EngineAI / 众擎机器人 public evidence packet v1
Date: 2026-06-12 Owner: Hugo / Genius Team Research owner: Finance / Charlie AGT-002 Visibility: PUBLIC Status: source-backed evidence packet Public-safety flag: evidence map only; no trade recommendation; no private portfolio information.
1. One-line answer
EngineAI / 众擎机器人值得从 robotics-companies.md 的 open validation pool 提升为“China OEM watch card”：截至 2026-06-12，它已经有官网可验证的多产品矩阵、2023-10 成立锚点、PM01 / SE01 / T800 / SA01 等产品参数、以及 2026-05-22 “首批 T800 humanoids officially rolls off the production line / toward 10,000-unit scale delivery” 新闻标题；但这仍是 S3/S4 产品化 + 量产意图证据，不是 S5 scaled commercial economics。🟢 EngineAI official website / product / news pages, accessed 2026-06-12.
2. Core question
EngineAI 是一个能把高动态 humanoid 产品矩阵转化为真实客户经济性的中国 OEM，还是一个展示能力强、但尚未披露客户 / 交付 / 收入 / 毛利验证的产品化平台？
3. Market-definition lens
- Layer: robot body / OEM + embodied AI hardware platform.
- Adjacent layer: developer / open platform, because SA01 和 PM01 官方页面都强调开放、可扩展、低层接口 / code support。🟢 product-sa01 / product-pm01 pages, accessed 2026-06-12.
- Not counted here: component supplier economics, software ARR, or public-equity expression. EngineAI is private and no reviewed filing-level financials were found in this pass.🟠 evidence packet scope.
4. Current evidence
4.1 Company base
| Claim | Evidence | Grade | As-of |
|---|---|---|---|
| 公司主体为 EngineAI Robotics Technology Co., Ltd. / 众擎机器人，总部深圳，成立于 2023-10。 | About page says headquartered in Shenzhen and established in October 2023. Development-history page says “October 2023: Company Establishment / 2023-10-07”. | 🟢 primary | accessed 2026-06-12 |
| 业务覆盖 humanoid robots and related products 的研发、制造和场景部署，目标推进 intelligent robotics commercial application globally。 | About page company description. | 🟢 primary | accessed 2026-06-12 |
| 场景包括 scientific research and education, industrial manufacturing, commercial services, and home companionship。 | About page product/service description. | 🟢 primary | accessed 2026-06-12 |
4.2 Product matrix and specifications
| Product | Official positioning / specs | Evidence type | Signal read |
|---|---|---|---|
| PM01 | “lightweight, highly-dynamic, and fully-open embodied intelligent agent”; homepage lists 138 cm height, 24-DOF total DOF, 130 N·m/kg peak torque density. Product page says integrated joint module peak torque 130 N·m, ≥23 total DOF, >300° waist rotation, low-level hardware interfaces and code support for training/deployment. | 🟢 primary product page, accessed 2026-06-12 | S3 developer / platform signal; not customer economics. |
| SE01 | “world's first general-purpose humanoid robot to achieve a human-like walking gait”; homepage lists 170 cm height, 2 m/s normal walking speed, 32 pcs total DOF. Product page says LiDAR + depth camera + infrared camera perception matrix and five-finger dexterous hand. | 🟢 primary product page, accessed 2026-06-12 | S3 motion / product capability signal; “world’s first” is company claim and should not be treated as independently verified market fact. |
| T800 | “full-size high-dynamic general-purpose robot”; homepage lists 173 cm height, 29-DOF excluding dexterous hand, up to 450 N·m joint torque. Product page says active cooling enables 4 hours continuous operation, 360° LiDAR, 7 DOF per dexterous hand, 5 kg hand payload, tactile sensing. | 🟢 primary product page, accessed 2026-06-12 | S4 if production-line news is confirmed by official article; S3/S4 productized full-size platform signal pending delivery/customer details. |
| SA01 | “open-source, highly-expandable bipedal robot platform”; homepage lists 15 Ah lithium battery, 12 total DOF, S160 proprietary self-developed motors. Product page says fully open-source design, complete hardware architecture and interfaces available, supports compatible modules such as robotic arms. | 🟢 primary product page, accessed 2026-06-12 | S3 developer/research platform signal. |
| JS01 | “heavy-duty robotic hound ‘Leopard’”; product page says load capacity up to 80 kg and top speed 3 m/s. | 🟢 primary product page, accessed 2026-06-12 | Adjacent robotics signal; not humanoid-specific commercialization proof. |
| S2 | “Original IP Service Robot”; product page says stay tuned. | 🟢 primary product page, accessed 2026-06-12 | S2/S3 placeholder; little material evidence. |
4.3 Development-history and production-intent signals
| Date | Official evidence | Grade | Signal level | Why it matters | What it does not prove |
|---|---|---|---|---|---|
| 2023-10-07 | Company establishment. | 🟢 | S3 | Gives founding-date anchor; young company moving quickly. | Does not prove manufacturing quality or customer adoption. |
| 2024-07-29 | SA01 launched; “first open-source platform”. | 🟢 | S3 | Developer-platform / research access angle. | Does not prove paid deployment or margin. |
| 2024-10-24 | SE01 flagship product launched; company claims human-like natural gait via end-to-end neural network model. | 🟢 / 🟠 implication | S3 | Motion-control and model-stack positioning. | Company claim is not customer-site KPI. |
| 2024-12-24 | PM01 launched; Hefei branch launch “accelerated R&D and large-scale assembly progress, achieving industry-leading benchmarks in mass production and delivery.” | 🟢 for text / 🟠 for implication | S3/S4 | Suggests manufacturing setup beyond demos. | No units, yield, customer count, or revenue disclosed in the extracted text. |
| 2025-01-07 | CES debut with SA01 / SE01 / PM01. | 🟢 | S3 | International visibility and product-cadence evidence. | Trade-show presence is not deployment economics. |
| 2025-02-23 | Company says humanoid front flip stunt achieved. | 🟢 for company claim / 🟠 implication | S3 | High-dynamic control signal. | Stunts are not factory reliability or customer ROI. |
| 2025-05-22 | JS01 released at National AI Expo. | 🟢 | S3 | Broadens product family. | Not humanoid commercialization. |
| 2026-05-22 | News page headline: “EngineAI Robotics embarks on a new milestone toward 10,000-unit scale delivery as the first batch of T800 humanoids officially rolls off the production line.” | 🟢 headline from official news page | S4 watch item | This is the most material current signal: a first-batch / production-line / 10,000-unit scale-delivery intent claim. | Headline alone does not disclose delivered units, accepted customer units, contract value, recognized revenue, yield, gross margin, uptime, intervention rate, or customer ROI. |
5. Current classification
Current label: S3/S4 productized-platform + production-intent evidence, not S5 scaled commercial economics.
Why not S5 yet:
- No public financial disclosure reviewed for robot revenue, gross margin, backlog, cash burn, or customer concentration. 🟠 based on this source pass, as of 2026-06-12.
- The official site gives product specs and production-intent language, but not customer-accepted units, contract value, uptime, intervention rate, customer ROI/payback, repeat orders, warranty/service cost, or production yield. 🟢 official pages for disclosed fields; 🟠 gap analysis.
- “Toward 10,000-unit scale delivery” is directionally stronger than a demo, but cannot be charted as actual 10,000 delivered units without a detailed article, filing, customer acceptance data, or audited shipment/revenue evidence. 🟢 official headline; 🟠 interpretation.
6. Signal vs noise
Signal
- S4 watch: T800 first-batch roll-off / “toward 10,000-unit scale delivery” official news headline dated 2026-05-22.🟢
- S3/S4: T800 product specs — 173 cm, 29-DOF excluding dexterous hand, up to 450 N·m joint torque, active cooling for 4 hours continuous operation, 360° LiDAR, 7-DOF dexterous hands, 5 kg hand payload.🟢
- S3: PM01 / SA01 open platform positioning, interfaces, and code / hardware openness.🟢
- S3: SE01 humanoid gait and perception / dexterous hand product claims.🟢
Noise / do not overstate
- Viral videos, stunts, CES presence, and “world’s first” language should stay S3 unless tied to deployment metrics.🟠
- Product breadth is not customer traction.🟠
- First batch / production-line headline is not full commercialization unless units, acceptance, revenue, margin, and customer economics are disclosed.🟠
7. What would upgrade EngineAI to S5?
EngineAI would become S5 only if public sources show at least several of the following:
- Delivered / accepted humanoid units by customer or application, not just production intent.
- Contract value, revenue recognized, backlog, or recurring service/software economics.
- Customer-confirmed runtime, uptime, intervention rate, failure rate, safety record, and ROI/payback.
- Production yield, first-pass quality, warranty / service cost, and gross margin.
- Repeat orders from named customers or multi-site deployments.
- Financing runway and cash-burn transparency if growth is capital intensive.
8. Counter-thesis / bear case
- EngineAI may be excellent at high-dynamic demos and product cadence, but still far from reliable real-world deployments.🟠
- The product matrix may fragment resources across humanoid, bipedal platform, quadruped, and service robot lines before a single economic wedge is proven.🟠
- If 10,000-unit scale delivery language is mostly production ambition rather than accepted customer shipments, the evidence remains S4 at best.🟠
- China humanoid OEM competition is crowded; low prices, rapid demos, and product launches can compress margins before durable customer demand appears.🟠
9. Common misconceptions
-
Misconception: “T800 first batch rolled off line means 10,000 robots have shipped.” Correction: the reviewed official news page headline says “toward 10,000-unit scale delivery” and first batch T800 humanoids rolled off the production line; it does not, from the extracted headline alone, disclose 10,000 accepted deliveries, revenue, or customer economics.🟢/🟠 as of 2026-06-12.
-
Misconception: “High-dynamic movement proves commercialization.” Correction: movement capability is useful, but commercialization requires uptime, intervention rate, maintenance cost, customer workflow fit, payback, and repeat orders.🟠 framework.
-
Misconception: “Open-source / developer platform means software value capture is proven.” Correction: openness can expand experimentation, but value capture still needs developer adoption, paid tools/services, attach rate, or deployment economics.🟠 framework.
10. Think Deeper questions
- If EngineAI can truly move toward 10,000-unit scale, which bottleneck appears first: production yield, battery / thermal reliability, customer deployment, or working capital?
- Is EngineAI building a broad robot product company, or is T800 the only path that matters for humanoid commercialization?
- Does China’s low-cost high-cadence humanoid OEM cluster create more value for OEMs, suppliers, data/model layers, or deployment integrators?
11. Public-safe site draft section
EngineAI / 众擎：产品化速度很快，S5 还要等客户经济性
EngineAI is one of the China humanoid OEMs worth tracking because the public evidence is no longer just a video feed. The company’s official site, accessed 2026-06-12, shows a multi-product robot family: PM01, SE01, T800, SA01, JS01, and S2. The strongest humanoid evidence is T800: 173 cm height, 29 DOF excluding the dexterous hand, up to 450 N·m joint torque, active cooling for 4 hours continuous operation, 360° LiDAR, and 7-DOF dexterous hands with 5 kg payload. The most material news signal is the 2026-05-22 official headline saying the first batch of T800 humanoids rolled off the production line and the company is moving toward 10,000-unit scale delivery.
That is stronger than a pure demo cycle, but it is not S5 yet. The missing evidence is still customer-accepted units, contract value, runtime, uptime, intervention rate, service cost, production yield, robot revenue, gross margin, and repeat orders. Treat EngineAI as a productized-platform / production-intent watch item, not as proof that humanoid economics are solved.
Footer: Evidence map only. No winner ranking. No trade recommendation. Production-intent language ≠ delivered units or customer economics.
12. Source list
- EngineAI official homepage, https://en.engineai.com.cn/ and https://www.engineai.com.cn/, accessed 2026-06-12. 🟢 primary.
- EngineAI About page, https://en.engineai.com.cn/about-us.html, accessed 2026-06-12. 🟢 primary.
- EngineAI development-history page, https://en.engineai.com.cn/about-process-sa01.html, accessed 2026-06-12. 🟢 primary.
- EngineAI news page, https://en.engineai.com.cn/about-news.html, accessed 2026-06-12. 🟢 primary.
- EngineAI PM01 product page, https://en.engineai.com.cn/product-pm01.html, accessed 2026-06-12. 🟢 primary.
- EngineAI SE01 product page, https://en.engineai.com.cn/product-se01.html, accessed 2026-06-12. 🟢 primary.
- EngineAI T800 product page, https://en.engineai.com.cn/product-t800.html, accessed 2026-06-12. 🟢 primary.
- EngineAI SA01 product page, https://en.engineai.com.cn/product-sa01.html, accessed 2026-06-12. 🟢 primary.
- EngineAI JS01 product page, https://en.engineai.com.cn/product-js01.html, accessed 2026-06-12. 🟢 primary.
- EngineAI S2 product page, https://en.engineai.com.cn/product-s2.html, accessed 2026-06-12. 🟢 primary.
13. Public-safety note
This artifact uses only public official pages and internal framework interpretation. It contains no private portfolio weights, no trade recommendation, and no supplier-customer inference.