Orbital Data Centers: The Space AI Thesis, Explained

Orbital data centers have moved from science fiction to FCC filings in about two years, and the current newsletter cycle is racing to keep up. The pitch driving this round of interest is Dylan Jovine’s “Elon’s 100X AI Stock” from Behind the Markets, which builds an investment case on orbital AI compute and then pivots to an unusual place: the chip supplier behind it. Here is the thesis explained layer by layer, with the players, the numbers, and an honest verdict on the timeline.

The Problem That Creates the Opportunity

The setup is the AI power squeeze, and it is verifiable. “The four biggest names in tech, Microsoft, Amazon, Google and Meta, are pouring a combined $725 billion into AI infrastructure in 2026 alone,” with 16 gigawatts of new U.S. data center capacity opening this year. The constraint is electricity. Morgan Stanley is quoted projecting “the U.S. will be short 49 gigawatts of electricity for data centers by 2028.” Google, the pitch says, “now waits up to 12 years to get a new data center hooked up to the grid.” Local resistance is real: Denver, Oklahoma City, and Tulsa have paused new data centers, and “Maine’s legislature voted to ban them outright.”

The Orbital Answer

Jovine’s presentation of Musk’s solution is quotable: “Because Elon figured this out years before anyone else. The answer to this problem is NOT on Earth… Build it up there.” Two Musk statements anchor it: “The cheapest place to put AI will be in space within 36 months or less,” and “Space-based AI is obviously the only way to scale.”

The physics claims are worth stating because they check out on their face. “A solar panel in space generates roughly eight times the energy of the same panel on the ground,” thanks to uninterrupted sun and no atmosphere. “Space is 250 degrees below zero, and it never warms up,” which provides a cold sink for radiator-based cooling (though heat can only leave by radiation, through large panels, since a vacuum carries nothing away by convection). The concrete artifact is AI1: “Its solar wings stretch 230 feet from tip to tip. That’s longer than the wingspan of a Boeing 747.” And the scale request, from a January 30, 2026 filing the pitch cites: “SpaceX asks the FCC for permission to launch one million satellites,” versus “about 10,000 satellites in orbit” today.

The Players, Verified

The orbital data center race is a genuine multi-front story, and the pitch names the field accurately: the Nvidia-backed Starcloud has trained the first AI model in space, a proof of concept that moved the thesis from paper to hardware; Bezos is “pouring $1 billion of his own money into Blue Origin every year”; Google’s Project Suncatcher is underway; Meta has a space-solar deal in motion; and Musk’s AI1 plus the one-million-satellite FCC request represent the largest stated ambition. Musk’s epigraph in the pitch is the market-size claim: “We have identified the largest actionable total addressable market in human history,” with SpaceX’s own materials citing a $28.5 trillion TAM, which the pitch notes is more than eight times bigger than everything Musk has built combined.

Is the 36-month timeline credible? Honest answer: credible as direction, aggressive as schedule. The grid bottleneck deepens every quarter, which pushes demand toward orbit regardless. But first-generation orbital facilities are experiments, and the gap between 10,000 satellites and one million is the entire bet.

The Investment Twist: The Supplier

Here is where this campaign differs from a generic space-AI pitch. Jovine does not point at the satellite operator. His case is that “There is one AI company that Elon needs for his massive satellite fleet… There’s just one thing he can’t build himself. The chip that runs them.” The licensing logic: “Elon doesn’t own that chip. He licenses it. That means, every time a Tesla rolls off the line… Every time SpaceX launches an AI1 satellite… This company gets paid.” Jovine says the relationship dates to 2012, and that a 100X satellite fleet means “a 100 times bigger bill to this company.”

The company is unnamed; its name and ticker sit inside the first research report of a $49 subscription. The clues point toward a chip-IP licensor on a royalty-per-unit model. That identification is unconfirmed, and we hold it as a fingerprint, not a fact. The royalty business model itself is real and well documented, and the switching-cost argument, that leaving an architecture means “rewriting every line of software in every Tesla and every satellite he owns,” is economically sound.

The Calendar Device

The pitch’s urgency hangs on December 8th, the final SpaceX lockup tranche: “the last block of roughly 800 million insider shares comes loose, all at once. That’s more than $100 billion at the IPO price.” Jovine’s thesis is that freed insider capital flows toward the suppliers insiders know best, hedged with “I believe” and “in my opinion.” We covered that event in detail in our SpaceX lockup expiration analysis, and we have tracked orbital data centers as an investment theme across several campaigns now.

Our Honest Read

This is one of the better-built pitches in the genre: real bottleneck, real race, real business model, event-driven urgency tied to a genuine calendar date, and a presenter with dated, checkable calls (Rocket Lab under $4, Palantir at $7, Micron at $54). The unverifiable core is the unnamed supplier and every specific claim about it. If the thesis interests you, the full presentation is here, backed by a six-month full-refund guarantee on the $49 subscription.

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.