Designing multi-billion transistor chips for hyperscale data centers and autonomous systems requires far more than basic generative AI models.
In this featured interview on CNBC's Mad Money with Jim Cramer, cadence design systems President and CEO anirudh devgan deconstructs why advanced eda software is indispensable to the global semiconductor ecosystem. As hyperscalers and traditional chipmakers accelerate custom silicon production, electronic design automation software serves as the foundational engine powering physical microarchitecture layouts.
The discussion highlights how artificial...
What Anirudh Devgan said
Written from the verified transcript and checked against it. Every figure links to the moment it was said.
Anirudh Devgan, CEO of Cadence Design Systems, argued that AI is a tailwind, not a threat, to his company. He stated that all chips designed today use Cadence software, with 45% of business from system companies and 55% from semiconductor firms like Broadcom, Nvidia, and Intel. He described AI as a 'turbocharger' on top of physically accurate base tools, citing a 10 to 40x improvement in verification turnaround time for Nvidia using Cadence's AI agents with ChipStack. Devgan discussed increased collaboration with Intel under Lip-Bu Tan, aiming to co-optimize products for both Intel's own chips and its foundry business. He expressed excitement about physical AI, predicting multi-trillion dollar markets in cars, drones, and robots, with automotive semiconductor content expected to grow 10x. He noted that 70-80% of Cadence's business comes from 70-80 customers, emphasizing a 'win with the winners' strategy.
Key takeaways
45% of Cadence's business now comes from system companies designing chips, with 55% from semiconductor companies.
Nvidia saw 10 to 40x improvement in verification turnaround time using Cadence's AI agents with ChipStack.
Cadence is working more closely with Intel under Lip-Bu Tan to co-optimize products for both Intel's own chips and its foundry business.
Physical AI markets in cars, drones, and robots are projected to be multi-trillion dollar, with automotive semiconductor content expected to grow 10x.
70-80% of Cadence's business comes from 70-80 customers, all major companies across geographies and segments.
Special treat. Okay, right now, except for today, we have an incredible bull market in semiconductors and semiconductor capital equipment. But what about Cadence Design Systems? This is a software company that helps design chips and other electronics that I have liked since probably 1989. Okay, here's a stock that's down more than 10% over the past 12 months, victim of AI displacement worries. Don't worry, we're going to explain that. Doesn't seem to matter that Cadence tools are used by customers like Nvidia to create the chips at the heart of the data center. This stock keeps getting lumped in with the software cohort and punished for it. Didn't even matter when the company reported a very strong quarter at the end of July. If anyone actually bothered to listen to management over the past year, they know that AI is in fact a major tailwind for Cadence. They help design the chip, for heaven's sake. This is serious technology that can't be vibecoded out of existence. Don't take it from me. Let's go straight to the source: Dr. Anirudh Devgan. He's the president and CEO of Cadence Design Systems. Dr. Devgan, welcome back to Mad Money.
Okay, Anirudh, it's always great to see you, and I know that I always ask you to educate people. There are many people who aren't quite sure what Cadence does, but somehow, maybe because of analyst notes or what they hear, they hear software and they say whatever they think they're doing, it's not going to exist once Claude takes over. It couldn't be more wrong, correct?
Absolutely. We are an irreplaceable part of the semiconductor industry, and there are two ways we play in the semi industry. One is all the chips that are designed in the world today, whether they're data center chips or phone chips or computing chips, are designed with Cadence software. The demand for those chips is only growing because the chips get bigger, there are more of them used in all these applications, and all the data center companies are now designing chips, as you know. So 45% of our business now is system companies—these new companies designing chips—and 55% is the semi companies, like Broadcom and Nvidia and Intel and all. So the demand for our products is only increasing and will continue to increase for the next 5 to 10 years. All right. And then the second part is when we apply AI to our own products to improve them.
Yeah. Right now I think that people have to understand that if you apply AI to your own product, you do it faster, you can do it smaller. And we know we have this problem with Intel and Moore's Law, we'll get to that in a moment. But the fact is, without AI you can't be as great as you'd like to be, and with AI, any software company would still be way over their heads. And you even give examples in your conference call. It's just you wished it worked, but it doesn't.
A good analogy for AI plus Cadence is like a turbocharger. Our base tools are like the V6 or V8 engine, and AI by itself cannot do them because it's physically accurate. These chips are at 200 billion transistors at 3 nanometer, and they have to be first-time right. So there's a lot of physics and mathematics involved in that which AI per se cannot do, so you still need our main tools. But AI can be a turbocharger. It can give more scenarios because what the customers want is to improve the performance of the chip. If it's 3 GHz, they want 3.5 GHz; if it's 100 watts, they want 90 watts. So it's an optimization problem, and AI can give more scenarios that we can run and improve the performance of the tool.
Okay. So when Jensen Huang comes out and says, "Listen, we've got Vera Rubin, and then after Vera Rubin, we've got Richard Feynman," he has to do this in conjunction with you or it just won't work.
8 more exchanges in this transcript
Sign in free to read the rest of this interview. No card required.
Devgan, A. (2026, August 24). Cadence CEO Anirudh Devgan on AI, Semiconductor Demand & the Future of Chip Design | CNBC Mad Money [Interview transcript]. Cadence Design Systems. CEOInterviews.AI. https://ceointerviews.ai/interview/1267128/
MLA
Anirudh Devgan. "Cadence CEO Anirudh Devgan on AI, Semiconductor Demand & the Future of Chip Design | CNBC Mad Money." Cadence Design Systems, 24 Aug. 2026. Transcript, CEOInterviews.AI, https://ceointerviews.ai/interview/1267128/.
BibTeX
@misc{devgan2026_1267128,
author = {Anirudh Devgan},
title = {Cadence CEO Anirudh Devgan on AI, Semiconductor Demand \& the Future of Chip Design | CNBC Mad Money},
howpublished = {Interview transcript, Cadence Design Systems. CEOInterviews.AI},
year = {2026},
month = {aug},
url = {https://ceointerviews.ai/interview/1267128/},
note = {Speaker-attributed transcript with timestamps}
}
