Industrial Next was founded in 2021, with its global headquarters in Suzhou Park and its team in San Francisco, USA. In 2017-2019, the team completely delivered the Tesla Model3 production capacity improvement and production line upgrade solution, applied driverless technology to manufacturing, reduced manufacturing costs and BOM costs, and shortened product production cycle.
This issue of "Seeing New Projects" interviewed Chen Yiming, COO of Industrial Next, to learn how Industrial Next empowers traditional workstations and turns workstations into "marginal brain nodes".
Industrial Next completed the Pre-A round of financing of about US$ 12 million in June last year. The main investors are Lenovo Venture Capital and Xiaomi Group, and it is expected to start the A round of financing in the third quarter of 2023.
Open up R&D and production with the new generation production system of smart factory as the core.
At present, intelligent factories are being popularized in various industries, among which new energy vehicles, 3C and semiconductor industries are the industries with the highest penetration rate of intelligent factories, and they are also the industries mainly targeted by Industrial Next.
For these target customers, Industrial Next thinks that they will put forward three new requirements in the future: the new iteration of products will be accelerated, the production capacity will expand flexibly with market demand, and the higher gross profit margin level will support continuous innovation. However, in traditional manufacturing, R&D and manufacturing are separated, and manufacturers adopt the method of developing first, and then manufacturing after the product is finalized, which makes the communication between R&D and production not smooth, and the production line cannot accept the continuous iteration of the product, resulting in the inability to adjust the production capacity once the product is finalized, and the gross profit of the product cannot be improved. Such a production mode will no longer match the current intelligent development.
In view of the new needs of future customers, Industrial Next thinks that it is urgent to connect R&D and production, and it is necessary to connect R&D and manufacturing more efficiently and more agile, so that production can drive R&D and jointly reduce the cost of BOM (accounting raw material cost according to standard cost).
Chen Yiming said: "At present, traditional factories mostly adopt rigid production lines, R&D, production and sales are independent of each other, and they can only be used for large-scale orders and single-category production. Such a production line can only increase the gross profit margin by reducing the manufacturing cost, and cannot change the BOM cost. In our eyes, the future factory will be flexible and intelligent, which will realize global optimization, internal and external circulation linkage, and "R&D-production-sales" integration. Such a smart factory is intelligent throughout the factory and can realize small batch and multi-batch production and real-time optimization of products and costs. "
Chen Yiming also said: "In the future, we will make three improvements: the product R&D and production processes can be quickly and iteratively adjusted, and the low-cost rapid factory production and flexible production line can be realized, and the product parts and components will be continuously integrated and optimized, thus reducing the BOM cost."
According to this, Industrial Next determines a brand-new thinking mode. The key is to fuse parts with similar functions through modular technology and software-defined ability, thus reducing the number of parts and BOM cost. First, AI system analyzes the multi-dimensional parameters of 15,000 parts from CAD/CAE, generates Prefect Sample, and judges which parts can be fused into a new part according to the fusion degree; Second, the AI system combines production parameters and production technology to judge whether the new parts after fusion can be produced; Finally, AI system evaluates the ROI of new parts and judges whether the scheme reduces the total cost of BOM.
Combining the experience in Tesla and insight into the market, Industrial Next designed an innovative product -MaaS system (a new generation production system of smart factory), which includes three parts: system, software and hardware.
The system is an AI system that integrates R&D and manufacturing. Through system simulation, the parameters of parts are classified, and the system simulation is divided and fused to judge whether parts with similar parameters can be fused. Compensate production parameters, process parameters, production errors and other data to the simulation system, and comprehensively judge whether the fused parts can be produced; Through the cost calculation engine, the updated value of each part is calculated. The software can provide the core competence of the production scene through the integration of software and hardware of multi-sensor and low-cost production of universal AI model and automatic calibration driven by AI. The hardware sensing module is a self-developed multi-sensor fusion edge computing module, the learning module is a self-developed production scene high-performance SoC that can solidify the neural network, and the control module is a self-developed extensible modular control platform. Through MaaS system, we can get through R&D and manufacturing, and help enterprises improve their competitiveness from three aspects: reducing parts, simplifying technology and reducing manufacturing costs.
The production mode of software-defined hardware will become the direction of intelligent production.
For manufacturers with existing production lines, Industrial Next provides hardware and SaaS to improve production efficiency. For example, upgrade EV, 3C and other manufacturing stations to provide real-time quality inspection and dynamic negative feedback adjustment capabilities; Merge similar items, reduce the number of production line stations, reduce costs and improve delivery; Understand and analyze the system through AI, and solve 90% manufacturing problems that affect the productivity in OEM. At the same time, Industrial Next optimizes R&D and manufacturing from the architecture, improves the efficiency of workstations in five dimensions: personnel, machines, materials, methods and environment, and feeds back to upstream parts suppliers and downstream production lines in a data format that can be seen in manufacturing and R&D to assist in product design adjustment.
For companies without production lines, Industrial Next has the delivery capability from 0 to 1, providing a highly flexible, configurable, modular and software-driven production line, including software services, system platforms and hardware devices in MaaS system.
Industrial Next introduces autonomous driving technology into the production line, and projects the "perception-decision-execution" of autonomous driving to the production line to realize the integrated cycle iteration of "perception-decision-automation", and the adaptive ability is spiraled. Collect data and monitor the status of the station in real time; Analyze the data to find the optimization space of products and production lines; Respond to the execution command immediately and optimize the product design and production line scheme. Under the trend of intelligent transformation of manufacturing industry, Industrial Next plans to expand this scheme with MaaS system to industries with "complex product composition and fast product iteration speed" to promote the construction of a new paradigm of the next generation manufacturing industry.
For the future factory, Chen Yiming also mentioned: "The production mode of software-defined hardware will become the core of intelligent production, and the AI control platform will become the central brain of the future factory." Industrial Next will focus on the broader next-generation manufacturing industry, hoping to improve and upgrade the production line technology that has been verified in Tesla and promote its application with the help of Tesla’s genes to help realize "software-defined manufacturing".
This article was originally produced by Alibaba Cloud Innovation Center. If you need to reprint it, please note the source.