Cadence Design Systems (CDNS -1.19%) -- a leader in semiconductor and computing system design software -- has been booming. Shares are up nearly 60% in the last 12 months, including a more than 12% increase so far in 2024. Its latest foray is with some help from artificial intelligence (AI).

In early February, Cadence announced a new supercomputer system called Millennium M1 that it says is powered by graphics processing units (GPUs) "from leading providers" -- which could certainly mean Nvidia, the de facto leader in GPU systems and a longtime Cadence collaborator.

Cadence's announcement is part of a broader trend in chip design, and AI (really, just think "accelerated computing" when you see those two letters) is driving the growth. Can Cadence continue its hot streak?

Cadence and its brand of CFD

The Millennium M1 AI supercomputer is really a server, a large computing unit that can be installed either on a company's property (in an office, a private data center, etc.) or deployed in the cloud and accessed via an internet connection. Specifically, Cadence designed this computing platform for accelerating computational fluid dynamics, or CFD.

It's a type of analytics and simulation software used in the design of all sorts of things -- including auto engines and how cars will behave driving down the freeway, jet engines and how a plane will fly, and industrial and manufacturing processes using liquids and other materials. Really, anything that "flows" can be analyzed and simulated using CFD.

Cadence's Millennium Platform, powered by the Millennium M1, is the first CFD and multi-physics simulation product that uses a purpose-built supercomputer for performing such complex computations. Think of it as a turbocharged "digital twin" platform, enabling engineering teams to visualize how a product or project will behave in the wild, before ever manufacturing or building it.

Nvidia CEO Jensen Huang has been talking about the massive potential for digital twins for some time, especially when the metaverse was all the rage a couple of years ago. But CFD and multi-physics simulation actually has some of the most promising potential in the business world.

Cadence has been gearing up for this, even hinting at a big Nvidia product release event a year ago that it was working on big things in CFD.

For reference, these big moves in CFD -- really a type of convergence between the digital and physical worlds -- are why Cadence peer Synopsys is acquiring simulation software provider Ansys, and why hardware testing company Keysight Technologies made a simulation-software acquisition of its own last year.

Does Cadence have a special advantage?

So does Millennium give Cadence a leg up on the competition? It certainly could.

Cadence and Synopsys are the two dominant players in semiconductor design software (called EDA, or electronic design automation). Lots of new devices with chips embedded in them coming to market refill the sails of these two industry giants.

However, Synopsys' move on Ansys will mean significant debt will need to be paid off if the deal is finalized (perhaps in 2025). Meanwhile, Cadence is developing its own services that converge chip design with more mechanical processes in-house. Put another way, its clean balance sheet and best-in-class profit margins will remain intact, while Synopsys will have to digest its acquisition (again, if it happens).

CDNS Revenue (TTM) Chart

CDNS Total Long Term Debt (Quarterly) Chart

Data by YCharts; TTM = trailing 12 months.

After a stellar 2023 that featured soaring profitability, though, Cadence stock is expensive at nearly 50 times expected earnings for 2024. With a valuation like that, even a significant 20%-plus pullback in the share price won't make Cadence cheap.

But innovations like its new AI supercomputer, powered by GPUs and other high-end hardware and software algorithms, make Cadence a premium stock for good reason. Be patient with this investment in 2024, and consider a dollar-cost average plan if long-term ownership of Cadence Design Systems fits your portfolio needs.