The "Magnificent Seven" have been market leaders since 2023 but are not quite an artificial intelligence (AI)-focused group. While many within the cohort have AI products, they aren't pure plays.

Instead, I'm proposing a new group of three that informs investors on how strong AI demand is. That group, consisting of Nvidia (NVDA 6.18%), which is also a member of the "Magnificent Seven," Super Micro Computer (SMCI 8.89%), and Taiwan Semiconductor (TSM 1.26%), gives investors a look at the hardware side of AI.

So, if you're looking for a basket of stocks to track AI infrastructure buildout, this "Terrific Trio" is the way to go.

This trio has their hands in most AI products

Starting at the base, computing hardware requires semiconductor chips, and Taiwan Semiconductor produces the vast majority of them. To produce more powerful chips, manufacturers place the conductive traces closer together, which allows more components to be packed onto a board. To describe these chips, chip manufacturers name them based on how far apart the traces are. Currently, the smallest TSMC can go is 3 nanometers (nm), but it's already working on its 2 nm technology as well.

Taiwan Semiconductor makes chips for some of the world's biggest companies and is one of Nvidia's key suppliers.

Nvidia gets the most attention of this trio, as its parabolic rise since early 2023 resulted in the stock increasing nearly 500%. This excitement is due to Nvidia's graphics processing units (GPUs), which are best-in-class for training AI models. GPUs are built to process complex calculations like training AI models.

But what benefits Nvidia is the number of GPUs purchased when a supercomputer is built. These devices require thousands of GPUs to provide users with top-end computing power, and with the rumored price of an H100 GPU coming in at around $30,000 a piece, that generates a lot of revenue for Nvidia.

Lastly is Super Micro Computer. Haphazardly assembling a few hundred or thousand GPUs together without the networking or architecture experience is a recipe for disaster. Many companies without the experience necessary to get the most out of their GPUs go to Supermicro to build their servers.

With Super Micro Computer's highly modular system, it can spec in a server of just about any size and focus that a client could want.

Together, these three offer a great pulse on the AI movement, as their products are essential table stakes for getting into the AI game. But they've each had a strong 2024, so is this a good trio to buy right now?

The trio has had a strong 2024

This group has been a fantastic one to own in 2024. Supermicro's led the charge, rising more than 280%.

NVDA Chart

NVDA data by YCharts

After those numbers, many investors may be worried about overpaying for any of the stocks in this group. However, that argument breaks down if you examine the forward price-to-earnings (P/E) ratio (which uses analyst projections over the next 12 months).

NVDA PE Ratio (Forward) Chart

NVDA PE Ratio (Forward) data by YCharts

The forward P/E ratio is important when the company you're assessing is going through a massive transformation, like these three are due to AI demand. Additionally, when comparing this metric to a more typical price-to-earnings ratio (which uses trailing earnings to value the company), it's best to take a few points off because these are only projections.

With that in mind, it's clear that Nvidia and Taiwan Semiconductor are still within the realm of buying, although Supermicro has gotten quite expensive. However, Nvidia's stock also traded for more than 50 times forward earnings in 2023, so Supermicro could grow into its valuation through sustained outsize growth.

So, should you buy the "Terrific Trio?" I'd say yes if you want an absolute pulse on the AI market. But, if you're looking to maximize your portfolio returns, I'd only buy Nvidia and Taiwan Semiconductor, as Supermicro might be due for a pullback after its stellar run to start the year.