There may be no better industry event for semiconductor design than DAC, or the Design and Automation Conference. Now in its 62nd year (yes, sixty-two, that’s no typo), this is the signature event for semiconductor design. And everyone who is anybody designing silicon chips is here: Cadence, Synopsis, Siemens, and more.
Siemens, as we all know, acquired Mentor Graphics a few years ago (for $4.5 billion) and renamed it Siemens EDA. Then they added a number of other products that may have sealed the deal as a leadership position.
Siemens was all decked out at DAC 2025. They had bought the platinum sponsorship and were clearly positioning themselves as EDA (electronic design automation) leaders.
At a media event featuring key executives from the Siemens EDA division, we learn about how Siemens will be able to create semiconductors like never before, leveraging its extensive EDA portfolio and a commitment to implementing industrial AI to give chip designers every possible advantage.
Siemens, more than any other design and engineering software company, appears to be intent on implementing AI for maximum effect. Where others may be timid and tentative, Siemens is charging ahead, taking the lead and going boldly forward. And I must say, the result could indeed be a design revolution.
DAC Attack
At DAC, Siemens introduced a comprehensive AI system designed to revolutionize semiconductor and electronic design automation (EDA). The announcement, with one AI-enhancement after another, could be a giant leap forward in how chips are conceptualized, designed, and verified. It made me think that Siemens, the normally staid German manufacturing conglomerate, could actually leapfrog longstanding, specialized EDA players, even those who exist for no other purpose than to create EDA.
At the heart of Siemens’ approach is the concept of “industrial-grade AI,” which for the EDA division has resulted in a multi-pronged set of tools. Amit Gupta, VP and GM , Custom IC Verification, who joined Siemens through the Solido acquisition, emphasized the critical difference. While consumer AI can afford to be approximate, industrial AI must be precise, verifiable, and immediately applicable to mission-critical applications.
Amit founded Solido in 2005, when it was a pioneer in machine learning in electronic design. Their breakthrough was developing a method to generate billions of simulation coverages using only a few thousand simulations. This approach allowed semiconductor designers to overcome manufacturing variability challenges by creating more efficient and reliable chip designs.
Comprehensive AI-Powered Design Ecosystem
Siemens’ new EDA AI system spans an impressive range of design tools, including:
- IC and PCB design tools
- Simulation tools
- Product lifecycle management solutions
- Digital industry software
The system integrates AI across multiple stages of chip design, from initial concept to final verification, offering unprecedented capabilities in:
- 3D chip design
- Thermal performance prediction
- Stress analysis
- Collaborative design
- Error detection and resolution
Chips Playing 3D Chess
Chip design used to be 2D, with all transistors arranged on a single level. We now go to another dimension with 3D chip design, also known as 3D IC (integrated circuit) or heterogeneous integration. This involves stacking multiple layers of silicon with transistors or chiplets arranged in layers and with circuits that go X, Y and Z directions. Come to think about it, why not go 3D? It can only perform better with reduced power consumption and save space. This architecture shortens interconnect distances between components like processors, memory, and accelerators, enabling faster data transfer and higher efficiency, essential for AI and high-performance computing.
What’s a Chiplet?
A chiplet is a small, modular integrated circuit (IC) that forms part of a larger system-on-chip (SoC) or multi-chip module (MCM). Rather than creating a single, complex chip with all features on a single piece of silicon, manufacturers can build several smaller chiplets and assemble them together in a package.
The advantages are that you can mix and match chiplets (CPU cores, GPUs, AI accelerators, IO, memory controllers) to customize chips for different uses. Chips are cheaper to produce and test. Technologies like 2.5D/3D stacking and silicon interposers enable high-speed connections between chiplets. Rather than going crosstown, signals can go up and down, travelling smaller distances.
Examples of chiplets include AMD Ryzen CPUs, which use chiplets: CPU cores on one chiplet, IO functions on another.
Also, NVIDIA and Intel are also adopting chiplet designs for GPUs and CPUs.
EE Times, the leading publication in the electronics space, devoted a stage to discuss them.
Calibre
Siemens is a key enabler of 3D chip design through its Calibre and Xcelerator portfolios. Calibre provides industry-leading electronic design rule checking and 3D-aware physical verification, essential for ensuring manufacturability across stacked dies. Siemens’ IC packaging tools, especially those from its chiplets product lines, support chiplet integration, thermal analysis, and testing strategies for advanced multi-die systems. The company also offers system-level modeling and digital twin simulations, allowing designers to evaluate performance, power, and reliability early in the design process. This comprehensive toolchain is crucial for next-generation chip development.
Transformative Tools Across the Design Workflow
Wei Lii Tan, Director of Product Management, Digital Design Creation Platform, Aprisa, demonstrated how Aprisa uses AI to transform the RTL (Register Transfer Language) to GDS (Graphic Design System) workflow. The AI-powered tool delivers remarkable improvements:
- 10x productivity increase
- More computationally efficient design processes
- Better power, performance, and area (PPA) metrics
The AI Design Explorer uses reinforcement learning to automatically optimize design flows, eliminating the need for manual expert tuning. This approach can improve design metrics by up to 10%, allowing engineers to focus on higher-level design challenges.
Calibre Vision AI: Intelligent Verification
Priyank Jain, Principal Product Manager, Calibre Interfaces, Calibre Design Solutions, introduced Calibre Vision AI, a tool that revolutionizes chip debugging and verification. Key features include:
- Ability to load and analyze billions of errors in under a minute
- Intelligent error clustering
- Automated root cause identification
- Collaborative debugging capabilities
In one example, the tool reduced error analysis from 5,000 individual checks to just 200 grouped clusters, dramatically accelerating the verification process.
Custom IC Design: AI-Powered Precision
The custom IC design tools now incorporate AI assistance for:
- Setup guidance
- Analysis support
- Debugging help
- Results generation
Addressing Industry Challenges
The Siemens EDA AI system addresses critical semiconductor industry challenges including increasing chip complexity, reduced design cycle times and the need for more efficient design processes.
Unique Approach to Data Privacy
Recognizing the sensitive nature of semiconductor design, Siemens has developed a privacy-first approach. The system utilizes retrieval-augmented generation (RAG), enabling companies to leverage their proprietary data without exposing their intellectual property.
Nvidia Partnership and Ecosystem Support
Nvidia is supporting the Siemens EDA AI system, providing infrastructure and framework support, as it is with all major EDA companies. This partnership underscores the system’s potential to transform semiconductor design across the industry.
“Modern manufacturers face mounting pressure to boost efficiency, enhance quality and adapt swiftly to changing market demands,” said Jensen Huang, founder and CEO of NVIDIA, at VivaTech, France’s big tech event. “Our partnership with Siemens is bringing NVIDIA AI and accelerated computing to the world’s leading enterprises and opening new opportunities for the next wave of industrial AI.”
The Future of Semiconductor Design?
The Siemens team envisions a future of software-defined, AI-powered, silicon-enabled systems. This approach recognizes that software is increasingly driving hardware innovation, with AI acting as a critical enabler.
Industry estimates suggest AI could improve designer productivity by up to 50%. Siemens’ comprehensive approach positions them at the forefront of this technological transformation.
Siemens’ EDA AI system represents more than just a technological upgrade – it could be all the difference, a fundamental reimagining of semiconductor design. By combining industrial-grade AI with a comprehensive design ecosystem, Siemens EDA is setting new standards for innovation, efficiency, and precision in chip development.