By Marc Berte, Founder and CEO of Overview Energy
AI is making energy valuable enough that we’re reconsidering where infrastructure should live. Space is increasingly where that conversation leads.
Orbital data centers are the most visible expression of that shift right now. They’ve attracted real capital, and SpaceX’s endorsement alone moved the conversation. But they’re one answer to a larger opportunity: continuous energy in space is now commercially viable, and the industry is still working out what that unlocks. Not just compute. The whole thing.
That’s the orbital energy economy, and it’s still early enough that the shape of it isn’t settled.
Launch costs alone undersell the shift
The obvious explanation for why orbital energy is suddenly serious is launch costs. Starlink demonstrated something more fundamental though too: that space infrastructure can scale like a manufacturing business rather than a flagship mission. SpaceX has launched over 10,000 Starlink satellites, replenishing failed ones and upgrading generations on orbit. The whole system gets better over time.
That shift from bespoke to industrial is what opened the door to everything else, and once you believe infrastructure in orbit can be built and operated at scale, you start asking what else belongs there. The energy has always been there. The economics are what changed.
Meanwhile, the ground-side pressure has become impossible to ignore. Grid interconnect queues in the U.S. run up to 10 years in some markets, and developers are canceling data center projects not because of land or capital, but because they can’t get power fast enough. Here’s one way to feel the scale of that shift: data centers used to be measured in square footage, and now they’re measured in megawatts. The building is the easy part. The plug is the problem.
Not all the ways to monetize orbit are energy plays. But mapping the full landscape makes clear why energy is now becoming the most consequential one.
Communications is the proven model. Starlink converts orbital solar energy into internet connectivity. The reason communications worked first is that information has extraordinary value relative to the energy required to move it. You don’t need much power to generate enormous revenue. That’s a favorable economics profile, and it’s why Starlink is already profitable while most of the orbital economy is still pre-revenue.
Manufacturing is the longest horizon. Some physical processes like crystal growth, certain alloys, and fiber optics produce powerful structures in microgravity that are difficult or impossible to replicate on the ground. Varda Space has already made that case in orbit for commercial drug development. But the ecosystem needed to scale it doesn’t exist yet, and manufacturing tends to need the supply chains around it before it does.
Compute is the current bet. Starcloud, SpaceX, and now many others are building toward orbital data centers on the premise that AI’s energy demand is growing faster than the grid can respond. If you can’t get power to where the chips are, move the chips to where the power is.
That framing is what makes compute something new: it’s not just another satellite application, it’s a thesis that energy in space is more valuable than energy on the ground. The harder questions are operational and engineering. GPU generations turn over every few years, thermal management remains a meaningful challenge, and servicing hardware in orbit introduces costs that terrestrial data centers don’t face. The economics depend on whether those costs and engineering decisions can be brought down to something manageable.
Electricity is the bigger bet. It takes that same thesis and follows it to a different conclusion. Rather than relocating demand to where the energy is, it delivers the energy to where demand already exists, which is directly to the terrestrial grid. Communications, compute, and manufacturing all convert orbital energy into a distinct product before it reaches a customer. Space solar energy skips that conversion.
Compute is one market. Electricity is every market.
Marc Berte, Founder and CEO of Overview Energy
At Overview Energy, we’re designing our system to use existing utility-scale solar projects as the receiver so they can generate at night. Meta has already signed a capacity reservation agreement for our first power, an early signal that the demand side of this market is real. Energy flows into grids that are already built rather than requiring entirely new sites, which matters in a world where companies are deploying gas turbines just to get faster access to power.
The remaining challenges are manufacturing and scale so energy from orbit gets to the ground at a cost that competes with other sources. The reason it’s worth solving is the size of what’s on the other side. Compute is one market. Electricity is every market.
Why energy is the defining question
Data centers will consume between 9-17% of U.S. electricity by 2030. That number keeps getting revised upward, and the pressure that creates is real. But electricity also powers the other over 80%, and the industrial loads, transportation networks, and desalination systems that will define the next wave of energy demand don’t fit neatly into the compute framing at all.
The GPS analogy is useful here. The system wasn’t designed for financial settlement networks or the synchronized timing that the modern internet depends on. It became foundational to those things because positioning, navigation, and timing (PNT) turned out to be a general input, not a specific product. Continuous orbital energy has the same character. At a sufficiently low price, people will use all you can produce. There’s no natural ceiling on that demand the way there is on any particular application.
The orbital energy economy is still being defined. Communications established the first major commercial model. Compute and space solar energy are making the case for what comes next, approaching the same underlying resource from different points in the value chain. Manufacturing addresses a different set of opportunities beyond that. None of these are mutually exclusive, and the capital and launch capacity exist to pursue them in parallel.
What matters most at this stage is whether the approaches being funded are designed around the actual hard constraints or whether those are being treated as problems to solve later. The orbital energy economy will be built by companies that took those constraints seriously from the beginning.
Related