AI / LLMChina
← Sticker Wall
JOURNEY · DETAIL
Taishi Navigation: The Autonomous Driving Veteran Entering Embodied AI with Wire Harness Assembly
Founded: Yilun Chen, Tongqing Chen · Taishi Navigation
JOURNEY
Key Fields
FIELD STAMPSIndustryAI / LLM
RegionChina
ScaleMid-size
ChannelOther
Origin
The founding team comes from autonomous driving companies such as Huawei, having experienced the entire process of intelligent driving from demos to mass production. In 2025, recognizing the explosion window for embodied AI, they decided to migrate the perception, decision-making, simulation, and supply chain experience accumulated in the autonomous driving field to robots. They chose to cut in from automotive wire harness assembly—a high-frequency, essential, and high-precision segment—hoping to bypass the generalization difficulties of general-purpose humanoid robots and prioritize feasible commercial scenarios.
Milestones
2025
Founded Turning Point
In 2025, autonomous driving veterans including Yilun Chen and Tongqing Chen founded Taishi Navigation in Shanghai, with core team members largely originating from mass production projects like Huawei Intelligent Driving. In the early entrepreneurial stage, the team attempted to directly reuse autonomous driving perception algorithms, but discovered that robot body control, gripper mechanisms, and simulation pipelines were completely different. They were forced to pivot to a self-developed hardware-software integration route, marking a critical node in the company's transformation from an autonomous driving team to an embodied AI company.
2025
Early Financing Growth
Less than a year after its establishment, the company successively completed two rounds of financing totaling nearly $700 million, breaking early-stage financing records in the embodied AI field. Investors include Sequoia China, Hillhouse, Meituan, Qiming Venture Partners, and others. Capital continued to pour in despite high valuations, driven fundamentally by the recognition of the 'dimensional reduction strike' narrative of an autonomous driving team pivoting to robot wire harness assembly.
2026
Pre-A Round Growth
In April 2026, Taishi Navigation announced the completion of a Pre-A round financing of $455 million, with a valuation exceeding 20 billion RMB, led by Sequoia China, Hillhouse, and Meituan, with follow-on investments from Qiming Venture Partners and others. This is one of the largest single Pre-A round transactions globally in the embodied AI sector. The funds will be used for the mass production of wire harness assembly robots and the construction of a model data closed-loop, signaling the company's transition from the laboratory to factory validation.
2026
Technical Route Selection Turning Point
In 2026, while mainstream robot companies heavily bet on VLA large models, Taishi Navigation publicly stated that the VLA route was more like an outsider's approach, insisting on self-developing embodied models based on autonomous driving end-to-end frameworks. The team believes that data structures in automotive scenarios and robot manipulation tasks share deep homology, and operation data should be extended on top of driving data rather than directly copying language-action models. This judgment sparked intense debate across the industry.
2026
Scenario Validation Transition
In 2026, alongside automotive wire harness assembly, Taishi Navigation began experimenting with applying robots to flexible and precision manipulation tasks such as embroidery to validate the generalization capabilities of its technical solution. Public reports indicate the company is seeking to prove that its end-to-end embodied model can replace human labor in high-precision, unstructured scenarios, expanding its commercialization path from a single scenario to multiple scenarios and providing more landing options for subsequent mass production.
Turning Points
- Crossing over from autonomous driving to robotics in 2025 and utilizing the wire harness assembly scenario to open up the market entry point became the company's most critical strategic pivot.
- Securing $455 million in Pre-A round financing in April 2026 with a valuation exceeding 20 billion RMB propelled the company from an early-stage project into a top-tier embodied AI player.
- Publicly questioning the mainstream VLA route and insisting on migrating the autonomous driving end-to-end framework to robots established its distinct technical tag amidst industry controversy.
Failures & Pitfalls
- Directly reusing autonomous driving perception algorithms on the robot body during the early entrepreneurial phase resulted in wasted development time upon discovering that truss structures and gripper force control differed completely from automotive control.
- The yield rate of the wire harness assembly scenario initially struggled to meet the requirements of automotive production lines, with clients only willing to use robots for auxiliary workstations. The team was forced to return to actual production lines to recapture data and redesign actuators.
- To keep pace with the financing rhythm, a generational gap emerged between product demos and mass production versions, leading to a passive situation of budget overruns in supply chain costs under 2026 mass production pressures.
关键成功要素
- The team possesses mass production experience in autonomous driving, enabling the migration of perception, decision-making, and simulation modules to robotic scenarios.
- Taking essential demand segments like automotive wire harness assembly as the entry point to bypass general-purpose robot generalization difficulties.
- Securing joint lead investments from Sequoia China, Hillhouse, Meituan, and others in the Pre-A round, establishing strong capital endorsement.
- Insisting on self-developing end-to-end embodied models without blindly following mainstream VLA routes, thereby building a technical barrier.
Lessons
- Entrepreneurship in a new track requires leveraging experience migration from the original industry, but must confront the differences between robot hardware and autonomous driving.
- High financing heat does not equate to commercial maturity; validation in B-end scenarios like wire harness assembly still takes time.
- Maintaining independent judgment on mainstream technical routes helps companies build differentiation amidst bubbles.
- Capital can accelerate R&D, but mass production yield and cost remain the long-term variables determining victory or defeat.
Core Data
- 一年融资总额：Nearly $700 million (Based on public disclosures; independent verification unverified)
- 第一轮前融资额：$455 million (Based on public disclosures; independent verification unverified)
- 估值：Over 20 billion RMB (Based on public disclosures; independent verification unverified)
- 成立至第一轮前的时间：Approximately 1 year (Based on public disclosures; independent verification unverified)
- 领投机构数量：3 (Based on public disclosures; independent verification unverified)
Competitors / Peers
Key competitors include embodied AI startups such as Unitree Robotics, Agibot, and Galbot, as well as overseas players like Tesla Optimus. Unlike competitors that cut in from quadrupedal robot dogs or complete general-purpose humanoid robot systems, Taishi Navigation relies on its autonomous driving team background, choosing precision operation scenarios like automotive wire harness assembly and embroidery as its landing entry points to build differentiated barriers using end-to-end embodied models. However, such scenarios have long validation cycles and stringent customer acceptance standards, lagging behind some competitors in humanoid robot hardware and commercialization scale.