NeuroAI NEUROAINEUROAI.SITE
China AI, reported clearly
1,300+ verified stories on models, robotics, BCI and the business of AI.
Browse stories
ESC
Home · en

One production line, 36,000 robots' worth of force sensors

Daily AI/BCI industry briefing (EN edition, 2026-11-18).

2026-11-18 · 3787 words · NeuroAI
One production line, 36,000 robots' worth of force sensors

On November 17, Bluepoint Touch announced a Series D3 round in the hundreds of millions of yuan, with 14 institutions participating.

The company makes six-axis force sensors — the sense of touch in a robot's wrist and joints.

15% of cost, 80% of the market

According to Securities Times, force sensors account for about 15% of a humanoid robot's bill of materials. One robot uses four six-axis sensors and 28 joint torque sensors.

A CCID report puts Bluepoint's share of China's humanoid six-axis force sensor market at 80.3% in 2025.

Together the two numbers describe the position: the part is cheap, but without it the robot cannot move; and the market is concentrated enough that latecomers struggle to enter.

The sensors carry three hard specifications. The company cites 0.1% full-scale accuracy, a 10kHz response rate, and 500% overload tolerance. The first two decide whether it can hold an egg; the third decides whether it survives a knock on the factory floor.

Dividing capacity by content per robot

The company has built a facility of more than 10,000 square metres in Jiangmen, Guangdong, running what it describes as China's first fully automated force sensor line, with annual capacity of 1 million joint sensors and 200,000 end-effector six-axis units.

At 28 joint sensors and 4 six-axis sensors per robot: 1 million divided by 28 is about 35,700 units; 200,000 divided by 4 is 50,000. Taking the smaller, the line at full utilisation corresponds to roughly 36,000 humanoid robots.

That conversion is mine, based on per-unit content. It is not an order book. The company has not disclosed orders in hand or actual shipments.

Delivery time is the moat

Another number matters more: overseas suppliers quote 8 to 12 weeks for custom delivery; Bluepoint says it delivers in 2 to 3 weeks.

Founder Liu Wuyue told Securities Times that China's supply chain responds faster, and customisation can be compressed to a month.

For a humanoid maker, that difference decides whether a volume milestone is met. A sensor two weeks late means a robot two weeks late, and customers do not wait.

The customer list makes the point: AgiBot, Galbot, Xiaomi Robotics, UBTech, EngineAI and Noetix are all buying in volume, with several accounts at the 10,000-unit level.

A billion in half a year, buying position

The company has raised more than 1 billion yuan over the past six months. Its shareholder register includes CATL, SAIC and GAC, along with AgiBot and Galbot, and SMIC Capital.

Industrial capital, local state funds and financial investors rarely sit together like this. Local funds bring regional orders and factory landing; industrial backers bring joint development; financial investors bring a listing path.

Liu's thesis is that robot makers will fragment while component suppliers consolidate, much as Bosch sits beside carmakers.

Whether that holds depends on two things: whether capacity can ramp with orders, and whether overseas customers accept the price. He says the next step is competing directly with established suppliers such as ATI; at home the products have already been through a price war, and abroad the contest is delivery and cost.

The bet underneath is timing: when humanoid robots move from thousands of units to hundreds of thousands. On that day, capacity becomes share.

The arithmetic on a robot

Thirty-two sensors per robot is a lot of sockets. It also explains why sensor makers scale faster than robot makers: every humanoid shipped pulls through a multiple of components, and the same sensors go into industrial arms and collaborative robots that already ship in volume.

That is the reason a component supplier can raise at this pace while the humanoid market is still measured in thousands of units. The sensor business does not depend on humanoids winning; it depends on them not losing.

The risk is symmetrical. If humanoid designs converge on fewer sensors, or on cheaper non-six-axis alternatives for non-critical joints, the content per robot falls. The company's answer — moving from hardware to sensor plus algorithm plus data — is an attempt to hold value per unit even if unit content drops.

The order of magnitude is worth holding onto. A component that costs a small slice of the bill of materials can still gate an entire production plan, because assembly cannot finish without it. Scarcity, not price, is what gives the supplier leverage.

*Disclaimer: This article is based on publicly available information and is for reference only. It does not constitute investment advice.*

Related coverage

Cite this article

NeuroAI (脑机网), "One production line, 36,000 robots' worth of force sensors," 18 November 2026, https://neuroai.site/a/06-na-capital-force-sensor-1118

Machine-readable: BibTeX · RIS · JSON · Markdown

More in “en” → · Back to home · Markdown version

China AI, reported clearly
1,300+ verified stories on models, robotics, BCI and the business of AI.
Browse stories