🗞 6 Critical Suppliers Powering the AI Boom

Happy Sunday! 👋 

This week we’re taking a look at 6 critical suppliers that are quietly powering the AI boom. 🚀 

Let’s dive in!

Fiscal.ai Update

Adjusted Metrics 📊 

This week, the Fiscal.ai team launched Adjusted Metrics!

Through our Adjusted Tab, you can now track dozens of adjusted metrics for all US and Canadian companies.

Fiscal.ai also pulls in consensus analyst estimates for adjusted metrics so you can easily pair historical data with forward assumptions.

Industry Comps 🏭

This week we completely re-vamped our industry comps list.

Now for all global companies, you can automatically see a list of the closest peers/competitors by simply selecting the “Industry” tab.

The comps list comes pre-populated with Market Cap, Gross Margin, and Forward P/E, but you can add any metrics you prefer from our massive catalog.

Featured Story

6 Critical Suppliers Powering the AI Boom

During the 19th century, thousands of hopeful prospectors in the U.S. traveled west in hopes of finding gold. This eventually became known as the California Gold Rush.

For those familiar with the story, you likely know where I’m going with this.

While some individuals were lucky enough to strike gold and make a fortune, most were unsuccessful. In the end, those who benefitted the most weren’t the prospectors themselves, but instead the merchants who supplied tools, food, and other necessities to the masses of miners. These merchants have since become known as the “picks and shovels” providers.

McKinsey estimates that the world will need to spend roughly $7 trillion on AI infrastructure by 2030 in order to keep pace with the demand for computing power. Whether you agree with that estimate or not, the direction is clear.

In what has become sort of a modern gold rush, the world’s largest companies are pouring capital into data centers to win the AI race. To put some numbers on it, Amazon, Microsoft, Alphabet, and Meta have spent nearly $300 billion in combined capital expenditures over the last 12 months!

And just like the merchants of yesteryear, there’s a wave of suppliers reaping the benefits of this spending. These seemingly boring suppliers don’t get nearly the same notoriety as the big tech giants, but they’re critical enablers of data center buildouts.

Here are 6 of the critical suppliers powering the AI boom:

1. Amphenol

You might not recognize the name, but Amphenol is one of the largest manufacturing businesses in the world.

They make thousands of seemingly simple products like point-to-point cables, power distributors, and sensors that several end markets (including IT and Data Communications) rely on.

High performance data centers in particular require Amphenol’s high density connectors and cables to transfer massive amounts of data at extremely high speeds.

This has created a massive acceleration in revenue growth for Amphenol. During the company’s Q3 earnings report this week, they reported 53% revenue growth compared to a year ago which is the company’s fastest growth rate in more than two decades.

2. Quanta Services

When a company decides it wants to build a new data center, Quanta Services is one of the first calls they make.

Among other things, Quanta is one of the largest designers and installers of electric transmission and distribution lines. And with the recent surge in data center power consumption putting a strain on the existing electric grid, data center operators are turning to Quanta early in the design process to ensure that they will have the actual power and electricity that they need.

Quanta’s CEO Duke Austin described the advantageous position that they’re in during the latest conference call: “If we're going to lead the country and the world, you have to have power, and we're right in the middle of the infrastructure”.

3. Arista Networks

Arista makes high-speed switches and routers, as well as specialized operating software, that connects servers and storage systems within data centers.

They are often referred to as a “tollbooth” for AI since data centers depend on Arista’s switches and operating system for speed, reliability, and low latency.

Training large AI models requires moving tons of data between GPUs, which traditional networks aren’t fast enough for. This is why Arista has seen a huge acceleration in revenue since the LLM race began.

4. Vertiv Holdings

Vertiv specializes in powering and cooling solutions.

It’s no secret that GPUs consume a lot of power. That power consumption generates a tremendous amount of heat. To offset the heat, data centers have to use cooling solutions from specialized providers, and one of those providers is Vertiv Holdings.

Vertiv’s liquid cooling systems have been in high demand since the data center race began. This demand has not only led to an acceleration in revenue growth, but has also helped Vertiv more than double its operating margins over the last 3 years.

5. Micron

CPUs and GPUs use a memory system known as Dynamic Random-Access Memory (DRAM). Advanced GPUs in particular require 5x-6x more DRAM capacity than traditional servers.

Micron is one of the largest providers of DRAM chips in the world.

As more and more money continues to be poured into high-performance data centers, demand for Micron’s DRAM chips is outpacing supply. This is leading to increased pricing power for Micron.

Micron CEO Sanjay Mehrotra on the Q4 conference call: “Our HBM performance has been strong, and robust demand, tight DRAM supply, and disciplined execution have significantly strengthened the profitability of the rest of our DRAM portfolio.”

6. Comfort Systems

Comfort Systems provides mechanical, electrical, and plumbing (MEP) contracting services.

In other words, they’re the company that makes massive industrial buildings (like data centers) actually functional, livable, and as the name implies, comfortable.

Given the power distribution and ventilation requirements, data centers require particularly complex MEP systems.

The hyperscalers (AWS, Azure, and Google Cloud) are Comfort Systems’ largest customers, so this recent race to buildout AI infrastructure as fast as possible has led to a massive jump in bookings.