Cadence has been rolling out new “Super Agents” at an impressive pace to provide chip design teams with agentic AI support across the chip design workflow. The latest iteration will cut the time needed to design the Printed Circuit Boards (PCBs) and advanced multi-chip packaging in half, according to the company. Once again, Cadence named Nvidia as an early adopter. (Note that Cadence and Nvidia, like many players in the semiconductor industry, are both clients of my company, Cambrian-AI Research.)
With this announcement, Cadence claims to offer autonomous AI coverage across the full chip design spectrum. Built on Cadence Allegro AI Studio and accelerated by Nvidia, AuraStack offers a unified, AI-native environment. It aims to accelerate time to market by 2X, boost productivity by 15X, and improve quality through early multiphysics co-optimization, significantly reducing costly design iterations and respins.
What is a “Super Agent”?
A Cadence Super Agent is an agentic platform that helps translate a design team’s intent into working silicon using AI. Cadence has now rolled out four agentic AI suite that automate design and verification, analog chip design, digital chip implementation, and now PCB and Advanced Packaging design.
At the core of Cadence Super Agents are what the company calls “Mental Models”, knowledge graphs that aggregate existing design data, specifications, diagrams, libraries, and any other human-readable content that represents design intent. Mental Models serve as sources of ground truth on which agents can plan and direct their actions. These models then enable the stepwise planning, orchestration, and execution monitoring that engineers would have otherwise burdened the engineering team’s time.
While chip design is indeed complex, the boards and packages onto which chips are mounted are also design challenges. PCBs can warp, crack, or even melt under the intense thermals that chips can produce. And the advanced packaging technologies such as TSMC CoWoS (Chip-on-Wafer-on-Substrate) are the largest current supply chain bottleneck that plagues the AI semiconductor industry. Consequently, the multi-physics aspect of AuraStack is critical to a successful project.
The Bigger Picture
Applying AI agents to speed the chip design process is perhaps the most significant advance the EDA world has seen in many years. Cadence, as are its primary competitors, Synopsys and Siemens, is quickly developing and deploying agentic AI to their customers to speed time to market, improve productivity, and increase system-level quality. This is one of the reasons major chip companies like Nvidia and AMD have been able to move from a historic two-to-three-year design cycle to an annual product cadence, adding fuel to the raging AI fire.
The ChipStack AI Super Agent is available now for customer deployments. The Cadence ViraStack and InnoStack AI SuperAgents are in early engagements with development partners.
Disclosures: This article expresses the opinions of the author and is not to be taken as advice to purchase from or invest in the companies mentioned. My firm, Cambrian-AI Research, is fortunate to have many semiconductor firms as our clients, including Baya Systems BrainChip, Cadence, Cerebras Systems, D-Matrix, Flex, Groq, IBM, Infleqtion, Intel, Micron, NVIDIA, Qualcomm, SImA.ai, Synopsys, Taalas, Tenstorrent, Ventana Microsystems, and scores of investors. I have no investment positions in any of the companies mentioned in this article. For more information, please visit our website at https://cambrian-AI.com.