Can Elon Musk Really Put ONE MILLION Data Center Satellites in Orbit?

Now that you've read the headline, I'll show you the numbers, and you'll probably say, "OH NO WAY!"

But indulge me, won't you?

Ars Technica's senior space guy Eric Berger looked at how much it might cost SpaceX to fulfill Elon Musk's stated goal of putting ONE MILLION DATA CENTER satellites into Low Earth Orbit (LEO) for the company's artificial intelligence division, xAI.

I don't want to bother you with all the background math — you can click the link for the full MEGO effect, should you desire — but you do need to see the final totals for putting all those birds in orbit:

• Optimistic: $350 billion in launch costs; $1 million per satellite; all-in cost of $1.45 trillion
• Neutral: $1.85 trillion in launch costs; $1.5 million per satellite; all-in cost of $3.45 trillion
• Pessimistic: $7.7 trillion in launch costs; $2.0 million per satellite; all-in cost of $9.8 trillion

The median scenario is $3.45 trillion dollars, or roughly 70% of what the U.S. spent waging World War II.

Honestly, I felt the need for a shot of tequila after going through those figures. "Daunting" doesn't even begin to cover the expense and complexity of such lofty goals. "Unimaginable" perhaps comes closer.

But things become less daunting and more imaginable when viewed from the proper perspective.

Enter, stage right: Your Strangely Optimistic Friendly Neighborhood VodkaPundit.

The question isn't, "Can Elon Musk really put one million data center satellites in orbit?" The correct question is, "Can he put enough data center satellites in orbit at a low enough cost to generate positive cashflow?"

It's all about the money, honey — because once you're making more than you spend, the sky's the limit.

Literally, in this case.

And while one can say for sure whether Musk can make the economics work, this is an area where SpaceX enjoys hard-won expertise. You might have heard of a little space-based internet service provider called Starlink.

The first operational batch of 60 satellites went up in May 2019, and Starlink beta service went live near the end of 2020 with fewer than 1,000 in orbit. By the end of 2024, with around 6,400 satellites, Starlink turned cash flow positive. 

That was the key metric, not whether Starlink had achieved its longterm goal of having a constellation of 42,000 operational satellites. 

“The key takeaway I want everybody to walk away with is, if SpaceX was building the Starlink system to pay for a Mars colony, we’ve got evidence that the company will generate the type of free cash flows from the business that could pay for said endeavor," Quilty Space president Chris Quilty said at the time.

Today the end goal has changed — the Moon and data centers, not Mars — but the means remain the same: using orbital operations-at-scale to turn a profit doing the previously impossible.

It took roughly 6.5 years for SpaceX to launch the 10,000th Starlink bird, and for the division to start raking in serious profits. So I tend to think of roughly 10,000 satellites — an unimaginable number not that many years ago — as the number required for sustained operations at a return worthy of the effort. 

And Another Thing: Naysayers were all doom-and-gloom that SpaceX lost money in 2025, entirely on data center buildouts. But now the company leases server space to companies like Google and Anthropic for about $2 billion a month. xAI isn't quite cash flow positive, but it's closing in.

Assuming SpaceX makes Starship work, the launch economics improve considerably. If the company doesn't make it work, the entire project falls apart.

Also, Starlink enjoys a built-in audience, as it were, of people in rural or underdeveloped areas who had no other means of getting affordable high-speed internet service. Plus previously unfulfilled demand from airlines, cruise ships, and the Department of Defense. There's no similar built-in demand for orbital data, so it's my assumption that xAI will have to compete with terrestrial data centers on price.

If Musk can overcome the serious technical and financial challenges to achieve positive cash flow, then SpaceX can continue putting data center satellites into LEO until they've hit one million... or 100,000, or 10,000. Or whatever number the market will allow.

Recommended: Come Meet Britain's Next Worst Prime Minister Ever

I can't even put one satellite in orbit. But I can make you one fantastic offer.

PJ Media VIP members get tons of exclusive goodies, including podcasts and video live chats with your favorite writers. You can support alternative conservative news and save 60% with promo code FIGHT.

Join today.