AI Is Starting to Design the Machines That Will Replace It

Last week, I wrote about how AI is rewriting the semiconductor market.

The numbers involved are shockingly large. McKinsey says the chip industry could reach roughly $1.6 trillion by 2030.

And the biggest driver of that growth is computing and data storage. It’s already changing how the next generation of AI gets built.

But there’s something else happening today that might have an even bigger impact on the future of AI technology.

AI isn’t just increasing the demand for chips.

It’s starting to help design them.

The Self-Evolution of AI

Chip design is one of the most complex engineering problems in existence.

Modern processors pack tens of billions of transistors into a single piece of silicon. That means every design involves tradeoffs between power, performance, area, heat and manufacturability.

If you get any of it wrong — you’re looking at delays, lower yields or a costly redesign.

What’s more, a lot of the work is repetitive. Engineers can spend up to 70% of their time writing and testing design code.

But that’s exactly the kind of repetitive, optimization-heavy work that AI is already good at doing.

And that’s why the biggest design software firms are moving quickly to hand this work over to AI.

It can design code, build testbenches, run validation cycles, debug issues and even fix them automatically.

Cadence says its underlying AI-driven design tools have already been used in more than 1,000 tapeouts. That’s industry shorthand for the moment a chip design is finalized and sent to manufacturing, including new chips and updated versions of existing ones.

And this month, Cadence rolled out another AI agent for the later stages of design, where circuits get laid out physically on silicon. It brings AI even deeper into the physical side of chip design.

Synopsys (Nasdaq: SNPS) is moving in a similar direction.

In March, the company introduced a system where multiple AI agents work together across design, testing and simulation. It’s also linking more closely with tools that model heat and other physical effects, which are becoming essential as chips run hotter and more densely packed than ever.

So this is way more involved than having AI simply assist with writing code. It’s moving AI deep into the engineering stack.

We saw an earlier version of this with Google’s AlphaChip.

Image: Google

It says AlphaChip has generated layouts used in every generation of Google’s TPU since 2020. In other words, AI has already helped design some of the hardware used to train and run modern AI systems.

But there are limits to AI’s capabilities today.

AlphaChip’s performance claims have been debated in the research community, and even the companies selling the newest agentic tools admit that humans are still in the loop. Cadence’s own customer examples describe an engineer-in-the-loop workflow, not a completely autonomous one.

And there’s a reason for that.

Chip design is full of tradeoffs that don’t always translate cleanly into code. Timing constraints, edge cases and system-level decisions still require human judgment.

This means AI isn’t about to replace chip designers overnight. But it is taking over the middle of the process.

And that’s already making the design process much more efficient.

Once AI is able to handle more of the repetitive design and verification work, smaller teams can explore more architectures, run more iterations and get to tapeout faster.

That compresses development cycles.

It also creates a feedback loop. AI helps build better chips, and better chips help train and run better AI. Then those systems improve the design process all over again.

It’s a clear example of a compounding advantage, just like we saw with Nvidia’s Ising AI models that could help advance quantum computers, which in turn could advance AI.

And it could have a huge financial impact.

The Semiconductor Industry Association says the market is on pace to reach roughly $1 trillion in 2026. Gartner is even more aggressive, forecasting more than $1.3 trillion in semiconductor revenue this year, with AI chips making up roughly 30% of the total.

That’s a staggering level of concentration. Roughly $0.30 of every semiconductor dollar this year is expected to come from AI chips.

And AI is increasingly helping design the very products driving that growth.

Here’s My Take

AI is rewriting the chip market from the demand side.

And it’s starting to rewrite the supply side too.

It doesn’t mean chip engineers are going away any time soon. But the teams that lean into AI will be able to run more experiments, move faster and bring products to market sooner. And in a business where timing can translate to billions in revenue, that’s the kind of edge that could make or break businesses.

Because this semiconductor boom is no longer just about who can manufacture the most advanced chips.

It’s also about who can design them fastest.

And as AI keeps moving deeper into that process, the chip industry could be entering a phase where machines accelerate their own evolution.

Regards,

Ian KingChief Strategist, Banyan Hill Publishing

Editor’s Note: We’d love to hear from you!

If you want to share your thoughts or suggestions about the Daily Disruptor, or if there are any specific topics you’d like us to cover, just send an email to [email protected].

Don’t worry, we won’t reveal your full name in the event we publish a response. So feel free to comment away!