AI Is Coming, How Will EDA Evolve?

The semiconductor industry is undergoing profound changes brought about by AI. On March 19, during his speech at GTC 2025, NVIDIA CEO Jensen Huang pointed out that AI has gone through three technological paradigm shifts, from Perception AI to Generative AI, and then to the current Agentic AI. The next phase will be Physical AI, marking the era of robotics.

AI is becoming increasingly intelligent and widely applied, requiring more computational support. Embracing AI has become a consensus among all semiconductor companies. As the "origin of birth" for almost all semiconductor products, how will EDA (Electronic Design Automation) tool software face the continuous and vigorous industry transformation, and how will it evolve?

Is It Time to "DeepSeek"?

Earlier this year, after the surge in popularity of DeepSeek, the semiconductor industry began an "all-hands-on-deck" mode. Between late February and early March, within two weeks, almost all CPU and computational chip companies announced their compatibility with DeepSeek. In contrast, EDA compatibility with DeepSeek is still relatively rare.

On February 11, Hangzhou GrandMicro Electronics Co., Ltd. (hereinafter referred to as "GrandMicro") announced that its SemiMind platform is integrated with DeepSeek, achieving three functionalities: integrating industry know-how and massive process data to build a domain knowledge base; supporting users in quickly building customized functional modules through low-code/no-code methods; and upgrading data analysis software platforms to provide personalized recommendations, automated process management, and real-time data analysis.

On February 28, the domestic FPGA chip design EDA software eLinx announced its integration with DeepSeek, supporting rapid generation of FPGA functional modules, code syntax and logic error detection, and performance improvement.

On March 18, domestic industrial software company Daisy Software, which possesses CAD/CAE/EDA/HPC/RVC self-development capabilities, announced a deep technical integration with the large model DeepSeek and Tongyi Qianwen. This will support optimization of simulation parameter design through complex data analysis and predictive models, reducing trial and error costs. For example, in automotive crash simulations, AI can automatically recommend material combination schemes, shortening the verification cycle. Additionally, this cooperation will support interdisciplinary simulations, automatically identify conflicts in multi-disciplinary simulations (such as structural strength and thermodynamic contradictions), and provide optimization suggestions.

Looking at overseas EDA manufacturers, news about their compatibility with general large models on the market is quite rare.

Many companies have chosen to develop native AI application routes, and many functions have already been used in production. In an interview with "China Electronics News," Beijing Huada Jiutian Technology Co., Ltd. Chairman Liu Weiping stated that some tools have already started using AI technologies, optimizing engineers' efficiency significantly. Some tasks that used to take 10 hours now can be completed in just 1 hour.

"Acceleration" and "efficiency improvement" are answers the reporter often heard when asking about the impact of AI on EDA business. However, industry professionals still seem to doubt higher-level, more intelligent EDA AI applications.

There are three reasons for this. First, AI training requires a large amount of industry data, which needs long-term and massive data accumulation. Without such accumulation, the model's reliability cannot be discussed. Second, the EDA industry requires high precision, while the basic operating logic of models is inference, which makes precision difficult to achieve. Third, the chip industry's trial-and-error costs are extremely high. Without confidence in AI reliability, no manufacturer would risk using it at the expense of declining yield rates.

Different Paths for Domestic and Foreign Companies

In the EDA industry, domestic and foreign companies are advancing along two different paths.

Overseas EDA manufacturers have formed a relatively stable oligopolistic pattern. The three major EDA leading companies—Synopsys, Cadence, and Siemens EDA (formerly Mentor Graphics)—have completed the layout of full-process EDA tool solutions, occupying over 70% of the global market. Some research institutions predict that after the three companies complete their respective mergers and acquisitions, the overall market scale may reach over 90%. For these companies, aside from continuously expanding the scope of acquired companies and broadening business areas, they also focus more on helping customers adapt to the rapidly changing market environment through various means.

At the Synopsys Developer Conference, the company outlined three major business development focuses. First is to supplement the IP core product line, providing customers with a comprehensive IP core portfolio. The second is EDA platform innovation, using reinforcement learning and other methods to upgrade EDA tools. The third is launching packaging solutions.
In contrast, domestic leading EDA companies generally have high-performance products that have achieved significant market recognition, but no single company can achieve a complete toolchain layout. Domestic companies tend to see product line expansion as an important development strategy: expanding product lines in their areas of expertise or striving to develop full-process solutions.

Based on this, industry representatives have given three suggestions on promoting the sustained and healthy development of China's EDA industry. First, focus on R&D. The core vitality of the EDA industry lies in continuous technological innovation and breakthroughs. As a typical high R&D investment industry, some EDA companies' R&D expenditures account for more than 70% of total costs. Second, strengthen intellectual property protection. Chip design companies should actively reject pirated software to create a good market foundation for the commercial application of EDA software. Third, focus on the ecosystem. Resist homogenized vicious competition, insist on developing more competitive products through continuous R&D investment, and form a virtuous cycle.

With the deep application of AI technologies like DeepSeek and the intelligent upgrading of the semiconductor industry, global semiconductor manufacturing is accelerating towards high efficiency, high precision, and low carbonization. This transformation not only reshapes the technological landscape of the semiconductor industry but also transmits new demands and challenges to upstream raw material manufacturers of copper processing, copper rod, copper strip, copper wire, enameled wire, and magnetic steel through the industry chain. For copper processing companies, the trends of semiconductor equipment miniaturization and high chip integration will drive their transition to high-end products such as ultra-fine copper wire, high-purity copper strip, and high-temperature resistant enameled wire. Magnetic steel manufacturers will need to improve the magnetic performance and stability of materials to meet the stringent requirements of core components in semiconductor equipment, such as precision motors and sensors.

Jintian Copper, leveraging its leading technology and rich experience in the copper processing field, actively responds to market changes and demands, dedicating itself to providing high-performance, high-precision copper products for the global semiconductor industry. As a key industry supplier, Jintian Copper focuses on the production and processing of copper rod, copper strip, copper wire, enameled wire, and magnetic steel, continuously promoting product technological innovation and upgrades.

Meanwhile, AI-driven intelligent production models (such as process optimization and defect detection) also provide cost reduction and efficiency enhancement pathways for traditional material companies. Facing the dual waves of technological iteration and market demand, companies need to center on technological innovation, conducting intelligent production line transformations, collaborative R&D within the industry chain, and green low-carbon process upgrades to build future-oriented competitiveness. It is foreseeable that the deep integration of the semiconductor and materials industries will foster a comprehensive innovation from technical standards to business models, ultimately propelling China's high-end manufacturing towards the top of the global value chain.