AI interprets factory orders, Cloudpi Intelligent Computing launches the factory AI Brain.
Bringing AI from the chat box to the workshop is what Beijing Yunpai Intelligent Computing Technology Co., Ltd. (hereinafter referred to as "Yunpai Intelligent Computing") is doing. The company recently launched the "Factory AI Brain" product for small and medium-sized manufacturing enterprises — Yunpai Intelligent Engine Factory Brain. It connects the enterprise's existing ERP, MES, WMS and other systems through the locally deployed AI execution layer, and hands over the entire business chain from order acceptance, material matching, production scheduling to delivery and settlement to AI for management, with deployment and launch completed in as fast as 3 to 7 days.
Orders of small and medium-sized factories are scattered in Excel sheets and WeChat groups
After visiting a large number of small and medium-sized manufacturing enterprises, the Yunpai Intelligent Computing team found that the most urgent problem for these factories is often not replacing workers with robotic arms, but the operational problems that occur every day. When a customer order arrives, relevant materials are scattered in Excel sheets, ERP systems, WeChat groups and employees' computers, and no one can quickly find all of them; the versions of BOM lists, engineering drawings and SOP process documents are not unified, and the wrong drawings are not discovered until production starts. A more typical scenario is the scheduling link: the PMC planning of the whole factory relies on one experienced veteran technician, and if he takes a day off, the production arrangement of the whole factory will be completely messed up. There is clearly inventory backlog in the warehouse, but key materials are often missing on the production line; procurement staff are busy inquiring for prices, urging delivery and tracking delivery dates every day. These seemingly trivial problems are consuming the delivery capacity, inventory capital and profits of enterprises every day.
Yunpai Intelligent Computing locks the target users of Factory Brain on trade-manufacturing integrated enterprises and small and medium-sized manufacturing enterprises. This group has a huge number, with uneven informatization foundations. Most of them have already used ERP, but the data in the system still requires manual query, judgment and operation. Over the past two decades, ERP has helped enterprises record businesses, but the data has always been "lying" in the system, and its value has not been truly released.
Not an ERP substitute, but an AI execution layer on top of enterprise systems
The product idea of Factory Brain is an "AI execution layer", rather than another set of management system that requires enterprises to demolish and rebuild. It is like a layer of neural network, connecting the enterprise's existing ERP, MES, WMS, e-commerce platforms, financial systems, as well as Excel and document materials accumulated over many years. In terms of specific business processes, when a customer order enters the system, AI will automatically identify the order content and match the corresponding products, BOM and engineering drawings; then check the existing inventory and in-transit materials to judge which are sufficient and which need to be purchased immediately; procurement tasks are automatically generated and supplier delivery dates are continuously tracked, and the risk of delay will be warned in advance. In the scheduling link, AI integrates order priority, material arrival time, production process and equipment capacity to assist in generating the optimal production plan, and finally links the delivery logistics and automatically aggregates cost and revenue to complete the closed loop of financial accounting. The system can also output native AI CAD drawings, covering the digital connection from drawings to processes.
Supporting this execution layer is the self-developed software-hardware integrated technology stack of Yunpai Intelligent Computing. At the top level is the Maple multi-agent orchestration architecture, which does not preset fixed processes, but organizes AI into a self-adjustable "team": agents of various roles are assigned to the Agent pool by department for call at any time, a tree-shaped reporting structure is dynamically formed when tasks come, new members are "recruited" on site if there is a shortage, adjustments require approval, and each step is traceable and rollbackable. This architecture has completed 316 automated tests, is open source under the Apache 2.0 license and has zero third-party dependencies. At the bottom level is the self-developed full liquid cooling hardware topology and dual-system heterogeneous intelligent computing system. The entire software-hardware integrated system is deployed on the enterprise's internal AI server, so core data will not leave the factory area, and it can run 7×24 hours.
In terms of business model, Yunpai Intelligent Computing adopts a three-layer structure of "software-hardware integrated delivery + software subscription + continuous service": the Factory AI Brain software system and local AI server are delivered in an integrated manner, application modules such as AI scheduling and AI drawing are charged on an annual subscription basis, and supporting deployment, implementation, operation and maintenance iteration services are provided.
Starting from understanding products, and gradually moving towards equipment collaboration
According to the technical path planned by Yunpai Intelligent Computing, Factory Brain is promoted in three stages: the first stage is to let AI understand products and business processes, and open up the whole link from order to delivery, which has been realized; the second stage is to access MES and equipment layer data, so that AI can deeply understand the production process and on-site status; the third stage is to let AI participate in equipment collaboration and production automation on the premise of safety and controllability. The company clearly stated that it will not pursue "unmanned operation" for the sake of concepts, but gradually increase the degree of automation starting from the most practical operational problems of enterprises.
In terms of the team, the company is the AI Intelligent Computing (Beijing) Center of Shanghai Zhu Guangya Strategic Science and Technology Research Institute. Chief Scientist Wang Yuanpeng is responsible for the technical route design of Factory AI Brain, and the core team has a compound background in computing power hardware and industrial intelligence. In September this year, Factory Brain made its debut at the 16th Smart City and Intelligent Economy Expo. Wang Yuanpeng delivered a keynote speech titled "Let Every Small and Medium-sized Manufacturing Enterprise Have Its Own Factory AI Brain"; the company's self-developed AGC architecture micro-super node Agent all-in-one machine previously won two honors at the 2026 Global Digital Economy Conference: "Excellent Case of Internet of Things and Smart City" and "Innovation Achievement of Industrial Intelligent Computing Application Scenario".
Contact: Teacher Gao