Agentic AI engineering is emerging for use in Electronic Design Automation (EDA), a process already already used in building edge applications requiring multiple sensors like robots, drones and autonomous vehicles. However, industry pros believe EDA today is too slow to handle designs for the new and coming AI accelerators and 3D Integrated Circuits.
On Sunday, Siemens announced it is taking the concept of agentic engineering to a higher level, by combining its existing EDA tools with Nvidia AI infrastructure and software. The result is higher productivity and design quality across the EDA lifecycle, Siemens declared, including for semiconductors and printed circuit boards (PCBs) used in electronics, including edge devices.
Siemens named STMicroelectronics as a testing and validation partner in its ongoing design activity in non-volatile memory work. “The ability to relate layout insight directly to electrical behavior is critical,” said Gianbattista Lo Giudice, design manager at STMicroeletronics in a statement.
 Rather than discovering layout dependent effects late in the process, Siemens’ new Solido Layout Analyzer, one part of the total Siemens enhancements, helps bring ST’s analysis earlier into the design flow, “improving confidence and helping cut down the time we spend debugging complex design blocks by weeks,” he said.
Siemens recently introduced Fuse EDA AI Agents which are now combined with Nvidia AI tech. AI agents will reason and act and validate engineering decisions based on physics-based EDA engines, Siemens said. Fuse EDA is also integrated into Siemens’ Intelligence Center X for coordinating processes across design, manufacturing and supply chain working with Digital Twin.
Nvidia’s Nemotron open models will be employed in the process. EDA agents are built with Nvidia NeMo Gym, an open library. Nvidia’s Timothy Costa, vice president and general manager of computational engineering, claimed Nvidia accelerated computing and CUDA-X libraries will reduce signoff-quality results on semiconductors and PCBs to hours instead of days.
In a presentation, Siemens said its strategy with agentic engineering is to speed up design tools by 1000x, resulting in overall productivity improvements of up to 50x. Design risks will be reduced. Â
Why the news matters to the sensors universe
With regard to sensors used in robotics and other applications, EDA tools are almost always essential. Â If a robot company designs its own motor-control boards, sensor interface boards, power management systems, battery management systems, embedded controllers, customer AI accelerator boards and safety and communication electronics, then company engineers will use EDA tools to create schematics, simulate circuits, lay out PCBs and verify electrical designs before manufacturing. Siemens EDA competes with Cadence Altium and others.
“We can help customers accelerate development, improve design quality and increase confidence in trusted engineering systems,” said Amit Gupta, senior vice president and chief AI strategy officer for Siemens EDA. Siemens said its EDA will work with multiple products used in the design process; they include synthesis with Catapult and allow verification with Questa One and Veloce and custom IC design/verification with Solido, physical implementation with Aprisa, signoff with Calibre and design-for-test with Tessent. Also Siemens now supports advanced 3D IC integration using Innovator3D IC and PCB design with Xpedition.
Siemens asserted the growth in AI chip complexity, chiplets and 3D integrated circuits has outpaced traditional verification methods. GPUs and ASICs are now approaching 1 trillion transistors, and transistors in AI systems could reach a quadrillion by 2030, according to IDC.
“Agentic AI is the natural path forward. Agents can orchestrate multi-domain verification, reason across billions of test scenario and delivery production-quality results in timeframes that were impossible just two years ago,” said Abhi Kolpekwar, general manager of digital verification tech for Siemens EDA.Â
Siemens said its portfolio of expanded AI-driven EDA will become available in forthcoming releases of its AI-native EDA portfolio.