Cadence Design Systems has launched AuraStack, an AI agent that combines low-precision artificial intelligence with high-performance computing to accelerate printed circuit board and advanced packaging design. The tool aims to speed up complex simulations by using AI to handle less precise tasks, freeing HPC resources for high-accuracy calculations.
How AuraStack works
AuraStack integrates AI directly into the design workflow, acting as an intelligent assistant that can quickly explore design options and predict outcomes. This allows engineers to iterate faster on PCB and advanced packaging layouts, which traditionally require lengthy simulation runs. The agent uses machine learning models trained on design data to provide near-instant recommendations, while the HPC backend handles the precise electromagnetic and thermal simulations needed for final verification.
Implications for chip design
By melding AI with HPC, Cadence offers a glimpse into how the industry can balance speed and accuracy. Low-precision AI can rapidly narrow down design choices, and high-precision simulations then validate the best candidates. This hybrid approach could reduce development cycles for complex chips and packages, which are increasingly critical for AI, data center, and mobile applications. The tool is part of a broader trend where EDA vendors are embedding AI into their software to handle growing design complexity.
Industry context
Cadence's announcement comes as chipmakers push for faster time-to-market and more sophisticated packaging technologies like chiplets and 3D stacking. AuraStack competes with similar AI-assisted design tools from Synopsys and Siemens EDA. The company has not disclosed pricing or availability details, but the agent is expected to be integrated into Cadence's existing Allegro and OrCAD platforms.