Recently, I was reading a market analysis about investments in AI infrastructure and found a pretty interesting phenomenon—while everyone is watching chip companies, some companies hidden within the supply chain are quietly cashing in on the benefits.

First, let’s talk about the background. The U.S. stock market in February was indeed a bit rough: the Nasdaq fell 3%, the S&P 500 dropped 1%, and investors were somewhat concerned about the sustainability of AI spending. But this actually reflects a bigger issue—AI infrastructure buildout goes far beyond just chips.

I’ve noticed that the real opportunities may be in companies that provide the infrastructure for AI data centers. For example, Comfort Systems USA, which mainly operates in commercial and industrial HVAC business. That may sound boring, but don’t be fooled—cooling systems for data centers are a high-end demand. With AI compute capacity exploding in demand, these data centers need precision liquid-cooling systems and modular cooling solutions. This niche market has margins that are surprisingly high, and it’s also attracting a large amount of M&A activity. Comfort Systems USA’s expected revenue growth rate this year is 20.3%, profit growth is 28.2%, and its recently raised weekly earnings outlook was increased by 20.9%. Behind these figures is a surge in HVAC demand.

Now, let’s look at the chip side. NVIDIA is, of course, the clear leader. Recently released the Vera Rubin chip delivers 10 times the performance of its previous generation, with shipments starting in the second half of the year. But even more worth paying attention to is NVIDIA’s forecast that by 2030 AI infrastructure spending will reach $3–4 trillion. What does that mean? It means the entire industrial chain will be pulled along.

In addition to cooling, advanced packaging is another key area. Onto Innovation focuses on this field. Its 3Di technology has already passed certifications from two high-bandwidth memory customers, and it is now pushing forward orders for 2.5D AI packaging. This year, the expected revenue growth is 19.5% and profit growth is 29%.

Another angle is raw materials. Rio Tinto is a global mining giant, and AI infrastructure requires large amounts of raw materials such as copper and aluminum. They expect revenue to grow 10.7% and profits to increase by 21.8% this year, reflecting an optimistic outlook for demand for base metals.

Finally, there’s the capital side. Alternative asset managers like Brookfield Asset Management are making extensive moves into infrastructure and renewable energy. AI data centers have massive electricity demand, and these firms are investing in related energy infrastructure. They expect revenue to grow 12.5% and profits to rise by 15.2% this year.

So you see, this AI wave isn’t just chip companies making money. From cooling systems and advanced packaging to raw materials and energy infrastructure, the entire industrial chain is benefiting. While these companies may not be as well known as NVIDIA, their growth potential should not be underestimated. If you’re looking for the second wave of investment opportunities in the AI era, these supply-chain companies are definitely worth paying attention to.