Cadence Adds AuraStack AI Agent For PCB And Advanced Packaging Design
The Register reported that Cadence Design Systems introduced AuraStack, an agentic AI system for PCB and advanced packaging workflows. Cadence cited a 15x productivity claim and named Nvidia among customers, but The Register did not include pricing, availability dates, full customer names or independent benchmark results.
Cadence Design Systems has introduced AuraStack, an agentic AI system for printed circuit board and advanced packaging design.
It uses natural-language instructions to plan and orchestrate circuit design and testing workflows that still run through Cadence's existing simulation tools.
Unveiled on Wednesday, AuraStack targets work that has historically depended on high-precision simulations.
The product serves as an interface for coordinating design and test suites rather than replacing those engineering tools.
AuraStack Orchestrates PCB And Advanced Packaging Workflows
For electrical engineers, AuraStack is intended to support printed circuit board design, testing, and advanced packaging design.
The system can connect with a range of open and proprietary AI models while the underlying workflows run on CPUs, GPUs, and other accelerators.
Michael Jackson, corporate vice president of Cadence's system design and analysis division, stated that AI is "amplifying the value" of the company's engineering products and technologies.
He outlined a reliability workflow in which the system identifies power-management components, creates a simulation-ready power tree, runs the simulation, and gives feedback to the designer.
The Register described AuraStack as an engineering tool rather than a general AI assistant.
It works with Cadence's existing product stack, which already automates parts of circuit design and simulation.
Cadence Cites A 15x Productivity Claim
Jackson indicated that a PCB or package design can require thousands of tasks through development, with "Sixty-five percent" of an engineer's day spent navigating and dealing with those tasks.
Cadence claims AuraStack can deliver a 15x productivity boost by moving more of that orchestration work away from designers.
That performance claim remains vendor-owned; independent benchmark methodology, customer-level productivity data, or a third-party engineering validation report were not included.
Among the companies signed up for the service are several large electronics groups, including Nvidia.
The report did not list the full customer roster, contract terms, or deployment scope for those users.
Nvidia Is Named As A Customer, But Deployment Detail Is Limited
Beyond this launch, Cadence has also built similar AI agents for digital and analogue chip design.
AuraStack fits a broader pattern of using lower-precision AI models to coordinate more precise single- and double-precision physics simulations.
Separate earlier work at the Department of Energy's Sandia National Laboratories involved researchers using AI agents to develop and test hypotheses in a self-driving lab setting.
Those tests used mature architectures such as variational auto-encoders rather than large language models.
Cadence did not provide independent benchmark results for its 15x productivity claim.
