So, here's a thing about building the future: sometimes you just need a really, really big computer. And when you're Microsoft (MSFT) and you're trying to fuel the artificial intelligence boom, you go lease a roughly 700-megawatt data center in Texas. This isn't just any plot of land; it's a site in Abilene that was originally supposed to be for Oracle Corp. (ORCL) and OpenAI. According to reports, negotiations between those two fell apart over financing and shifting requirements, leaving the door open for Microsoft to swoop in.

The deal, secured from developer Crusoe, puts Microsoft right next door to Oracle and OpenAI's planned "Stargate" campus. It's a physical manifestation of the intensifying competition to build the computing backbone for AI. Meta Platforms Inc. (META) reportedly looked at the site too but passed. For Microsoft, this is part of a much bigger push. The company committed about $50 billion in new data center leases just last quarter, reversing a slowdown from a year ago. When demand for AI compute is surging, you build—or in this case, lease—where you can.

This aggressive infrastructure expansion is exactly why Wall Street is getting more bullish on Microsoft's role in what analysts are calling the "AI supercycle." On Tuesday, Bank of America Securities analyst Tal Liani reinstated coverage of Microsoft with a Buy rating and a $500 price target. That implies about 31% upside from recent levels.

Liani's thesis is pretty straightforward: Microsoft isn't just participating in AI; it's sitting at the center of it. He sees the company delivering sustained, multi-year growth as it monetizes AI across its entire stack. On one layer, you have Azure, which is becoming the go-to backbone for enterprise AI workloads. On another, you have products like Microsoft 365, Dynamics, GitHub, and Windows, all of which are weaving in AI features designed to make users stick around and spend more.

The numbers back up the optimism. Liani forecasts total revenue growth of 15% to 17% over the next three years. The engine for that will be the Intelligent Cloud segment, which he expects to grow by 24% to 28% as AI workloads scale up. To justify the $500 price target and a valuation multiple above peers (24x 2027 earnings versus a roughly 19x peer average), Liani points to Microsoft's sheer scale and its enviable positioning across the AI ecosystem.

But here's the catch, and it's a big one: building the future is expensive. Liani expects margins to feel the pinch in the near term from all this AI-related spending. Gross margins could decline by about 340 basis points through fiscal 2028. The company's strong software mix should help keep operating margins above 46%, but free cash flow margins are another story. They could drop into the low-20% range from about 30% in fiscal 2024.

Why? Because capital expenditures are projected to skyrocket from $44 billion to about $143 billion by fiscal 2028. That's a staggering increase, but Liani views it as a necessary, if temporary, investment phase. You have to spend money to make money, especially when you're trying to capture a generational shift in technology.

Of course, it's not all smooth sailing. Liani highlighted several key risks. First, there's the pace at which Microsoft can convert its massive AI backlog into actual, recognizable revenue. Then there's the partnership with OpenAI—a crucial relationship, but one with evolving economics that could change. Finally, there's the overarching question everyone is asking: how long will this AI investment cycle last? Is this a durable supercycle or a more fleeting boom?

Microsoft shares were down 2.72% at $372.58 at the time of publication on Tuesday. In the grand scheme of a multi-year AI build-out, a single day's move is just noise. The real story is whether Microsoft's billions in data center leases and its sprawling software empire can indeed make it the central hub of the AI revolution. Based on the latest lease in Texas and the analyst conviction on Wall Street, the company is certainly betting that way.