CATL and AgiBot co-invest in BluePoint, a leader in six-axis force sensors, to lock up core tactile technology and prepare for mass production.
As embodied intelligence races at full speed into the deep waters of mass production in mid-2026, the battle among major humanoid robot makers for core sovereignty has shifted decisively from high-level whole-unit launches and "brain" algorithm leaderboards down to the muddiest and hardest-fought terrain: end-effector component supply chains.
Bluedot Touch (Beijing) Technology Co., Ltd., the domestic leader in six-axis force/torque sensors, recently completed a highly targeted capital restructuring. According to the latest business registration changes captured in the Tianyancha App, Zhiyuan Innovation (Shanghai) Technology Co., Ltd., an affiliate of Zhiyuan Robotics, alongside the Fujian Times Zeyuan Equity Investment Fund, backed by CATL, have formally taken positions as new shareholders in the company's equity structure. At the same time, the company's registered capital edged up from approximately RMB 5.003 million to RMB 5.255 million.
This seemingly uneventful micro capital increase—adding just RMB 250,000 to the table—would barely register as a financial move worth a dedicated research note under conventional venture capital logic or in the secondary market. But when viewed against the backdrop of Bluedot Touch's just-announced Series C+ round of over RMB 100 million completed this April, not to mention its staggering market share of over 80% in six-axis force/torque sensors for humanoid robots, that RMB 250,000 "technology premium" reveals its true nature: a joint "ecosystem protection fee" paid by the two most powerful forces in the embodied intelligence track to the "robot tactile brain" on the eve of full-scale mass production.
Many technology analysts accustomed to tracking the whole-robot market and inflating valuations for large-model unicorns tend to interpret this move as routine early-stage investor positioning, or a big company's lightweight test of locking in one more ordinary supplier in its supply chain. Such superficial and ultimately hollow commentary completely misses the hardware substitution bottleneck the entire industry is now confronting as robots make the leap from "seeing" to "doing."
In past industry explorations, a humanoid robot taking a couple of steps outdoors or grabbing an apple in a PowerPoint demo could get by with machine vision alignment and rigid mechanical arm programming. But when Zhiyuan Robotics attempts to push its Expedition series onto heavy industrial assembly lines, and when CATL tries to deploy embodied intelligence to take over its extremely hazardous and tedious power battery cell assembly lines, the hallucination-prone nature of vision systems and the blunt insensitivity of rigid mechanical arms become fatal vulnerabilities.
The robot's ability to flexibly adjust its wrist posture, grip with its fingertips, and even perceive micro-level forces during assembly demands end-effector sensors capable of exceptional performance under tight spatial constraints.
Bluedot Touch—a company built by Liu Wuyue, a leading expert in aerospace-grade precision sensing technology who sold his house to fund the startup—derives its deepest value-chain positioning from mastering the purest domestic core algorithm platform for six-axis force/torque sensors and joint torque sensors, with precision rivaling the entrenched overseas monopolists. The fact that Zhiyuan Robotics and CATL have now aimed their investment dollars squarely at this high-tech enterprise founded in 2019 marks, from a technology inflection standpoint, a hard-nosed transition from mere "customer buying product" to "sharing strategic sovereignty."
CATL wants the closed-loop control capability of absolutely safe flexible production lines and long-tail battery assembly; Zhiyuan Robotics, meanwhile, needs to ensure its end-effector supply chain cannot be poached by any competitor as the first mass-production year arrives.
That RMB 250,000 in newly added share capital visible through Tianyancha represents, in the capital marathon crowded with actuaries, a vastly amplified "craft vault." It does not carry the burden of heavy industrial land acquisition, because Bluedot Touch's existing capacity and million-scale automated upgrades were already underwritten by its earlier hundreds of millions in funding. The latest footprint this money leaves in Tianyancha is the legal stitching together of two giants' interests with Bluedot Touch's process patents and multi-axis synchronous calibration technology.
In this 2026 end-game defined by real delivery efficiency and the purity of hardcore componentry, whole-robot teams that pile on flashy visual effects while failing to build hard walls around end-effector micro-actuators and precision tactile sensing will ultimately pay the steepest exit bill under the weight of industrial-landing compliance. The two investment coordinates Zhiyuan and CATL have left in the Tianyancha system are a sober statement of how giants are completing their productivity iteration in the physical world. When the bubble of speculators profiting off AI narratives and lacking underlying hardware cash-flow capability is finally punctured, the ones who can set aside ego first—and achieve pixel-level meshing between their whole-robot ecosystem and the most hard-won sensor sovereignty—will be the ones who genuinely tighten the safety line that secures long-term corporate recovery in the coming shakeout.
