In this week’s Fish Fry, I’m excited to welcome back Rob Knoth from Cadence Design Systems! Rob and I dive into a major shift in electronic design automation: the direct integration of artificial intelligence into the design workflow. Rob and I discuss the details of Cadence’s newest innovation – the ChipStack AI Super Agent and Cadence’s overarching strategies of ‘design for AI’ and ‘AI for design. We also explore the core challenges currently facing verification engineers, and examine how the ChipStack AI Super Agent is set to bring predictability and acceleration to your next design process.
Links for April 3, 2026
More information about Cadence Design Systems
More information about Cadence Design Systems ChipStack AI Super Agent
Click here to check out the Fish Fry Archive.
Click here to subscribe to Fish Fry via Podbean
Click here to get the Fish Fry RSS Feed
Click here to subscribe to Fish Fry via Apple Podcasts
Click here to subscribe to Fish Fry via Spotify
Amelia’s Weekly Fish Fry – Episode 675
Release Date: April 3, 2026
Amelia Dalton:
Hello everyone, and welcome to Episode 675 of Amelia’s Weekly Fish Fry, brought to you by EEJournal.com and written, produced, and hosted by yours truly, Amelia Dalton.
Did you miss me? Yes—last week, for a variety of reasons, this Fish Fry didn’t quite make it out of the pan. But this week, I’m back with the one and only Rob Knoth from Cadence Design Systems!
Rob and I are diving into a major shift in electronic design automation: the integration of artificial intelligence directly into the design flow. That’s right—we’re talking all about Cadence’s newest innovation, the ChipStack AI Super Agent.
We’ll explore Cadence’s “design for AI” and “AI for design” strategies, discuss the core challenges facing verification engineers today, and look at how this new solution aims to act as an orchestrated virtual silicon organization—bringing predictability and acceleration to the most risky parts of the design process.
So, without further ado—please welcome Rob to Fish Fry!
Interview Segment
Amelia Dalton:
Hi Rob, thank you so much for joining me!
Rob Knoth:
Hey Amelia, thanks for having me. It’s always great to be back on Fish Fry.
Design for AI vs. AI for Design
Amelia:
We’re talking about Cadence’s new ChipStack AI Super Agent today, but first—can you explain Cadence’s “design for AI” and “AI for design” strategies?
Rob:
Absolutely—it’s an exciting time, and this concept is core to who we are.
“Design for AI” refers to the work we’ve been doing for decades: partnering with industry leaders to create the chips, data centers, and devices that power the next generation of AI.
But the other side is “AI for design.” Once AI exists, it’s not just for writing poems or giving recommendations—it has a powerful role in engineering itself. At Cadence, we’re applying AI directly into our design tools to help engineers build the next generation of infrastructure.
It becomes a virtuous cycle:
You design for AI → which improves AI → which improves your tools → which helps design the next generation.
Challenges in Front-End Design & Verification
Amelia:
ChipStack AI Super Agent focuses on front-end silicon design and verification. What are the biggest challenges there today?
Rob:
Verification has become the long pole in the tent.
Historically, EDA tools focused on physical design challenges—placement, routing, optimization—especially during the FinFET era. But as that became more automated, designers could build more complex systems.
Now the challenge is verifying those systems actually work.
Verification is never fully “solved.” You aim for as close to 100% coverage as possible, but ultimately you still have to build and test the chip.
Another key point: verification involves a lot of natural language—from specifications to test creation. That makes it a great fit for AI, especially large language models.
But it’s also a domain that requires deep expertise—so combining AI with core EDA tools is critical to ensure accuracy and avoid hallucinations.
How the ChipStack AI Super Agent Works
Amelia:
Walk me through how this solution works.
Rob:
It starts with multimodal inputs—design documents, Verilog code, diagrams, even whiteboard images.
All of that is fed into the “super agent,” which acts as an orchestrator—not a single AI, but a system coordinating many specialized agents. Think of it as a virtual silicon organization.
From there, it builds a structured understanding of the design—what we call a “mental model.” This becomes the single source of truth.
That’s critical because AI is probabilistic. Without structure, it can hallucinate. With a mental model, it has a grounded, hierarchical understanding of the design.
Once that’s established, the system can:
- Generate RTL
- Create and debug tests
- Assist with verification strategies
This dramatically accelerates workflows and allows even non-experts—like architects—to participate in verification earlier.
Flexibility & AI Models
Amelia:
And this system is flexible in terms of models and workflows, right?
Rob:
Exactly. No two design teams work the same way, so flexibility is essential.
The system supports:
- Proprietary in-house models
- Cloud-based frontier models
- Hybrid approaches
Interestingly, less than 0.1% of typical LLM training data is HDL (hardware description language). That’s why combining AI with EDA tools—and fine-tuning models with domain-specific data—is so important.
Key Benefits
Amelia:
What are the biggest benefits?
Rob:
Three major ones:
- Faster, more effective verification
Reduces schedule risk and improves quality. - “Left shift” of verification
Enables non-experts to perform high-quality verification earlier. - Training the next generation of engineers
Faster design cycles mean more learning opportunities. More iterations = faster skill growth.
Fish Fry Fun: Midwest Edition
Amelia:
Before we wrap up—let’s talk fish fry! We both have Midwest roots. What’s the difference between a regular fish fry and a Midwestern one?
Rob:
Oh, fish fries are near and dear to my heart. I grew up in Wisconsin and went to Purdue.
Sure, fish and chips are good—but a true Wisconsin fish fry is different. It’s not just the food—it’s the experience.
It’s Friday night, community gathering, bowling alleys, rye bread, coleslaw, perch, walleye… but most importantly, it’s the heart and soul behind it. That’s what makes it special.
🎤 Closing
Amelia:
I love it. Rob, it’s always a pleasure—thanks so much for joining me!
Rob:
Thanks, Amelia—looking forward to next time.
Amelia Dalton:
Well folks, that’s all I’ve got for this week’s Fish Fry!
For more information, check out this week’s Fish Fry page on EEJournal.com or our YouTube channel.
Be sure to follow EEJournal on social media—Facebook, LinkedIn, BlueSky, Mastodon—and subscribe to our YouTube channel for tons of tech content, including our popular Chalk Talk series.
Got a cool technology to share? Or just want to chat?
Email me at amelia@eejournal.com or post on our forums.
For the week of April 3rd, 2026—I’m Amelia Dalton, and you’ve been fried.
