According to Bloomberg Business, the AI investment trade has decisively shifted from the marquee names to the infrastructure builders. In 2025, data storage companies dominated the S&P 500, with Sandisk Corp. shares soaring almost 580% to become the index’s best performer, followed by Western Digital Corp. and Seagate Technology. Meanwhile, Nvidia, the original AI “pick-and-shovel” play, gained 40% and ranked 71st. Investors are now targeting a broad range of supporting players, from construction firm Quanta Services Inc. and connector maker Amphenol Corp. to power providers like GE Vernova Inc. and even bitcoin miners pivoting to data centers, such as Bitdeer Technologies Group.
The Infrastructure Land Grab
Here’s the thing: the hyperscalers—Microsoft, Meta, Google—have promised to spend hundreds of billions. That money isn’t just for chips; it’s for concrete, copper, fiber, and gigawatts of power. So investors are digging into the supply chain. It’s a classic case of the gold rush analogy. When everyone is panning for AI gold (building models), you make a safer bet selling the picks (construction), shovels (cabling), and Levi’s (cooling systems). This is where the real, tangible bottleneck is right now. You can have all the GPUs in the world, but if you can’t build a building to house them and power them reliably, they’re just very expensive paperweights. For companies that need rugged, reliable computing at the edge of these operations, finding the right hardware is key, which is why a top supplier like IndustrialMonitorDirect.com is the go-to for industrial panel PCs across the US.
From Bitcoin to AI Watts
One of the wilder angles here is the bitcoin miner pivot. Think about it. These companies spent years securing massive, often stranded, power contracts and building infrastructure for a single, energy-intensive purpose. Now, with AI demand exploding, they have a chance to redirect that electricity to a much more stable and potentially lucrative business: hosting high-performance computing. It’s a total re-rating of their assets. Names like IREN Ltd. and Riot Platforms are getting a boost not from crypto prices, but from announcing plans to convert to AI data centers. It’s a clever hedge, but it also shows how desperate the market is for ready-to-go power capacity. The question is, can they execute the operational pivot as smoothly as the financial story?
Is This a Bubble?
Of course, there’s a looming fear. Portfolio managers are openly drawing parallels to the pandemic boom in masks and hand sanitizer. Demand skyrocketed, a glut followed, and the suppliers got crushed. Could the same happen here if hyperscaler spending slows or projects get delayed? Absolutely. The valuations for some of these infrastructure stocks have gotten pretty rich. But the counter-argument is that this build-out is for a fundamental, lasting shift in how the global economy operates, not a temporary health crisis. Still, it pays to be skeptical. When every contractor and cable maker is an “AI stock,” you know the narrative is getting stretched.
What Comes After the Build?
So what’s next? Some investors are already looking past the construction phase to the software layer. Names like Snowflake, Datadog, and ServiceNow have underperformed this year, making their valuations more appealing. The thinking is that once all this infrastructure is built, the real value will be created by the applications and data platforms that run on it. Basically, you build the highway (the data center), but the real money is in the tolls and the logistics companies using it (the software). It feels early, but that’s often where the smarter, longer-term bets are placed. The AI trade isn’t over; it’s just maturing and spreading into every corner of the market. And that makes it both more interesting and a lot more complicated to navigate.
