36Kr (36氪)→ original

Шэсюй (设序科技) закрыла раунд B — всего 300 млн юаней на промышленный AI для проектирования

Китайская компания Шэсюй закрыла раунд B на сумму более 100 млн юаней (около $14 млн) и довела общий объём инвестиций до 300 млн юаней. Её платформа «Цзэсин AI» на базе промышленной world-модели генерирует 3D-модели и 2D-чертежи для инженеров. За первое полугодие 2026 года контракты выросли на 70%, среди клиентов — Honda, Dongfeng и CRRC. В 2026-м компания выходит в Европу, начав с Германии.

AI-processed from 36Kr (36氪); edited by Hamidun News
Шэсюй (设序科技) закрыла раунд B — всего 300 млн юаней на промышленный AI для проектирования
Source: 36Kr (36氪). Collage: Hamidun News.
◐ Listen to article

Chinese startup Shexu (设序科技) closed a Series B round worth more than 100 million yuan (about $14 million) in July 2026, bringing its total funding raised to 300 million yuan (~$42 million). According to 36Kr, the funds will go toward European expansion and the development of an industrial AI model.

What the Zexing AI platform can do

Zexing AI (则形AI) is a cloud-based agent platform for hardware engineering: from an engineer's text prompt, it generates a three-dimensional part model and two-dimensional drawings. At its core is a proprietary industrial "world model" linked to a large language model. Shexu was founded in 2020, and its product is already used in automotive, electronics (3C), and energy, as well as in commercial spaceflight and robotics.

  • Series B — more than 100 million yuan; over 300 million yuan raised to date
  • Investors: Shenzhen Capital (深产投), Heding (合鼎共), and existing shareholder Yonghua Investment (涌铧投资)
  • Founded in 2020; CEO — Wu Yongrong, PhD from Tongji University in advanced manufacturing
  • Clients: Honda, Dongfeng, Yutong, CRRC, China Automotive Engineering
  • Contracts in the first half of 2026 grew by roughly 70% year over year

How AI designs a part

The platform fulfills an order in three steps, each driven by a separate type of AI. Geometric AI (3D Agent) builds the precise shape of the part, physical AI calculates its strength, stiffness, flows, and magnetic fields, and manufacturing AI (2D Agent) assesses manufacturability. The physical AI that links them is still at the POC validation stage.

"We want to create a universal AI entry point for hardware engineering.

The user states the requirements, and AI carries them out in three steps — first geometric AI designs the precise shape, then physical AI works out its strength, stiffness, flow, or magnetic field, and finally manufacturing AI assesses manufacturability," — Wu Yongrong, founder and CEO of Shexu, in an interview with 36Kr.

Will AI replace engineers?

AI won't fully replace engineers, but it will redistribute the work, the company's founder believes. In his view, the ideal scenario is for AI to handle 80% of routine operations while the engineer focuses on the 20% that involves creative decisions and handling non-standard cases. Wu Yongrong expects the "GPT moment" for industrial world models to arrive within 3–5 years at the latest — when AI reaches L4–L5 autonomy, as in self-driving.

"Before AGI arrives, humans won't be replaced, but an engineer who can't work with AI will be replaced by an engineer who can," —

Wu Yongrong, CEO of Shexu.

Why the company is expanding into Europe

Shexu has declared 2026 the year of international expansion with a focus on Europe. The company has already signed its first, "seed," client in Germany; in the first half of 2026 it completed preparations for compliance with EU data requirements. In the second half of the year, Shexu expects to onboard about 10 clients, and in 2027 it plans to scale up in Europe while also preparing the markets of Japan, Korea, and Southeast Asia. A separate growth driver is the RAAS (Result as a Service) delivery model, which accounts for roughly a third of annual revenue; for the full year the company forecasts about 200 million yuan, nearly double last year's figure.

What this means

Industrial "world models" are moving out of labs and into commerce: generative AI is starting to design real parts for the automotive and energy industries, not just text and images. Shexu's round shows that investors are willing to put money into narrow engineering verticals of AI, not just universal chatbots.

Frequently asked questions

How much has Shexu raised?

The Series B round totaled more than 100 million yuan, and the company has raised more than 300 million yuan in total. Investors include Shenzhen Capital, Heding, and existing shareholder Yonghua Investment.

What is Zexing AI?

Zexing AI (则形AI) is a cloud-based agent platform that generates 3D part models and 2D drawings for engineers from a text description. It is built on Shexu's proprietary industrial world model and is already used in automotive, electronics, and energy.

Who are Shexu's clients?

Clients include Honda, Dongfeng, Yutong, CRRC, and China Automotive Engineering. In the first half of 2026, contract volume grew by roughly 70% year over year.

⧉ Story
ZK
Hamidun News
AI news without noise. Daily editorial selection from 50+ sources. A product by Zhemal Khamidun, Head of AI at Alpina Digital.

Need AI working inside your business — not just in your newsfeed?

I build production AI for companies — custom CRM, internal tools, autonomous agents, workflow automation. Owned by you, shaped to your process, no per-seat tax. Built by Zhemal Khamidun, CPO of AlpinaGPT (AI platform, 6,000+ users).

What do you think?
Loading comments…