DEPLOYDatabase

Robot model

Zhenwu 810E

otherproductionby Alibaba T-Head

T-Head's Zhenwu 810E is a general-purpose GPU (PPU) AI accelerator card with 96 GB HBM2e memory, 700 GB/s interconnect bandwidth, PCIe 5.0 x16 interface, and…

Track Zhenwu 810E
Manufacturer
Alibaba T-Head
Form factor
other
Maturity
production
Lifecycle
active
Deployments
2

The verified answer

As of Oct 2026, DEPLOY verifies Zhenwu 810E, a other by Alibaba T-Head (production): 2 verified deployments on record. 2 sources back the record.

Overview

T-Head's Zhenwu 810E is a general-purpose GPU (PPU) AI accelerator card with 96 GB HBM2e memory, 700 GB/s interconnect bandwidth, PCIe 5.0 x16 interface, and 400W power consumption. Performance is comparable to NVIDIA's H20, surpassing the A800. It became one of China's highest-shipping domestic GPUs in 2025, with T-Head delivering 470,000 AI chips cumulatively by February 2026.

Verified vs. claimed

Maturity stage
production(Full-scale production deployment with repeat customers.)
Sources on file
View all sources →

Key facts

Memory and interconnect

Yicai reported on Jan 29 2026 that T-Head published product information for its Zhenwu 810E AI chip: 96 GB of HBM2e and inter-chip bandwidth of up to 700 GB/s, for multimodal model training, inference, and autonomous driving. Treat those figures as T-Head-published, as relayed by Yicai (Yicai).

Qianwen use and customers

The same report says Yicai learned from Alibaba that Zhenwu 810E is already used for training and inference of Alibaba's Qianwen large language model, and that together with Alibaba cloud products it was serving more than 400 clients, including State Grid, the Chinese Academy of Sciences, and XPeng. Treat the deployment claims as company-stated via Yicai (Yicai).

Data & sources

Web sources

2

2 sources backing this record.View all →

Pricing

No verified price is on record for Zhenwu 810E. Physical-AI systems are often sold through enterprise contracts or operated as a service rather than at a public list price.

How DEPLOY tracks pricing

Deployments (2)

  • Zhenwu 810E at Qinghai Province, Chinaoperational

    China Unicom deploying T-Head (Pingtouge) AI accelerators at its Sanjiangyuan data center in Qinghai province, alongside chips from MetaX and Biren Technology.

  • Zhenwu 810E at Hangzhou, Zhejiang, Chinaoperational

    T-Head chips deployed across Alibaba Cloud data centers in Hangzhou for AI inference and training workloads including Qwen model training.

Zhenwu 810E on the deployment map

View the global map

Where Zhenwu 810E is verified operating. Explore the deployment map by place and type.

Free to use on your site, with a link back to DEPLOY.

Recent activity

Every change to this record is dated, sourced, and independently verified where marked.

Full change history →

Safety record

No incidents on record for Zhenwu 810E.

Only active incidents are counted. Retracted incidents are excluded from this summary but remain reachable at their canonical URLs.

Sources (2)

  1. Alibaba T-Head Unveils New AI Chip Said to Match NVIDIA H20 as IPO Speculation Builds · https://www.trendforce.com/news/2026/01/29/news-alibaba-t-head-unveils-new-ai-chip-said-to-match-nvidia-h20-as-ipo-speculation-builds/
  2. https://www.yicaiglobal.com/news/t-head-unveils-specs-of-its-ai-chip-after-rumors-alibabas-unit-is-preparing-for-ipo

Common questions

What is Zhenwu 810E?
T-Head's Zhenwu 810E is a general-purpose GPU (PPU) AI accelerator card with 96 GB HBM2e memory, 700 GB/s interconnect bandwidth, PCIe 5.0 x16 interface, and 400W power consumption. Performance is comparable to NVIDIA's H20, surpassing the A800. It became one of China's highest-shipping domestic GPUs in 2025, with T-Head delivering 470,000 AI chips cumulatively by February 2026.
How much does Zhenwu 810E cost?
Zhenwu 810E's price is not publicly disclosed. DEPLOY has no verified price on record for Zhenwu 810E from Alibaba T-Head. Physical-AI systems like this are often sold through enterprise contracts or operated as a service rather than at a public list price; check the manufacturer for the latest.
Is Zhenwu 810E actually deployed in the real world?
Yes. Zhenwu 810E is independently verified in real-world operation on the DEPLOY registry, confirmed at named deployment sites with primary sources: not a concept, render, or demo-only.
Who makes Zhenwu 810E?
Zhenwu 810E is made by Alibaba T-Head, based in Hangzhou, Zhejiang, China, founded in 2018.
Where is Zhenwu 810E deployed?
2 verified deployments of Zhenwu 810E are on the DEPLOY registry, including at Qinghai Province, China, Hangzhou, Zhejiang, China.
Can you buy Zhenwu 810E?
Zhenwu 810E is in commercial deployment and is sold to operators; contact the manufacturer to buy, and see the pricing section for any recorded price.
What are alternatives to Zhenwu 810E?
On the DEPLOY registry, comparable other robots to Zhenwu 810E include ASML EUV TwinScan, BrainOS, 10Beauty Manicure System, CleanSpark Bitcoin Mining Fleet.
How does Zhenwu 810E compare to ASML EUV TwinScan?
Zhenwu 810E and ASML EUV TwinScan (ASML Holding · 7 deployments) are both other robots on the DEPLOY registry. Zhenwu 810E has 2 verified deployments on record. Compare both records for specs, safety, deployments, and verified-vs-claimed autonomy.
Is Zhenwu 810E a top other?
On DEPLOY's intelligence score, which blends verified deployments, safety, adoption, media, and IP signals, Zhenwu 810E ranks in roughly the top 26% of other models tracked by the registry.
What is Zhenwu 810E's maturity stage?
Zhenwu 810E is at the production stage on the DEPLOY maturity ladder (research, prototype, pilot, commercial, production). Production stage means high-volume manufacture and commercial-scale deployments are sustained.
Is Zhenwu 810E safe?
Zhenwu 810E has no incidents on record in the DEPLOY registry. This does not constitute a safety guarantee; it reflects the incidents DEPLOY has tracked and verified to date.
Methodology: Verified · 2 sources (no primary) · last reviewed 2026-10-03

Verification posture

Verified

Low confidence

Review state

Stable

Last reviewed 2026-10-03

Maturity + lifecycle

Maturity stage: production

Lifecycle: active

Architectural position

Cohort: other

Sources by quality tier

2
unclassified
Unclassified source

The framework is documented at /methodology. Corrections at /corrections. Reviewer: DEPLOY editorial team.

Methodology surface for Zhenwu 810E.

Recent coverage

Zhenwu 810E in third-party press

Dark Horse Muka Robotics Charges WorldArena With Only 32 Zhenwu 810E GPUs: LJM World Model Achieves Global Second Place Defining the Breakthrough From Video Quality to Physical Logic

Jul 30, 2026