{"id":"0bb157bd-bd02-4317-9f2e-5151806abb0c","companyId":"f0cd4a91-4318-42ee-a342-d1c0932f1edd","modelName":"Zhenwu 810E","slug":"zhenwu-810e","description":"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.","formFactor":"other","industryVertical":null,"environment":null,"maturityStage":"production","lifecycleState":"active","supersededByModelId":null,"specs":{},"manufacturerSerial":null,"reviewStatus":"reviewed","sources":[{"url":"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/","label":"Alibaba T-Head Unveils New AI Chip Said to Match NVIDIA H20 as IPO Speculation Builds","title":"Alibaba T-Head Unveils New AI Chip Said to Match NVIDIA H20 as IPO Speculation Builds","mediaType":"article","sourceName":"TrendForce"},{"url":"https://www.yicaiglobal.com/news/t-head-unveils-specs-of-its-ai-chip-after-rumors-alibabas-unit-is-preparing-for-ipo","date":"2026-01-29","title":"T-Head unveils Zhenwu 810E specs","sourceName":"Yicai"}],"aliases":[],"collisionRisk":"low","reviewNote":null,"manufacturerTermForTeleop":null,"keyFacts":[{"label":"Memory and interconnect","value":"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](https://www.yicaiglobal.com/news/t-head-unveils-specs-of-its-ai-chip-after-rumors-alibabas-unit-is-preparing-for-ipo))."},{"label":"Qianwen use and customers","value":"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](https://www.yicaiglobal.com/news/t-head-unveils-specs-of-its-ai-chip-after-rumors-alibabas-unit-is-preparing-for-ipo))."}],"createdAt":"2026-08-25T05:03:05.211Z","updatedAt":"2026-10-03T18:19:00.325Z","unitsInFieldClaimed":null,"unitsInFieldVerified":null,"unitsManufacturedClaimed":null,"unitsManufacturedVerified":null,"unitsAsOf":null,"unitsSource":null,"salesModel":null,"priceUndisclosedSince":null,"autonomyClass":null,"autonomyStatement":null,"autonomyStatementSource":null,"demoForensics":null,"exportControlClass":null,"minCommitment":null,"warrantyTerms":null,"serviceModel":null,"productionRateClaimed":null,"productionRateEvidenced":null,"teleopRatioDisclosed":null,"interventionRateDisclosed":null,"jsonLd":{"@context":"https://schema.org","@type":"Product","@id":"https://registry.deploy.report/models/zhenwu-810e","url":"https://registry.deploy.report/models/zhenwu-810e","name":"Zhenwu 810E","description":"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.","identifier":"0bb157bd-bd02-4317-9f2e-5151806abb0c","category":"other","publisher":{"@id":"https://deploy.report/#organization"}},"framework_metadata":{"framework_schema_version":"0.1.0","verification_status":"verified","maturity_stage":"production","lifecycle_state":"active","architectural_position":{"cohort":"other","sub_cohorts":[]},"within_cohort_verified_vs_claimed_pair":null,"cap_flags":[],"verification_depth":{"sources_count":2,"primary_source_types":[]}}}