Cadence Introduces AuraStack AI Super Agent, the World’s First Agentic AI Platform for PCB and Advanced Packaging

The majority of chips you are using today—from Intel, AMD, Nvidia, or Qualcomm—are designed with tools from Cadence or its main competitor. The toolset for chip design and validation that Cadence enables can significantly accelerate time to market and save serious resources.

Today, the company is ready to talk about AuraStack AI Super Agent, the World’s First Agentic AI Platform for PCB and Advanced Packaging.

Cadence’s CEO Anirudh Devgan is very close to Jensen, and the two of them have been seen together at the Cadence Live conference a few times in a row. It is clear that Cadence is on top of the AI game and has been exploring ways to use AI to accelerate chip design.

They already have an in-house foundation model that can help with validation and chip design, which was announced a few months ago. Now, the company wants to tackle PCB design and, most importantly, advanced packaging.

Advanced packaging, whether you talk about CoWoS or EMIB / Foveros—has helped enable many of the SoCs that played a crucial role in the AI revolution. Monolithic cores have been hitting their limitations for a while, and the only way forward was to put more of the dies on the same chip and connect them with advanced packaging.

Running on Allegro AI Studio, the AuraStack AI Super Agent delivers up to two times faster time to market and increases productivity by as much as 15 times. Currently, AuraStack AI Super Agent is the industry’s only multiphysics-driven quality silicon-to-system agentic AI stack.

AuraStack AI Super Agent on Cadence® Allegro AI Studio is the world’s first agentic AI platform for printed circuit board (PCB) and advanced packaging design, taking designers from system planning to final product in a single AI-native environment.

The Cadence AuraStack AI Super Agent is accelerated by NVIDIA Blackwell and NVIDIA CUDA-X, and is designed to coordinate domain-specific AI agents across planning, implementation, and tightly integrated multiphysics analysis domains to compress the system-design cycle through manufacturing.

I had a chance to chat with Tim Costa, VP and GM of Computational Engineering at Nvidia, and learned that CUDA-X is an umbrella term for its entire collection of domain-specific libraries, microservices, and software tools built on top of the base CUDA programming model. I also learned that Nvidia helps customers such as Cadence fine-tune their products and serve exact customer needs.

With the AuraStack AI Super Agent, Cadence is now the only provider with agentic AI solutions spanning the full electronic system design flow, from digital and analog silicon design and advanced packaging through to PCB design, building on its ChipStack™, InnoStack™, and ViraStack™ AI Super Agents.

The next era of AI infrastructure—spanning data centers, automotive, aerospace and physical AI—will be defined not only by silicon, but by the systems that connect, power and cool it,” said Michael Jackson, corporate vice president of R&D for System Design and Analysis at Cadence. “As hyperscale data centers deploy massive AI clusters and other industries advance increasingly intelligent, high-performance systems, engineering teams face growing complexity in PCB and advanced package design. Agentic AI orchestration, combined with trusted EDA and SDA tools, enables customers to move from manual iteration to intelligent, automated design realization.

Agentic AI for Packaging and PCB Design

Building on the same architecture as Cadence’s ChipStack AI Super Agent, agentic AI is combined with principled simulation and optimization tools, leveraging a mental model of the design intent, to automate and orchestrate design exploration, realization, and signoff.

The AuraStack AI Super Agent brings together automation and optimization for system planning, constraints management, physical structure definition, IP creation and reuse, place and route, design for manufacturability, and multiphysics analysis across Cadence’s system design and analysis portfolio. It introduces a unified, AI-driven multiphysics foundation that concurrently models and optimizes electrical, thermal, and mechanical behavior—including SI/PI, thermal, mechanical stress, drop, vibration, and fatigue analysis—within a closed-loop environment, enabling earlier tradeoff evaluation and product-level optimization across the entire design flow.

This continuous multiphysics feedback loop enables real-time design convergence, reducing late-stage surprises and improving overall system reliability.

The AuraStack AI Super Agent’s key benefits include:

• Accelerating time to market by 2X, with 15X productivity and multiphysics-driven quality—automating complex tasks, expanding design exploration.

• Unifying separate engineering teams around a shared, multiphysics-aware design environment with a single source of truth across domains.

• Advancing early and continuous multiphysics co-optimization to reduce late-stage rework and costly design iterations. Unifies Cadence multiphysics signoff solutions, such as Celsius™ Thermal Solver, Clarity™ 3D Solver for 3D-EM, MSC Nastran™ and Marc™ Linear and Non-Linear Finite Element Analysis Solvers for mechanical analysis, and Sigrity™ X Platform for signal and power integrity.

• Limiting expensive respins by identifying system issues earlier in development.

• Enabling product-level optimization, including co-optimization with advanced packaging approaches.

Industry Leaders Advancing Agentic AI with Cadence

Cadence is collaborating with industry leaders to deploy AuraStack AI Super Agent workflows for real-world advanced IC packaging and PCB design challenges.

NVIDIA is using Cadence to help automate and optimize increasingly complex system design workflows for its engineering teams.

“The scale and complexity of modern AI infrastructure demands a new design approach,” said Tim Costa, vice president and general manager of computational engineering at NVIDIA. “NVIDIA’s collaboration with Cadence is advancing AI-powered engineering workflows that accelerate design convergence and innovation across the industry. The Cadence AuraStack AI Super Agent and the Millennium M2000 Supercomputer deliver up to 20X faster multiphysics performance, giving our engineers the capability to tackle the most demanding design challenges and bring the next generation of AI infrastructure to life.”

Cadence is partnering with TSMC to help customers accelerate advanced package implementation through AI-driven automation, enabling timely design convergence for increasingly complex multi-die systems.

“As advanced packaging complexity grows, customers need new levels of automation to achieve timely design convergence,” said Aveek Sarkar, director of the Ecosystem and Alliance Management Division at TSMC. “Our long-standing partnership with Open Innovation Platform® (OIP) ecosystem partners like Cadence to deliver advanced package design and verification solutions for TSMC 3DFabric® technologies helps customers accelerate the realization of next-generation multi-die systems for AI and high-performance computing applications. Through our multi-year collaboration on substrate auto routing, we are already enabling customers to boost productivity by 100X while delivering quality of results similar to manual routing.”

Socionext, FORVIA HELLA, and Schneider Electric are also among the early adopters.

The Cadence AuraStack AI Super Agent will be available in 2026 and if you are in the business of making an PCB or an SoC with advanced packaging, without any doubt, this product has all it takes to enable making it to market faster.