AI Design Bottlenecks Shift in Next 2 Years | McCord artwork

AI Design Bottlenecks Shift in Next 2 Years | McCord

MTS

November 10, 2026

Nominal co-founder and CEO Cameron McCord joins MTS to discuss raising $240 million at a $3 billion valuation to bring frontier AI agents to hardware engineering.

0:00

Cameron McCord

I think the bottleneck always has been validation in the real world. And I think that we're going to see over the next year, 18 months, two years, the design space is going to get really exciting from an AI agent perspective. But the bottleneck is going to shift to the area where we've been focused all along, which is testing, validation, verification. There's no substitute for proving that these systems work in the real world and then being able to actually feed, ideally, agentically those results back early in the design spaces. I think where a lot of this is going.

0:34

Sophia

Hello, everyone, and welcome back to MTS. Today, we are joined by Cameron McCord, the co-founder and CEO of Nominal, which just raised $240 million at a $3 billion valuation to bring AI to hardware engineering. Cameron, congratulations and welcome to MTS.

0:54

Cameron McCord

Thank you, Sophia. It's awesome to be here.

0:56

Sophia

Awesome to have you. Big day for people who aren't familiar with Nominal. Tell us a little bit about the origin story and the big problem you guys are working on.

1:04

Cameron McCord

Yeah. Today is really about sharing with the world, Nominal's hardware intelligence platform. But we started out actually solving a very simple problem. How do we give hardware engineers superpowers? How do we help them with analysis after they run a test for their hardware systems?

In the four years we've existed, that ambition has expanded pretty drastically. Today, our platform combines a lot of data infrastructure and other capabilities in a context layer to allow agents to do things that hardware engineering humans used to be the only ones that they could do. We've done a lot in these four years.

1:49

Sophia

Tell us a little bit about the hardware landscape. We're seeing software being completely transformed because of AI. Building anything is practically free, even testing and cybersecurity, whatever it might be. We're seeing a lot of innovation happen on that front. When it comes to hardware, it's not really the same.

Can you paint the picture of the wider landscape?

2:09

Cameron McCord

Yeah. Honestly, Frontier AI, as you said, is transforming so many industries. We talk about medical industry, legal, sales, revenue operations, marketing. All of these areas are being really transformed by AI, but that same luxury really isn't present in the hardware engineering world, and that's so much of what our mission is. I think when you look at AI companies that have been successful, is this diffusion layer of frontier model development capabilities into operationalizing it for a specific sector. I think the thing that you see is that whoever can help gather and maintain the data and context for decisions that humans are making is going to have a huge advantage. I think that's really difficult with hardware.

When I think about nominal, we've had to build almost three different businesses in one. We built a ton of zero-to-one infrastructure and platform to be able to get data off of hardware systems, be able to normalize it, aggregate it, and tag it, structure it. We built this agent layer now on top of that. Then we've also built more traditional applications that humans use to enrich data and the power of the hardware intelligence platform is like all of those things can come together to actually give hardware engineers a taste of what many software engineers have now had for several years.

3:35

Sophia

What's fundamentally different about hardware engineering? Why can't they just use some of the software engineering AI tools?

3:41

Cameron McCord

Yeah, it really goes back to so much of it.

We've invested a ton in our ability to do what's called data capture but also command and control. Being able to actually issue out commands to control real pieces of hardware that are performing often in mission critical applications. When we go into a typical industrial customer, data they have is often residing in 10, 15, 20 places. This is in manual spec sheets. It's things that a technician standing next to a piece of hardware is interpreting with a lot of human judgment. It's drawings, it's designs, it's telemetry and sensor data. So the data problem and challenge is really just so vast, and I think that's something that is very solved now, I think in software engineering, but still remains a big problem in the hardware space.

I'm really excited. I think we're seeing AI transform, we call it the hardware development life cycle. That's the term we use to describe everything that goes into building something in the real world. Then we're seeing AI really transform the ability to do generative designs and you think of CFD and modeling and simulation, but also CAD, so the physical design space, software and AI agents can now iterate drastically over that. That makes me really excited because I think the bottleneck always has been validation in the real world. I think that we're going to see over the next year, 18 months, two years, the design space is going to get really exciting from an AI agent perspective, but the bottleneck is going to shift to the area where we've been focused all along, which is testing, validation, verification. There's no substitute for proving that these systems work in the real world and then being able to actually feed ideally, identically those results back early in the design space is like, I think where a lot of this is going.

Read the other 9 minutes of this episode here.

$15 gets you this episode and 99 more. Credits never expire. After checkout, one link brings you back here to read it.

Not what you expected? Full refund within 14 days.

Thousands of transcripts fetched by people building searchable podcast archives

Fetch transcripts with code

The same credits work from a script or an AI agent. The demo key returns a sample episode in full, no card needed:

request
curl -H "x-api-key: pt_demo" \
  https://spoken.md/transcripts/1000651996090

Markdown with the speakers named, for your notes, your knowledge base, or anything that makes HTTP calls.

From $0.10 per transcript. No subscription. Credits never expire. Prices exclude VAT, added at checkout for EU customers. Not what you expected? Email us within 14 days with 20 or fewer credits used and we refund the pack in full.

Using your own key:

request
curl -H "x-api-key: YOUR_KEY" \
  https://spoken.md/transcripts/1000800912515