Industrial software is the cornerstone of the manufacturing industry and a pivotal sector. There are numerous types of industrial software, which can be broadly categorized into four major classes as shown in the figure below. Among them, embedded software is an extension of hardware devices and is not the focus of software investment. In the information management software segment, domestic vendors have gradually become mainstream players. Meanwhile, production control software is relatively easy to homogenize, leading to fierce competition, and thus is not our current focus.
From the perspective of investment and domestic substitution, R&D and design software possesses the highest technical barriers and entry thresholds, making it the most promising category in the next stage.
The scale of industrial software in each sub-segment might not be considered very large. Taking CAE software as an example, the current market size of China's CAE software is less than 5 billion RMB. However, it can support the R&D needs of tens of thousands of enterprises upstream, and further underpin an industrial output value of tens of trillions. In this article, we will focus on the investment opportunities in CAE, a specific sub-category of R&D and design software within industrial software.
In this article, you will learn: What is CAE? The current development status of domestic CAE; and the future opportunities for domestic CAE.
01. What is CAE? Reflecting the Physical World in the Digital Realm Replacing Physical Experiments in the Virtual World
CAE stands for Computer Aided Engineering. It uses virtual methods to reconstruct the entire engineering process of product design conception, drafting, simulation, and analytical testing, replacing real physical experiments with computer simulations. The main advantages of using CAE are the reduction of costly prototyping and testing, as well as the improvement of efficiency in the design process.
CAE is widely used in the R&D of various industrial products, and different CAE software emerges in every narrow niche of industrial products. For instance, a car generates noise due to vibrations during driving. Engineers need to analyze the noise issues of subsystems such as the engine, intake and exhaust systems, and transmission systems to reduce noise. In this case, software like Ansys Maxwell can be used for analysis.
Regarding the performance of aircraft under different flight conditions, wind tunnel testing, as a real physical flow field measurement method, can measure a vast amount of aerodynamic forces. However, wind tunnel testing is highly expensive; the total cost of a typical wind tunnel test and model processing reaches 5 to 10 million RMB. Benefiting from the development of CAE software, the number of wind tunnel tests required for aircraft R&D has been significantly reduced.
The wind tunnel testing time for the Boeing 787 was reduced by 30% compared to the Boeing 777 in 1990, and by 55% compared to the Boeing 767 in 1980. Boeing can use CAE to select a few design schemes with better performance and then put them into wind tunnels for test verification, which greatly reduces the cost of the final tests.
As the value of CAE is increasingly recognized by manufacturing enterprises worldwide, CAE software has become an indispensable part of the R&D and design process for these companies. Moreover, the growth rate of market demand is significantly faster than that of CAD software. It is expected that the market scale of CAE will eventually surpass that of CAD.
02. Current Development Status of Domestic CAE: Late Start and Low Market Presence
Although CAE plays a crucial role in modern manufacturing, the long-standing inherent mindset of valuing hardware over software in China has led to the neglect of industrial software development by enterprises and the government for a long time. The development of industrial software cannot be separated from the support of large manufacturing enterprises and the government. Even today, domestic CAE is only at the very initial stage in the market.
The leading domestic players are mainly the US-based Ansys, Germany's Siemens, and France's Dassault Systèmes. These three companies hold a combined market share of over 40% in China. Among them, Ansys is the absolute leader in the CAE field. Our industry research reveals that a vast number of universities and enterprises are using Ansys's standardized CAE products. In contrast, domestic CAE accounts for only about 10% of the overall market, and the vast majority of this revenue comes from software customization rather than product sales.
Over the past 20 years, Chinese enterprises have conquered numerous territories in the internet sector. Companies like Amazon, eBay, and Uber have retreated from China, while ByteDance's TikTok accounts for two-thirds of the monthly active users in the US market. Companies like Yonyou and Kingdee have also gradually established a firm foothold in financial and ERP software fields, competing on equal footing with SAP and Oracle.
China is the world's largest manufacturing country. Why is it difficult for Chinese enterprises to capture this specific CAE track? We believe the main obstacles to the development of domestic CAE are as follows:
-
- Technical Gap
- User Cognition
- Piracy Issues
- Customization-oriented Rather than Product-oriented Development
2.1 Technical Gap
CAE is generally divided into three modules: pre-processor, solver, and post-processor.
Pre-processor: Processes geometric meshes, sets business data, and prepares data required by the solver (meshes, properties, configurations).
Solver: Discretizes the physical model and sets parameters, then solves the system of equations through numerical algorithms, and finally generates analysis results.
Post-processor: Processes the calculation results from the solver and displays the corresponding results. Among these three modules, the most important is the solver, which uses program coding to objectively restore physical laws and mathematical principles.
Solver R&D involves three types of knowledge:
Theoretical Foundation: Includes various numerical theories, governing equations, and mathematical formulas, such as the classic heat conduction equation, wave equation, and Helmholtz equation. It requires the development team to be proficient in mathematics, physics, and industry knowledge.
Engineering Application: Theories and algorithms need to solve practical engineering problems in R&D. The characteristics of practical engineering problems are large and complex model scales, various parameters deviating from standard model assumptions, highly nonlinear features, and various unreasonable designs.
Software Development: Basic language development, requirements analysis, architecture design, process management, and technology selection.
Seeing this, you will realize that the underlying knowledge of a CAE solver is a complex comprehensive system, requiring cross-disciplinary comprehensive capabilities. The first type of knowledge, the theory, is relatively mature, and R&D has traces to follow, but it requires a certain amount of technical accumulation and has a considerable technical threshold.
The third type of knowledge has mature solutions, but the software industry itself is constantly and rapidly evolving. What is even harder is the second type of knowledge. Theoretical calculation and analysis actually involve many assumptions and premises, which do not completely match the actual models. For the selection of some calculation parameters, foreign CAE software has undergone detailed testing and user iterations. These parameters are encapsulated in the underlying layer as a black box. CAE requires a massive amount of engineering validation. Writing tens of thousands of lines of code to develop the first set of software is not difficult; the key lies in whether it can withstand the test of actual engineering problems.
Industrial software follows industry first and software second; its development cannot be separated from the feedback of the manufacturing industry. The development of the most top-tier industrial software is inseparable from years of engineering accumulation, with top leading manufacturing enterprises behind them. Only algorithms extracted from the most diverse and complex industrial environments can meet customer needs. The longer development history of foreign CAE has allowed them to accumulate leading engineering application experience.
History of Overseas CAE Development: The commercialization of CAE in the US originated in the aviation industry. In 1966, to meet the structural analysis needs of the aerospace industry, the National Aeronautics and Space Administration (NASA) proposed the Nastran program (NASA Structural Analysis Program).
The Nastran program was led by CSC, with MSC participating in the R&D. In 1969, NASA released the first version of Nastran, COSMIC Nastran. MSC, UAI, and CSAR all obtained licenses and continued development based on it. In 1971, MSC launched its own commercial product, MSC Nastran. The software quickly gained market recognition, thereby opening up this brand-new CAE track. MSC went public in 1982 and acquired UAI and CSAR in 1999, becoming the only company owning the commercial code of Nastran. In 2017, MSC was acquired by Hexagon, a well-known manufacturing enterprise. Today, the CAE giant Ansys also relied on the support of the manufacturing industry in its early development. Founded in 1970, Ansys's founder, John Swanson, worked at the Space Nuclear Propulsion Laboratory of Westinghouse Electric before starting his business. One of his tasks at the time was to conduct stress analysis for nuclear propulsion rockets.
In the early days of Ansys, all orders came from Westinghouse Electric. The nuclear industrial equipment business of Westinghouse Electric provided the best test samples and extensive application scenarios for the iterative upgrade of Ansys software. The V1 version of another well-known CAE software, Abaqus (acquired by Dassault Systèmes), was also specifically designed to serve a manufacturing customer—Westinghouse Electric, which used Abaqus to analyze nuclear fuel assemblies at that time.
Another important early customer for Abaqus was ExxonMobil. At present, a large number of giants and listed companies in the CAE field have emerged overseas. Besides comprehensive industrial software giants like Siemens, Ansys, Dassault Systèmes, Hexagon, Altair, and PTC have all achieved excellent performance in the capital market (with valuations generally around 10-15x PS, and PE multiples being even more exaggerated).
Recently, EDA giant Synopsys just announced its plan to acquire Ansys for $35 billion. The so-called EDA refers to the application of a series of software including CAD, CAE, and CAM in the specific field of semiconductors. Synopsys will benefit from the acquisition of Ansys at multiple levels, including underlying technology and customers. In China, the development of CAE started earliest in the academic community.
In the mid-1960s, scholars began to study CAE algorithms. In the 1970s, Dalian University of Technology developed the DDJ and JIGFEX finite element analysis software and the DDDU structural optimization software. Professor Li Mingrui from Beijing Agricultural University developed the FEM software, and the China Aircraft Strength Research Institute developed the aviation structure analysis software HAJIF. However, the CAE developed by universities was not for commercial purposes and lacked actual feedback and iteration from commercial customers. More importantly, from the 1970s to the 1990s, China's high-end manufacturing industry had not yet developed.
For a considerable period, the manufacturing industry leaned more towards the production process rather than process R&D and product R&D, which also led to the lack of engineering feedback from high-end manufacturing for domestic CAE. The US became an industrial power in the 1920s, with rapid industrial development. The Apollo program began in 1961. In 1964, Lockheed Martin developed the high-tech SR-71 Blackbird reconnaissance aircraft. In 1968, Boeing developed the subsequently popular Boeing 737. It can be said that a country's industrial level determines the level of its CAE.
CAE involves various fields such as structural simulation, fluid simulation, electromagnetic simulation, noise simulation, and other specialized simulations. Among them, structural simulation accounts for the largest proportion, followed by fluid simulation. Even leading manufacturers find it difficult to possess universal simulation products covering all physical fields. As mentioned earlier, well-known CAE software subsequently carried out numerous mergers and acquisitions through the power of capital, gradually completing the puzzle of various disciplines and establishing a complete product matrix.
2.2 User Cognition
In the mid-to-late 1990s, with the continuous and rapid development of domestic industry bringing increased simulation demands and opening up to the outside world, mainstream international simulation software vendors entered the Chinese market one after another, dealing a huge blow to domestic simulation software vendors. These foreign CAE tools occupied Chinese universities and research institutes of various enterprises, repeating the development path of other professional software (such as CAD, DCC, 3D engines, etc.). The technical advantages of foreign CAE allowed them to first capture Chinese universities and industrial users.
For engineers, the software they learn in school is Ansys, and the recruitment standards of enterprises after graduation also require proficiency in Ansys. Thus, foreign CAE has built a solid fortress from technology to user cognition. In many industries, using foreign CAE has become the industry standard. If the calculation results are wrong when using domestic CAE, who will take the responsibility?
2.3 Piracy Issues
As mentioned earlier, the R&D threshold for CAE solvers is high and requires significant investment. Without continuous financial support, enterprises find it difficult to sustain R&D. In the past, Chinese enterprises and individuals had low awareness of paying for software. In universities and state-owned enterprises, the procurement process for hardware was simpler than that for software. Even now, there is still a large amount of pirated CAE software in universities and state-owned enterprises.
For customers, if they can obtain industry-recognized CAE for free, why spend money to buy a product with slightly weaker functions? Foreign CAE software companies also turned a blind eye to piracy, using pirated versions to occupy the market in the early stages, and then starting to charge and harvest profits later when enterprises used more CAE software.
During our research on industry customers, we learned that a large electrical equipment manufacturing enterprise in Shanghai has been subject to Ansys rights protection actions since 2021. It has already been charged over 30 million RMB in copyright fees by Ansys, and this is still ongoing. Even a listed company finds it hard to bear and strongly hopes to find a cheap domestic alternative, only to find that the available options in the market are very limited. In the short term, it still has to continue paying high copyright fees to Ansys.
As overseas industrial software giants continuously intensify their copyright rights protection against domestic manufacturing enterprises, it is obviously unrealistic to rely on piracy for the continuous upgrading of domestic industrial enterprises in the future.
2.4 Customization-oriented Rather than Product-oriented Development
For software companies, product-oriented development is a difficult but correct path. Only product-oriented development can enable software companies to continuously invest the high gross profits obtained into R&D, and continuously enhance product barriers and core competitiveness from a medium to long-term perspective, forming a positive cycle. Although customization has certain advantages in scaling up and securing orders in the early stages, the continuous existence of customization in China—counting revenue by headcount, low gross margins, and the most severe accounts receivable problems—makes it difficult for such companies to maintain sustainable competitiveness in the medium to long term.
Foreign CAE companies are mostly led by standardized products, with gross margins generally achievable between 80% and 90%. In contrast, domestic companies, especially some CAE startups that have secured large amounts of financing, in order to quickly scale up and meet valuation expectations, can only go further and further down the narrow path of customization. This can be seen from a domestic listed CAE company, which has a gross margin of only 38%, and despite having just over 300 million RMB in revenue, it has over 500 million RMB in accounts receivable. It shows that the space and profits of such customization-oriented rather than product-oriented software companies will be very limited in the long run.
03. Future Opportunities for Domestic CAE
Although the development of CAE in China is arduous, we believe there is still huge room for development in the future.
Opportunity 1: Domestic Substitution
In August 2022, the US Department of Commerce issued a ban, cutting off the supply of EDA software necessary for designing chips with GAAFET (Gate-All-Around Field-Effect Transistor) structures to China. Although this ban only involves 3nm and below technology nodes, it has made many people in China more clearly realize the importance of independently developing core industrial software.
CAE is an indispensable software for the R&D of high-end industrial products. If the use of CAE is restricted one day, the R&D of everything from civilian automobiles to aerospace will be affected. The importance of vigorously developing domestic CAE software goes without saying. From our research results, the actions of state-owned enterprises and universities are already very clear, although there are no explicit internal documents requiring domestic substitution. Many state-owned enterprises have already launched domestic substitution projects internally. We believe this is a huge opportunity that the development of domestic CAE needs to seize.
Opportunity 2: Huge Future Market Increment
China's high-end manufacturing industry is developing rapidly. In 2023, the export value of high-end manufacturing reached 10.65 trillion RMB, accounting for 72.84% of the total annual export delivery value. Since 2020, the contribution of high-end manufacturing exports has remained above 70% for four consecutive years. As the number of high-end manufacturing enterprises continues to increase and enterprises increasingly recognize the value of CAE, we predict that the future CAE market space is huge, and this cake cannot be entirely taken by foreign CAE vendors.
From the perspective of the international market, any niche CAE can have customers. As long as the software has some unique features, there will always be customers willing to pay, and there will always be room for survival. China's manufacturing is moving towards both ends of the smile curve, which is a long-term trend. Continuous investment in R&D and design software is a necessary path. According to a survey by the Ministry of Industry and Information Technology, the current penetration rate of CAD products among Chinese manufacturing enterprises is estimated to be around 60%, while the penetration rate of CAE products is significantly lower than that of CAD. The future incremental market space remains huge.
Opportunity 3: More Cost-effective Products
The R&D of CAE products has extremely high technical barriers. At the current stage, domestic CAE vendors basically follow the path of customization for the military and state-owned central enterprises. Product-oriented CAE companies with their own solvers are extremely rare. However, once a breakthrough is made, it will form a huge product advantage. It will take a very long time and be extremely difficult for latecomers to make up for the core solver and standardized products.
Some high-end modules of foreign CAE have already been embargoed against China, and a basic set of CAE software costs millions of RMB, with hundreds of thousands of RMB required annually for upgrades and maintenance. For market-oriented customers, we have found that the exorbitant price of millions of RMB for foreign CAE products deters many small and medium-sized manufacturing enterprises. If there are similar products from domestic vendors that are relatively cheaper, many manufacturing enterprises we surveyed are willing to try them. I believe that the vast majority of domestic manufacturing enterprises have a dream of industrial upgrading to break away from low-end manufacturing.
Just like in the CAD field, domestic listed companies such as ZWSOFT and GstarCAD have emerged and gradually established a firm foothold in the CAD field. Although 3D CAD still awaits further major breakthroughs, this has already proven that the domestic design and R&D software industry possesses abundant opportunities at both the market and capital levels. We believe that domestic CAE can also gradually narrow the technical gap with foreign CAE, provide better cost-effectiveness and services, gradually occupy the domestic market, and even enter the overseas market someday through ultimate cost-effectiveness.
References
CFD Technology in Aerodynamic Design of Large Aircraft Wind Tunnel Test Requirements and Technical Prospects for Large Commercial Aircraft An Article to Get Started with "Solver" Development A Brief History of Industrial Software A Long Journey: The Independent Expedition of Simulation