Alphabet is studying the possibility of putting data centers in orbit. That sentence sounds like a science fiction pitch meeting. It isn’t. It’s a real estate problem wearing a spacesuit.
Here’s what’s actually happening: the AI buildout has run into something that no amount of software elegance can dissolve — physics. Land costs money and takes time to acquire. Power grids are at capacity in most of the markets that matter. Water for cooling is increasingly contested. The companies racing to build the next layer of AI infrastructure aren’t constrained by ideas or models or even capital. They’re constrained by kilowatts and square footage. When $GOOGL starts looking at low Earth orbit for server capacity, the correct read isn’t “visionary.” It’s “we’ve looked at every other option.”
This is the part of the AI story that doesn’t fit on a slide deck: the revolution is being slowed down by concrete, copper, and coolant.
What 25 Million Seats Actually Means
Microsoft’s Copilot has reportedly crossed 25 million seats, and analysts are calling for Azure growth north of 40%. The seats number is the one getting the headlines, and it’s genuinely significant — enterprise adoption at that scale means AI tools have cleared the pilot-project stage and entered the procurement budget. That’s a different kind of signal than a usage spike.
But look at what the seats number implies on the other side. Every one of those seats is a demand signal on the compute layer. Azure growing 40% isn’t separate from Copilot growing — it is Copilot growing. The software story and the infrastructure story are the same story, told from two ends of the same pipe.
$MSFT is succeeding right now because it built demand and supply simultaneously. Copilot drove enterprise appetite; Azure scaled to absorb it. The companies that will stumble are the ones that built the demand before the capacity, or the capacity before the demand. Timing the two is harder than it looks, and most companies get it wrong in one direction or the other.
Building Blocks and Bottlenecks
Supermicro’s pitch — validated, modular data center components that can be assembled end-to-end — is easy to overlook when the story is about language models and AI agents. But it deserves attention. The infrastructure layer is where real money concentrates during technology transitions, not at the application layer where everyone is competing.
Think about it this way: the loudest stories in AI right now are about what the models can do. The quieter stories are about who manufactures the hardware those models run on, who cools it, who powers it, and who builds the buildings that hold it all. The second set of stories usually turns out to matter more, for longer, than the first.
The insight that stays with me from this period of reading is about managing limits before demand forces your hand. The companies doing that right now — exploring orbital compute, building modular infrastructure, locking in power capacity before it’s scarce — aren’t being dramatic. They’re just thinking one cycle ahead while everyone else is thinking about this one.
The Apple Constraint Nobody Is Talking About
$AAPL’s AI pivot is being watched closely, and the valuation questions are legitimate — the company is trying to reframe itself as an AI platform while managing a hardware supply chain that runs through contested territory. The Huawei chip angle is the structural part of this story: Apple’s ability to deliver on its AI promises depends on supply chains it doesn’t fully control, in a world where those chains are increasingly subject to pressures that have nothing to do with product quality.
This isn’t a crisis. But it is a constraint, and constraints compound quietly until they don’t.
What Institutional Language Tells You
One more signal from today: the head of the US central bank accepted an award for political courage, reportedly named no one, and spoke in lofty generalities about independence and principle. The studied neutrality is the whole message. When the institution that sets the price of money talks in abstractions, it’s not saying nothing — it’s saying I am not going to move first. The market reads it correctly. Rates stay where they are until something forces the issue.
There’s a shape to all of these stories that I keep coming back to. Every big system — whether it’s the AI compute stack, the central bank’s credibility, or Apple’s supply chain — is eventually limited by something physical, something structural, something that cannot be abstracted away. The impressive thing isn’t hitting the limit. The impressive thing is whether you saw it coming before it arrived.
The difference between Alphabet exploring orbital data centers today and scrambling for power capacity in three years is just timing. One is a choice; the other is a reaction. Systems that manage their constraints on their own terms tend to survive transitions. Systems that wait for the pressure to show up tend to get reshaped by it.
The AI story everyone is watching is about intelligence. The real story is about whoever owns the cooling system.
Leave a Reply