Nvidia’s $44 billion quarter just redrew the map for commercial real estate.
While Microsoft, Amazon, Meta, and Alphabet are entering a more strategic phase of infrastructure planning, Nvidia is still in hypergrowth mode. CEO Jensen Huang highlighted the explosive growth in AI demand: “AI inference token generation has surged tenfold in just one year, and as AI agents become mainstream, the demand for AI computing will accelerate.”
But this isn’t just about chips. It’s about the infrastructure that supports them—and that’s where commercial real estate comes in. Huang recently described modern data centers as “AI factories” and emphasized their growing value in a 2025 keynote address: “You apply energy to it, and it produces something incredibly valuable…these things are called tokens.” In other words, the inputs are power and land. The output is AI at scale.
Big tech’s capital expenditures remain elevated—$14.9 billion for Microsoft, $17.2 billion for Alphabet, and $25 billion for Amazon—but the tone has shifted. Most are talking about optimization, modular growth, and smart deployments. Nvidia is still talking about global buildout.
For CRE professionals, this moment isn’t just about watching tech expand. It’s about enabling that expansion. Where AI can go next will depend on power, permitting, and real estate that’s ready to scale.
AI Is Still Booming, But Efficiency Is the Next Battleground
As Nvidia ramps up to meet surging demand, its peers are pressing pause to optimize portfolios.
Every major firm reaffirmed that infrastructure investment remains a top priority in Q1. But beneath the surface, a pattern is emerging—Microsoft, Amazon, Alphabet, and Meta are now favoring modular builds, phased deployments, and tighter cost controls. Meanwhile, Nvidia continues to accelerate, driven by a global race to build out what CEO Jensen Huang calls “AI factories.”
“We’re seeing incredible demand around the world,” Huang said. “Every data center in the future will be generative AI-enabled.”
That demand is pushing Nvidia to move fast—rolling out new chip platforms, announcing U.S.-based supercomputer production, and partnering globally to deliver the next wave of AI infrastructure.
Compare that to Microsoft, which is targeting $80 billion in data center investment this year—but has also reportedly canceled leases totaling several hundred megawatts to optimize its footprint. CEO Satya Nadella summed it up on the earnings call: “Demand continues to exceed supply,” but the strategy is shifting to “more intelligent infrastructure deployment going forward.”
Amazon is navigating a similar recalibration. In April, Wells Fargo analysts noted AWS paused certain lease negotiations with colocation partners to digest recent buildouts. AWS’s Kevin Miller framed it this way: “Some paths cost too much or take too long. That’s just part of scaling responsibly.”
Alphabet is also dialing in. CFO Anat Ashkenazi described a move toward “scalable, modular capacity expansion,” with server demand leading and physical builds following in phases. The company still plans to spend around $75 billion in CapEx this year, but warned the pace will fluctuate based on construction timelines and delivery windows.
The common thread? CRE decisions now hinge on timing, flexibility, and grid readiness—not just land.
Uniform site growth is giving way to retrofits, rolling entitlements, and stepwise power upgrades. “For developers and investors, winning in this environment will mean delivering not just square footage, but strategic infrastructure alignment,” said Dianne Crocker, research director at LightBox.
Electricity is the New Zoning
The next wave of AI growth won’t be determined by ambition—it’ll be decided by amperage.
Microsoft CEO Satya Nadella put it bluntly: “Power is the constraint, not ambition.” Add that to Nvidia’s push to manufacture AI supercomputers in the U.S.—part of a plan to generate up to half a trillion dollars in AI infrastructure—a move that will intensify demand for power-ready industrial sites.
On Nvidia’s Q1 2025 earnings call, Huang outlined the company’s strategy in clear terms: “Companies and countries are partnering with NVIDIA to shift the trillion-dollar traditional data centers to accelerated computing and build a new type of data center—AI factories—to produce a new commodity: artificial intelligence.”
That constraint is redefining site selection. In Northern Virginia, a July 2024 grid failure forced 60 data centers onto backup power, narrowly averting a blackout. In California, PG&E reported a 40% spike in 2025 data center power requests—driven by surging demand in the Bay Area. Across core markets, electricity has become the ceiling on development.
Ahead of yesterday’s earnings call, Huang previewed the growing constraint: “We’re working closely with partners to plan for capacity well in advance. It’s not just about chips. It’s about power, cooling, and real estate.”
CRE professionals have long measured opportunity by land and entitlements. But today, it’s megawatts that matter. “In many markets, electricity is the new zoning,” said Crocker. “It’s determining not just what gets built — but whether anything can be built at all.”
Because the next AI factory won’t break ground where the zoning is easy—it’ll land where the power is ready.
What CRE Clues to Watch in Q2
Data centers are no longer just a CRE niche—they’re ground zero for big tech’s next phase of growth. Power access, modular design, and development-ready land are now the gating factors for AI infrastructure—and CRE professionals are on the front lines.
The earnings calls from Microsoft, Amazon, Alphabet, Meta, and Nvidia all pointed to the same underlying truth: the speed and scale of AI growth will be defined by real estate decisions. From grid constraints in Virginia and California to phased buildouts and onshoring chip production, Q1 offered a preview of the new rules of site selection.
With Q2 earnings around the corner, CRE leaders should be listening for where hyperscalers are planning new capacity, how they’re addressing power procurement, and whether modular or retrofit strategies are gaining traction. Watch how tech firms talk about land—what’s being bought, entitled, or deferred. And pay attention to power: where it’s available, where it’s delaying deals, and how it’s shaping expansion timelines.
This next quarter will be less about ambition and more about execution. CRE firms that can anticipate tech’s infrastructure needs—and meet them with speed—won’t just participate in the AI wave. They’ll help direct it.
Subscribe To Insights
Be the first to get the latest news, updates and insights from LightBox.
