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

CATL Is the Sole Strategic Investor in an AI CAD Startup

The battery giant is the sole strategic investor in Ziqian Technology (子虔科技), a six-year-old platform that generates editable, manufacturable 3D models — a bet that the next productivity gain lies in the design office.

2026-11-19 · 1164 words · NeuroAI
CATL Is the Sole Strategic Investor in an AI CAD Startup

An engineer changes one measurement on a metal bracket. In a conventional design tool, a dozen related parts now need checking and adjusting by hand.

Ziqian Technology says its software can carry that change through the whole assembly and keep the result editable and simulatable.

On 17 November, the world's largest battery maker — by installed capacity, per SNE Research's 2025 ranking — decided it wanted a seat at that table, and it came alone.

CATL made an exclusive strategic investment in Ziqian Technology (子虔科技) that day, according to The Paper and other Chinese financial outlets. The amount was not disclosed. For a company that normally spreads its capital across many deals, going in as the sole strategic investor is a strong signal about where the battery giant thinks the next bottleneck sits.

Key takeaways

  • CATL was the sole strategic investor in a new round for Ziqian Technology, announced on 17 November; the round size was not disclosed.
  • Ziqian, founded in 2020, builds an AI platform for industrial research and development, including zCAD.AI and zCAD.Online, running on its own CAD kernel.
  • The company says its models produce editable, simulatable 3D assemblies rather than static pictures — the difference between a convincing demonstration and a part a factory can make.
  • CATL spent RMB 22.1 billion (≈ US$3.1B / HK$24.3B) on research and development in 2025 and employed about 23,000 R&D staff, according to reporting around the deal.
  • The investment treats the design office, not only the production line, as the next source of manufacturing productivity.

What Ziqian is actually selling

Most generative AI tools produce something that looks right. Industrial design has a harder standard: a model must obey dimensions, assembly relationships, material limits and manufacturing processes, and it must remain editable after an engineer changes one number.

Ziqian's answer is to generate parametric engineering designs that a person can keep working on. Its products sit on an in-house CAD kernel, and the company describes a workflow it calls "light-speed creation" (光速造物): a description in, a 3D assembly out, then import into zCAD.Online for editing. In one demonstration reported alongside the funding, a user described an aircraft-engine component and received a three-dimensional model that could then be revised.

The longer-term idea is to capture a company's accumulated design rules and engineering experience inside the platform. When an engineer adjusts a parameter, the system is meant to understand the knock-on effects across related parts, rather than leaving those to be found by trial and error.

Founder Gong Minyan (龚敏彦), an industrial-software entrepreneur, has been following machine-learning research for years. In interviews he has argued that the cost, supply-chain exposure and risk of a manufactured product are mostly fixed at the design stage. He has also said the toolchain requires physics, mathematics, materials and process knowledge woven together — not a single clever insight. That framing explains why the company sells to industry rather than to consumers.

Why a battery giant wrote the cheque

CATL is one of the most research-heavy manufacturers in the world. Its 2025 R&D spend of RMB 22.1 billion (≈ US$3.1B / HK$24.3B) and roughly 23,000 research staff were reported in coverage of the deal, and they explain the strategic logic.

A new battery is not just a cell. It pulls along production equipment, fixtures, structural parts and battery-pack casings, and those design tasks are spread across suppliers. When design work is slow, the whole chain waits. A manufacturer that already runs highly efficient production has a strong reason to care about the speed of the steps before production.

There is a second motive: control over tools. An exclusive position gives CATL influence over a design platform its engineering teams may come to depend on. It also fits a wider push in China to replace foreign industrial software with domestic alternatives — a shift with implications for manufacturers across Asia that still rely on Western design suites.

Ziqian says its customers span new-energy vehicles, aviation, intelligent equipment and power systems. One wearable-device client, the company says, cut its design-change cycle by about 40 percent and its rework cost at mass production by about a quarter after connecting design files and approval workflows. That case is a company claim, not an independently audited result.

The numbers, and what they do not prove

Two figures surfaced in the coverage. Ziqian said it served more than 100 enterprise customers in 2025 and more than 200,000 platform users across over 80 countries in 2026. It also said roughly a fifth of new sign-ups came through ChatGPT referrals.

These are self-reported growth numbers carried by an industry newsletter and other outlets, not verified statistics. They describe reach, not revenue or retention. A free or cheap tool can accumulate registered users without becoming part of anyone's production workflow. The more honest read is that Ziqian has distribution and a recognizable name in a niche, while its value to a manufacturer still has to be measured in rework avoided and cycles shortened.

The undisclosed round size is the other limit. Without it, there is no way to judge how much conviction CATL is expressing in money terms, or what valuation the deal implies.

Honest limitations

This article draws on The Paper (澎湃新闻), a Shanghai-based newspaper, plus Chinese financial-news reproductions and an industry funding newsletter. CATL and Ziqian have not published a joint filing, and I have found no Hong Kong or mainland exchange disclosure for this private round. The exclusive-investor framing and the 17 November date are consistent across the sources; the round size, user counts and customer case study are company claims.

I have not used zCAD.AI or zCAD.Online, so the description of the software comes from company material and reporting. Claims that the models produce manufacturable assemblies are unverified by any independent engineering test.

The claim that China is replacing foreign CAD software is a broad industry theme, not something this single investment proves. Nothing in the sources establishes that CATL will mandate the platform across its suppliers. Chinese company and product names are translated or transliterated roughly; I have kept the originals where the wording matters.

What readers can do now

  • If you work in product or mechanical design, test the editability claim, not the demo. Change one dimension in a generated assembly and see how many dependent parts update correctly. That is the difference between a toy and a tool.
  • If you supply a manufacturer, watch who owns the design platform. An exclusive strategic investor can make a tool free or cheap for partners — and can also make it the default standard you must design around.
  • Treat user and customer counts from startups as direction, not proof. Ask what percentage of registered users are paying and what share of designs reach production.

The more interesting part of this deal is not the money, which was not even revealed. It is the choice of stage. CATL has already optimized how things are made; now it is paying to make the thinking before that faster.

Related coverage

Cite this article

NeuroAI (脑机网), "CATL Is the Sole Strategic Investor in an AI CAD Startup," 19 November 2026, https://neuroai.site/a/na-capital-catl-ziqian-industrial-design

Machine-readable: BibTeX · RIS · JSON · Markdown

More in “Capital & Markets” → · Back to home · Markdown version

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