The archive · AI & Models · Strategic decision · 2025–
Simplexity Robotics bets Li Auto's playbook can win embodied AI: ¥2B in five rounds
Ex-Li Auto smart-driving team founded Simplexity in Jul 2025; five rounds in six months raised ¥2B at a $1B+ valuation, the youngest embodied-AI unicorn
Simplexity Robotics（至简动力）
What the business is
Simplexity Robotics develops embodied-AI robots for factories, supermarkets and logistics: a unified world-model/VLA 'brain' (LaST0, ManualVLA), an on-device data-training loop, and one in-house-designed robot body, moving from closed to open environments.
Starting capital：¥2 billion across five rounds in under six months (to March 2026), with Yuanjing, Lanchi, Hongshan, Legend Capital, Zhongke Chuangxing and Gaorong investing; Tencent and Alibaba as strategic backers; post-money valuation over $1 billion
How it started
Jia Peng, former head of Li Auto's smart-driving technology R&D, founded Simplexity with Wang Kai (ex-Li Auto CTO, previously a partner at Yuanjing Capital) and Wang Jiajia (former Bosch China R&D director who led Li Auto's full smart-driving mass production). The company registered in Hangzhou in late July 2025, and its first robot body went from first hire to prototype in under 45 days.
What happened
Within six months Simplexity closed five rounds totaling ¥2 billion — financial investors Yuanjing, Lanchi, Hongshan, Legend Capital, Zhongke Chuangxing and Gaorong, plus strategic backers Tencent and Alibaba — reaching a post-money valuation above $1 billion, the youngest unicorn in the embodied-AI race. It shipped LaST0 (world-model + VLA foundation model), ManualVLA (long-horizon tasks, a CVPR 2026 paper) and TwinRL (digital-twin reinforcement learning reaching 100% success on desktop tasks in under 20 minutes), completed two generations of robot bodies in small-batch production, and opened Beijing, Shanghai and Suzhou sites.
No ending yet — it is still running.
Background
Simplexity Robotics (至简动力) is a Hangzhou-based embodied-intelligence company registered in late July 2025 by the core team behind Li Auto's smart-driving program — CEO Jia Peng (former head of Li Auto intelligent-driving R&D), chairman Wang Kai (ex-Li Auto CTO) and COO Wang Jiajia (former Bosch China R&D director who ran Li Auto's smart-driving mass production). Its bet is that the playbook that won in cars can win in robots.
The company's architecture is 'one model, on device, one body': a single transformer fusing world model and VLA so the robot can reason about physics and act fast (LaST0), a long-horizon task model (ManualVLA, accepted to CVPR 2026), and TwinRL, a digital-twin reinforcement-learning framework that reached 100% success on desktop tasks in under 20 minutes. Compute is embedded on the robot so data collection, training and validation happen in the user's real environment, mirroring Tesla's shadow mode.
Between late 2025 and March 2026 the company closed five rounds totaling ¥2 billion — Yuanjing, Lanchi, Hongshan, Legend Capital, Zhongke Chuangxing and Gaorong as financial investors, Tencent and Alibaba as strategic backers — reaching a post-money valuation above $1 billion and becoming the youngest unicorn in the embodied-AI sector.
From first hire to first robot body took under 45 days, and by early 2026 it had completed two generations of bodies for B2B and consumer use, started small-batch production and PoC validation, and opened Beijing, Shanghai and Suzhou sites. Its roadmap moves from factories and supermarkets to logistics and homes, from closed to open environments.
What has to be true
- Team credibility: investors were betting on people who had survived the brutal smart-driving war and shipped at scale, not on an unproven prototype.
- Unified-model bet: fusing world model and VLA in one transformer aims for higher generalization and better scaling than modular feature-by-feature robotics.
- Data moat: on-device shadow-mode training lets each deployed robot collect real-world data continuously, compounding the model's advantage.
- Speed as strategy: 45 days to a first body and five rounds in six months turned momentum into a competitive barrier.
- Capital signals: Tencent and Alibaba as strategic investors de-risk enterprise and consumer channels.
What can be applied
A proven execution team compresses years of credibility into months: investors priced Li Auto's smart-driving record, funding five rounds before the product shipped at scale.
Aftermath
As of March 2026, Simplexity had announced five rounds totaling ¥2 billion and a post-money valuation above $1 billion, making it the youngest embodied-intelligence unicorn. It had completed two generations of robot bodies (B2B and consumer), begun small-batch production and PoC validation, opened Beijing, Shanghai and Suzhou operations, and was pouring the new capital into foundation-model training, body iteration, data collection and algorithm R&D. It plans to expand from factories to manufacturing and services and from China to overseas, following a closed-to-open roadmap.
Sources
spotted an error? The archive wants to know.
Your turn
You just read one. Describe what you are building, and see who is betting on the same thing.
Free account · 3 free questions · no card