The Filing That Started the Pitch

The centerpiece of James Altucher’s Musk Millionaire presentation is a document he calls a “hushed FCC filing” — something Elon Musk filed through SpaceX that Altucher says contains the blueprint for “the largest wealth creation event in history,” with 1,806,000 new “Musk Millionaires” minted starting September 25, 2026.

The dramatic framing is marketing, but the underlying substance is worth separating from the theater. The filing, as described in the presentation, lays out a plan to send one million satellites into orbit — a hundred times the roughly 10,000 SpaceX has launched to date. And critically, these are not Starlink-style internet satellites. Altucher describes them as “full blown AI data centers” loaded with advanced AI chip technology, each one drawing power directly from the sun.

The FCC’s role makes sense once you think about it. In the United States, the Federal Communications Commission regulates orbital spectrum use and satellite deployments. Any company planning a constellation of this scale has to file for it. That means the outline of Elon’s space ambitions, however quiet the rollout, exists in public regulatory paperwork — which is exactly the kind of primary-source artifact a newsletter analyst can build a thesis around.

The scale deserves a moment. Going from 10,000 satellites to one million is a 100X expansion of the largest private satellite array ever built. As Altucher puts it, “that number is so huge, it will change the economy forever.”

But the job description is what changes the investment story. Starlink beams internet to cell phones and takes pictures of the Earth. The new constellation’s purpose, per the presentation, is “giving mankind the ability to run full scale Artificial Intelligence from outer space.” SpaceX’s own quoted line from the filing: “Within a few years the lowest cost to generate AI [power] will be in space.”

The first of these orbital data centers already exists, according to the pitch. It’s called AI1, and with a wingspan of 230 feet it’s wider than a Boeing 747. Altucher says the xAI-SpaceX merger — the largest private merger in history — happened days after the million-satellite plan was filed, giving Elon both the rockets and the most advanced AI to put on them.

Why Orbit? Three Reasons

The presentation’s argument for moving AI compute to space rests on the physics:

Power. Solar panels in orbit work “up to eight times better than on Earth,” with uninterrupted sunlight more than 99% of the time. No clouds, no weather, no night.

Cooling. Data centers on Earth need enormous cooling infrastructure. Space “already sits at 250 degrees below zero.” Cooling is free.

Permission. No zoning boards, no town halls, no permitting fights, no communities banning the project. As the pitch puts it, “no town that can tell him no.”

Compare that to the terrestrial situation: Amazon throwing $20 billion at power including nuclear, Microsoft trying to restart old nuclear plants, and Meta funding seven gas plants for its biggest data center. Everyone else is scrambling for watts on the ground. Elon, Altucher argues, “went where the power is limitless.”

The Land Rush Nobody Noticed

There’s a scarcity mechanic buried in the filing thesis that gives it urgency beyond hype. Out of the three orbital zones — low, medium, and high Earth orbit — low Earth orbit is the most valuable for AI because proximity to Earth means faster data transmission. But scientists warn only about 100,000 satellites can fit in low Earth orbit before collision and debris risk becomes unmanageable.

The presentation’s read: “Elon is rushing to claim the rest of those 100,000 spots for himself, before anyone else.” That framing explains why Elon would file for capacity now rather than building gradually, and why rivals’ slower efforts matter. NVIDIA’s answer is a startup called Starcloud, whose big achievement so far is a single refrigerator-sized satellite. Google announced Project Suncatcher, two test satellites that won’t launch until 2027 “if they launch at all.” Blue Origin filed its own version, Project Sunrise, but the presentation notes the FAA grounded the program after flight failures, and Bezos himself admits it could take twenty years.

The Catch

One million satellites is a plan, not a deployment. A regulatory filing describes intent and requests permission; it doesn’t guarantee launches, timetables, or economics. Altucher himself concedes the engineering skepticism that greets every Musk ambition — the presentation opens with a montage of experts declaring Tesla, SpaceX, and Starlink doomed, all of whom were wrong. That pattern is genuinely relevant context. But it’s also true that Musk projects historically run years behind initial timelines.

The investable question isn’t whether one million satellites reach orbit on schedule. It’s whether the direction is right: is the future of AI compute partly orbital? For that angle, the transmission layer — the unnamed “AI MasterKey” supplier that supposedly links these data centers to Earth — is where Altucher directs investors. We cover the full thesis, the free Intel ticker, and the $49 offer in our Musk Millionaire review.

The energy side of the story is its own topic — the 840,000-homes-per-data-center math and what it means for power investing — which we break down in our AI energy bottleneck article. And for a different analyst’s take on AI infrastructure bottlenecks, see how Jason Bodner’s photonics thesis reaches a similar picks-and-shovels conclusion from the copper side.

NewsletterVetter is an independent publication. We receive compensation from some of the services we review through affiliate links. Nothing on this site is investment advice. Always do your own research.