A Five-Month-Old Startup Just Asked the Government for Permission to Launch 100,000 Satellites. The Entire History of Spaceflight Has Only Put About 15,000 in Orbit.

A company that didn't exist six months ago just filed paperwork with the US government to launch 100,000 satellites.

Not 100. Not 1,000. One hundred thousand. Into orbit. From a startup called Orbital that was founded roughly five months ago.

The Federal Communications Commission received the filing last week. Orbital's stated goal: deploy up to 100,000 data center satellites and deliver 10 gigawatts of computing power from space.

Read that again slowly.

Why Would Anyone Put a Data Center in Space?

The AI boom has an energy problem nobody advertises. A single large AI training run can consume as much electricity as 500 homes use in a year. Multiply that across hundreds of companies racing to build the next model — and you've got a grid crisis, a land crisis, and a cooling crisis all arriving at once.

On Earth, building a data center means acquiring land, running power lines, pumping enormous volumes of water to keep the servers from melting themselves. The costs are real and the timelines are long.

In orbit, the physics change completely.

The concept of space-based solar power and orbital computing has existed in research papers for decades. What's changed recently is launch economics — SpaceX's Falcon 9 reduced the cost of reaching orbit by over 90% compared to the Space Shuttle era. What was theoretically interesting is now, at least on paper, economically arguable.

Orbital is betting the argument holds. Whether it does is a different question entirely.

One Hundred Thousand. Let That Number Land.

The current largest satellite constellation ever assembled is Starlink. SpaceX spent years and an estimated $10 billion-plus getting roughly 6,000 satellites into orbit. That took the world's most prolific launch company, with its own rockets, operating at unprecedented cadence.

Orbital, at five months old, has filed for 100,000.

Every satellite ever launched by the United States, Russia, China, Europe, Japan, and India — across the entire history of spaceflight, from Sputnik in 1957 to this morning — hasn't reached 100,000 tracked objects. Orbital is asking the FCC to allow it to more than sextuple the entire orbital infrastructure of human civilization.

What the FCC Filing Actually Means — and What It Doesn't

Here's the part most coverage buries: this is paperwork, not hardware.

Filing with the FCC for orbital spectrum and slot authority is how companies stake a regulatory claim. It's the necessary first step in any satellite constellation plan — not the last. Amazon filed for its Kuiper constellation years before a single satellite launched. OneWeb filed, went bankrupt, was acquired by the UK government, and only recently got satellites flying. Filing is cheap. Building is not.

Once Orbital's application enters formal review, other satellite operators — SpaceX, Amazon, Telesat, international operators — can and likely will file objections. Regulators will scrutinize technical plans, debris mitigation commitments, orbital altitude, and coordination with existing users. At some point, Orbital will need to demonstrate it has the capital, the launch agreements, and the actual spacecraft to follow through.

The Debris Problem Nobody Can Ignore

Any serious conversation about 100,000 new satellites must address orbital congestion. Astronomers are already frustrated by the streaks Starlink's 6,000 satellites leave across telescope images. Scientists studying Kessler syndrome — the theoretical runaway collision cascade that could make low Earth orbit unusable — are already concerned about existing megaconstellations.

One hundred thousand additional objects would be an unprecedented pressure test for orbital debris management, active collision avoidance, and end-of-life deorbit requirements. Even a 1% failure-to-deorbit rate produces 1,000 uncontrolled objects. The global debris tracking network currently struggles to reliably track anything smaller than 10 cm. The math gets uncomfortable quickly.

Why This Story Matters Even If Orbital Never Launches a Single Satellite

Even an unrealized FCC filing is a data point. It reveals where ambitious people — or at least well-lawyered ones — believe the next infrastructure wave is heading. The AI energy demand crisis is real and documented. The physics argument for orbital solar power is sound. The launch cost trajectory is favorable.

Whether Orbital specifically executes on this plan is almost secondary. The fact that someone filed it — publicly, on the record, in a document regulators must respond to — means orbital data centers are now a live conversation in US telecommunications policy. Other companies, including ones with more capital and operating hardware, are watching how regulators respond.

The sky tonight looks the same as always. But on paper, it just got a lot more contested. Follow the SkyLens feed as the FCC review unfolds — the responses from existing operators over the next few months will tell us far more about whether orbital computing is a serious infrastructure play or the most ambitious cold call in regulatory history.

You can see every real satellite currently in orbit — all 15,891 of them — on the SkyLens live tracker right now. If Orbital ever launches even a fraction of what it's proposed, that number will look quaint.