Noen Intelligence’s ¥500M Angel++: Matrix China Bets on Foldable Home Robots
Matrix Partners China led Noen (Knowin) Intelligence’s 500 million RMB Angel++ round for GLOW embodied models and the sub-40cm foldable KNOWIN-X1 home robot — Redpoint among participants.
Noen Intelligence (诺因智能; product brand Knowin) closed a 500 million RMB Angel++ round reported August 10, 2026. Matrix Partners China led; Redpoint Ventures, SenseTime Guoxiang Capital, Walden International, and L2F Lighthouse Founders’ Fund participated. The money funds the GLOW generative embodied model stack and mass-production prep for KNOWIN-X1 — a foldable dual-arm home robot that collapses under 40 cm.
Key facts
| Field | Detail |
|---|---|
| Company | Noen Intelligence / Knowin (founded Aug 2025) |
| Round | ¥500M Angel++ (~$70M indicative) |
| Date | Reported August 10, 2026 |
| Lead | Matrix Partners China |
| Notable participants | Redpoint Ventures, SenseTime Guoxiang, Walden International, L2F |
| Team | Li Yinchuan (ex–Huawei Noah generative LM lead); Wang Yunjie (ex-DJI); Zhou Kaiwen (ex–Huawei Agent lead); ~190+ people as of Jul 2026 |
| Product | GLOW model system + KNOWIN-X1 foldable home robot |
| Launch target | 2027; “tens of thousands of RMB” price band |
Who uses the product — and for what job
Users (target): affluent Chinese households and geek early adopters who will tolerate a teachable robot — not a magical butler on day one.
Jobs:
- High-frequency home assistance (desktop sorting, item moves, appliance interaction)
- Companionship with long-term memory / greeting
- On-duty perception (safety reminders, elderly-care adjacent monitoring)
Company design choice: users demonstrate tasks (wipe a table once) rather than expecting full out-of-box autonomy — a deliberate honesty about generalization limits.
Hardware claim: fully foldable dual-arm form factor so the robot can live in a trunk or corner; three postures (lying / kneeling / standing) for floor, desk, and cabinet work; dual-arm max load 6 kg; 23 DoF whole machine (company specs via 36Kr).
Why now
- Embodied AI capital is rotating from industrial demos toward consumer home — harder environment, larger TAM if generalization works.
- Home long-tail tasks cannot be fully enumerated; generative data + world models + real-machine feedback is the emerging closed loop.
- Huawei/DJI talent density in Shenzhen/China robotics makes full-stack (model + body) teams fundable at Angel++ scale.
- CVPR 2026 EgoCross / Embodied Arena leaderboard claims (company via 36Kr) give investors a research signal before revenue.
Why Matrix China + Redpoint — portfolio fit
Matrix Partners China is underwriting the non-consensus bet: home is the hardest and most valuable embodied scenario. Their quote centers on continuous generalization and a model-to-product closed loop.
Redpoint Ventures participates as a global firm fluent in robotics and consumer platforms. The public Redpoint comment (via 36Kr) stresses learn-in-use flywheels — usage → data → model → experience — not skill count at unboxing.
Likely founder rationale: keep Day-1 China deep-tech partners (L2F says four consecutive rounds) and Matrix’s lead, while adding Redpoint for international pattern recognition and SenseTime-adjacent AI distribution.
| Investor | Fit |
|---|---|
| Matrix China | Lead conviction on China consumer embodied AI |
| Redpoint | Global robotics / product network |
| SenseTime Guoxiang | AI/model ecosystem adjacency |
| L2F / Walden | Early believer continuity + hardware fluency |
Matrix Partners China is not currently a /fund/ entity here — only Redpoint maps to our directory.
Competitive map
| Player | Lane |
|---|---|
| Industrial humanoids (Apptronik, Figure, etc.) | Factory/logistics first; different form factor |
| Vacuum / appliance robots | Narrow tasks; not general dual-arm agents |
| Other China home embodied startups | Same race; differentiation is foldable form + GLOW loop |
| U.S. home robot attempts (historical) | Cautionary: form factor + reliability killed categories |
Market signal
¥500M at Angel++ before a 2027 launch prices team + closed-loop thesis, not booked robot revenue. That is appropriate for hardware+model — and dangerous if treated like SaaS ARR.
When not to use this as a template
- Wrong if “world’s first foldable” marketing is treated as a moat without shipping reliability data.
- Wrong if 2027 launch guidance is modeled as 2026 revenue.
- Wrong if home robots are underwritten like warehouse AMRs with controlled floors.
Practical takeaway
- Founders (robotics): Sell foldability + teachability + model loop — not a demo reel of cherry-picked chores.
- Investors: Underwrite generalization failure modes and BOM cost at “tens of thousands RMB.”
- Operators / scouts: Compare to industrial humanoid capital; home is a different buyer and SLA.
Sources
- 36Kr (Aug 10, 2026): https://eu.36kr.com/en/p/3933076448558212
- Knowin tech page: https://www.knowinai.com/tech.html
- Related: /fund/redpoint-ventures · /startup/noen-intelligence · /2026-august-9-10-investment-news-chips-infra-biotech
Follow Venture Capital Tracker in Google
Add VCT as a preferred source to make our venture-capital coverage easier to find in Google Search.
Editorial note: AI tools assisted with research, structure, or drafting. Venture Capital Tracker retains human editorial responsibility for factual accuracy, relevance, and source quality before publication.
