AI-powered semiconductor design and manufacturing: Siemens

AI-powered semiconductor design and manufacturing was presented by Dr. Kolja Zakrezewski, Siemens. 

He stated that 29 of the top 30 global semiconductor companies are working with Siemens. India now wants to go even bigger. About 24% of  <20nm chips are required for automotive. We can do silicon-enabled, software-defined, and AI-powered chips to build the factory of the future. Siemens has been an industrial electronics player for very long. We are designing chips that are powering industrial electronics.

Our industrial electronics are also AI-powered. Chips are the foundation of all those. We need to combine digital and physical worlds. We need R&D and industry collaboration, chip design and verification, equipment and materials supply, and chip manufacturing, assembly, packaging, and testing. We need design, ecosystem, and production. We also need chemicals supplier, for eg., Merck. imec is defining roadmap for the years ahead for the semiconductor industry. 

Design of semiconductors is becoming increasingly complex. Decisions need to be taken at the very beginning of the stage. Sustainability and circularity decisions also need to be taken at the beginning. EDA reaches the next level through multifunctional AI. EDA digital twin consists of software, functional IC, physical IC, 3D IC, and electronic systems. 

Software has architecture exploration and IP selection. Functional IC has IC functional design and verification. Physical IC has IC physical design and verification. 3D IC has IC packaging, 3D IC, and chiplet integration. Electronic systems has electronic systems design, verification, and manufacturing.

You can simulate faster for highest productivity in chip design. Design verification with Questa One saw 10-1000x reduction in manual tests. MediaTek said GenAI can save weeks of engineering, and days of regression and debugging. Physical verification with Caliber Vision AI saw 2x productivity boost. It notifies design violations, and organizes errors through clustering.

About 60% of companies use six or more systems to manage their data. TeamCenter can deploy, use, extend cloud, or get on-prem PLMs. We can get accelerated new product introduction.

Lot of semiconductor parts are being buil across the world. We need to transform how we can produce much better semiconductors. You can stimulate and optimize without touching your production. Ensure optimal production setup at any time. Understand correlation in production better. Engineering is now powered by co-pilot. We can harmonize, contextualize, and orchestrate shop floor to top-floor data. We can optimize resource usage and increase asset lifetimes powered by AI.

For the ecosystem, we need to consider designing new chips and electronics systems. We need to bring that to the factory. The industry needs to increase collaboration. We are enabling the industry to develop silicon-enabled, software-defined, and AI-powered ecosystem. We can jointly unlock the opportunity. Technology can help us transform everything, everyday.