by its shareholders.

Where does Moore Threads operate robots?

Moore Threads is a robot manufacturer and does not directly operate deployments on the DEPLOY registry. Its products are deployed by customer operators; see the company page for the verified deployment record.

Is Moore Threads a top robotics company?

On DEPLOY's intelligence score, which blends verified deployments, funding, hiring, media, safety, and IP signals, Moore Threads ranks in roughly the top under 1% of companies tracked by the registry.

Are there any incidents involving Moore Threads?

1 active incident involving Moore Threads is on the DEPLOY registry. Each is a sourced, append-only record; retracted incidents are suppressed from this view.

Is Moore Threads safe?

Moore Threads has 1 active incident on record in the DEPLOY registry. 1 incident on record (1 serious). Most recent: Oct 2023. Retracted incidents are excluded from this count.