AI4Engineering innovative startup "Force Engine" has raised nearly 50 million RMB in its Angel+ round of financing.
Recently, Force Engine (Shanghai) Intelligent Technology Co., Ltd. (hereinafter referred to as "Force Engine"), a leading enterprise in the AI for Engineering (AI4Engineering) track, announced the completion of a nearly 50 million yuan Angel+ round of financing. This round of financing is jointly led by Shanghai Future Industry Fund under the Shanghai State-owned Investment Group and Shanghai Science and Technology Innovation Group, with Fudan University Science and Technology Innovation as an additional lead investor, Longson Venture Capital, Chengwei Capital and other institutions participating in the follow-on investment, and Lighthouse Capital acting as the exclusive financial advisor. The funds raised in this round will be mainly used for continuous R&D of generative simulation and intelligent control engines, device-level verification of benchmark scenarios such as nuclear fusion, expansion of industry clients in the energy and high-end manufacturing sectors, and continuous strengthening of the construction of the world's top interdisciplinary talent team.
Force Engine was founded by Professor Wu Tailin, who received his bachelor's degree from the School of Physics of Peking University, his doctorate in physics from the Massachusetts Institute of Technology, and later engaged in postdoctoral research in the research group of Professor Jure Leskovec at the Department of Computer Science of Stanford University; he is currently a distinguished researcher and assistant professor at the School of Engineering of Westlake University, and the head of the Westlake University Artificial Intelligence and Scientific Simulation Discovery Laboratory. He has long been engaged in research on large-scale scientific simulation, physical system control and scientific discovery. He once proposed with his doctoral supervisor Professor Max Tegmark a new learning paradigm of "theory learning" and a series of methods centered on AI Physicist, and has made a series of pioneering works in the field of generative AI empowering the simulation, control and design of complex physical systems.
In the broad concept of AI4S, AI4Engineering can be regarded as an important branch oriented to engineering physical systems and industrial implementation. It inherits the capabilities of AI4S in law learning and complex system modeling, and further extends AI to links such as simulation, state perception, design optimization and real-time control, and finally forms a closed-loop delivery that runs through the entire life cycle of engineering systems. Force Engine is committed to building an AI-native base covering simulation, control, diagnosis and design for high-dimensional, strongly coupled complex engineering physical systems. The company takes controlled nuclear fusion as its first highly difficult verification scenario, and expands its relevant capabilities to energy and high-end manufacturing fields such as oil and gas, nuclear power, 3C and aerospace. In the above scenarios, the model must not only "calculate fast", but also meet the engineering constraints such as high precision, generalization, real-time performance, safety boundary and on-site deployment.
From Numerical Solution to AI-Native Simulation and Control
Traditional numerical solvers and CAE software are important foundations for engineering R&D, but for multi-physics, multi-scale, multi-component coupled systems, it is usually necessary to build a solution process for specific geometry, boundary conditions and physical processes. When it enters the inner loop of parameter scanning, inversion design and control strategy training, the calculation and engineering costs will be further amplified.
Force Engine does not simply replace traditional solvers, but takes numerical solution and experimental data as a highly credible training basis, learns the evolution and conditional distribution of physical fields through methods such as neural operators and diffusion models, and completes fast prediction and coupled generation in the inference stage. The company has achieved tens to hundreds of times acceleration in complex physical simulation tasks such as fluid, oil reservoir and plasma. Some tasks can be compressed from the traditional hour or day-level calculation to second-level, and the model can be continuously calibrated through real devices and industrial site verification.
Verify the Technical Upper Limit with Nuclear Fusion, and Migrate to Complex Engineering Scenarios
Magnetic confinement fusion involves algorithm challenges such as high-fidelity simulation, state estimation and closed-loop control of complex plasma systems under extreme conditions. It spans the scales of macroscopic magnetofluid and microscopic turbulence, and involves the coupling of multiple complex physical processes in the core-boundary, which is one of the most representative highly difficult scenarios for AI4Engineering. In recent years, the academic community has verified the magnetic control of deep reinforcement learning for multiple plasma configurations on the TCV device, and realized active avoidance of tearing instability on the DIII-D device, indicating that AI is gradually entering the closed loop of actual fusion experiments.
Force Engine is promoting cooperation in plasma simulation, soft landing control and configuration control with domestic fusion scientific research institutions and leading commercial fusion enterprises in the fusion field. The generative AI control algorithm developed by the company has completed the experimental verification on the device and achieved superior control performance. The company hopes to use the strict requirements of fusion scenarios for model accuracy, real-time performance and safety to reversely temper reusable physical modeling and control capabilities.
This migration will reuse underlying methods such as neural operators, generative modeling, safety control and multi-agent collaboration, and then adapt to the geometric structure, boundary conditions, working condition data and safety specifications of various industries. The relevant algorithms developed by the founding team have been industrially deployed in the 10 million grid oil reservoir simulation of Saudi Aramco. The company will also promote industrial cooperation in directions such as nuclear power multi-physics simulation, 3C thermo-mechanical simulation, aerospace and other fields in the future.
Force Engine is Committed to Making AI Enter the Engineering Closed Loop and Reshape the R&D Paradigm of Complex Systems
The policy and industrial environment have also formed a joint force in the AI4Engineering field. The State Council's "AI+" action proposes to promote the integrated collaboration of AI-driven technology R&D, engineering realization and product implementation. The Outline of the 15th Five-Year Plan lists nuclear fusion energy as one of the future industrial directions for key cultivation. With the accelerated intelligent upgrading of major scientific devices and advanced manufacturing systems, the AI4Engineering capability that can enter the real engineering closed loop and measure value by system-level performance improvement is expected to become an important infrastructure for AI to move from the digital world to the physical world, reshaping the simulation, design, diagnosis and control paradigm of complex engineering systems.
From plasma control in fusion devices to multi-physics simulation in oil reservoirs, nuclear power and high-end manufacturing, Force Engine is targeting a kind of common fundamental engineering problem: how to make AI understand, predict and intervene in complex systems faster and more accurately on the premise of following physical laws and safety boundaries.
After the completion of this round of financing, the company will continue to polish the technical upper limit with highly difficult scenarios such as controlled nuclear fusion, promote the migration of simulation, control, diagnosis and design capabilities to more fields including energy and high-end manufacturing, and gradually precipitate into a reusable, deployable and continuously iterative AI4Engineering infrastructure, promoting engineering R&D from relying on high-cost calculation and empirical trial and error to a new paradigm driven by the collaboration of data, models and real systems.
Shanghai Future Industry Fund stated: "Shanghai Future Industry Fund will continue to carry out systematic layout around the entire industrial chain of fusion energy 'device - device - material - AI4Fusion - application' to support breakthroughs in key technologies and core links. The Fund attaches great importance to the deep integration of AI technology and the fusion industry, strives to promote the construction of 'fusion AI-native' capabilities, actively pays attention to and promotes the landing of high-quality and scarce AI for Fusion enterprises in Shanghai, empowers the intelligent upgrading and industrialization process of fusion devices in Shanghai, and helps Shanghai build a globally influential future energy innovation ecosystem."
Shanghai Science and Technology Innovation Group stated: "Shanghai Science and Technology Innovation Group fully affirms the cutting-edge breakthroughs of Professor Wu Tailin's team of Force Engine in the field of AI for Engineering. Starting from the difficult problem of intelligent control of controlled nuclear fusion plasma, the company's AI-native engine with both physical mechanism and data advantages greatly improves the simulation efficiency, and has extended to scenarios such as oil reservoirs and aerospace. As a long-term investor, we are optimistic about the global competitiveness of China's hard technology, will integrate industrial resources, support Force Engine to strengthen core technologies, promote the early implementation of nuclear fusion and industrial simulation applications, and help create a new pattern of AI4E in the global science and technology competition."
Fudan University Science and Technology Innovation stated: "We are optimistic about the technical accumulation and industrialization potential of Force Engine in the direction of AI4Engineering. The company verifies the technical boundary with highly difficult scenarios such as controlled nuclear fusion and aerospace, and builds AI-native capabilities covering simulation, control, diagnosis and design, which is expected to promote the R&D of complex engineering systems from being driven by experience and high-cost calculation to the collaboration of data, models and real systems. We look forward to working with the company to accelerate its implementation in energy and high-end manufacturing and other fields."
Longson Venture Capital stated: "As the first AI for Engineering company in China with generative AI as the base, the company takes the lead in cutting into the most difficult problem in the engineering field of controlled nuclear fusion — nuclear fusion involves high-dimensional strong coupling control of trillion-level variables. At present, the company has achieved millisecond-level control response to plasma instability in the simulation environment, and has reached in-depth cooperation with leading players of nuclear fusion devices. We are optimistic about its technical barriers and implementation capabilities. The technology polished by high-dimensional complex engineering scenarios such as nuclear fusion has extremely strong dimensionality reduction and reuse capabilities, and can be extended to broad markets such as nuclear power, aerospace, and high-end manufacturing. We look forward to the company continuously iterating the physical large model and engineer Agent, building the core base of domestic AI engineering simulation, and helping China's high-end industry and energy technology achieve breakthroughs."
Chengwei Capital stated: "Chengwei Capital has long been deeply involved in the cross-cutting field of cutting-edge technologies. We judge that the next breakthrough point of AI4Engineering is not to replace traditional CAE software at a single point, but to build an AI-native base covering the full link of 'simulation - control - diagnosis'. The Force Engine team has both top-level physical academic background and hard-core engineering implementation capabilities. It chooses the ultimate scenario of multi-physics strong coupling of controlled nuclear fusion as the touchstone, and takes the lead in completing device-level verification. This path of 'defining the technical upper limit with the most difficult scenarios' is highly consistent with our investment philosophy of 'less but better, heavy investment, long-term companionship'. Chengwei Capital looks forward to accompanying Force Engine to push AI4Engineering from the laboratory to the real industrial closed loop, reshape the R&D paradigm of complex engineering systems, and then witness AI taking human goals as the anchor point to provide a brand-new evolutionary path for understanding and safely intervening in complex physical systems."
Lighthouse Capital stated: "AI4S is moving from accelerating scientific discovery to modeling and intervention of complex engineering systems, and AI4Engineering is an important direction connecting cutting-edge algorithms and industrial value. Force Engine takes generative simulation and intelligent control as the core, cuts in from the highly difficult scenario of controlled nuclear fusion, and the team has the capabilities of AI research, physical understanding and engineering implementation. Lighthouse Capital is long-term optimistic about the platform opportunities brought by the deep evolution of AI4E into engineering, and looks forward to accompanying Force Engine to grow into a general base for AI to move to the physical world."