Musk’s hodge-podge

The core problem: Wall Street is pricing SpaceX’s 800 MHz spectrum purchase from Grain Management as a telecom bet: either leverage for an MVNO deal or a fourth national carrier. Both readings judge Starlink Mobile by carrier economics. But Starlink Mobile is a revenue business and a paid-for prototype of the local network a Mars settlement will need, and the “inefficient” satellite-plus-small-cell architecture that analysts dismiss on Earth is the only architecture that’ll work on The Red Planet.

On October 8, SpaceX agreed to buy Grain Management’s entire nationwide portfolio of 800 MHz spectrum, up to 14 MHz of paired low-band, which Grain bought from T-Mobile in August. The deal is pending FCC approval, but the next morning Verizon had its worst trading day since 2002; T-Mobile and AT&T weren’t far behind. By lunch, the industry had its villain.

Within 48 hours, the analyst community had sorted itself into two camps. Camp One, led by analyst Craig Moffett, says this is leverage: SpaceX wants an MVNO deal on good terms, and owning spectrum is how you make a carrier pick up the phone. Camp Two, led by Pivotal, says this is a credible fourth national carrier in the making, stacked on top of the $17 billion of EchoStar mid-band the FCC approved in May.

The skeptics in Camp One have the better spreadsheet. Moffett’s colleague Julie Zhu points out the big three hold 14 to 15 times more spectrum, even after this deal. Morningstar called a terrestrial buildout “a poor use of capital.” New Street dismissed the whole satellite-plus-ground-cell concept as a “hodge-podge.”

Here’s my problem with both camps. They’re grading SpaceX as a telecom company. Musk does not run SpaceX as a telecom company. He’s said it in plain English for years, and almost nobody in our industry is listening.

Starlink exists to pay for Mars

Don’t take my word for it. In 2015, unveiling the satellite business in Seattle, Musk described it as a long-term revenue source for the company that would help fund a city on Mars. He has repeated it ever since, most plainly on X: “Starlink is how we are paying for humanity to get to Mars.” Need more proof? Look at the illustration on Starlink routers. It’s the Earth-Mars transfer orbit. And the SpaceX board agrees with him. Part of his billion-share grant vests when a Mars colony reaches one million people.

Starlink is only the most visible piece. For twenty years, Musk’s companies have looked like a hodge-podge of their own—cars, tunnels, rockets, satellites—run by one person. Look at the whole portfolio through a Mars lens and every company snaps into place. Tesla is batteries, solar, and electric vehicles for a planet with no gas stations. The Boring Company is tunneling for a planet where habitats will have to be buried against radiation. SpaceX is the ride. Starlink is the money and the network. He has been assembling a Mars supply chain in public, and we keep reviewing each company as if it were a standalone business.

Starlink is SpaceX’s only profitable unit. Every dollar of Starlink margin is a dollar of Mars budget.

So when SpaceX moves from satellite broadband into mobile, the question I ask is: What does mobile do for Mars?

The revenue answer is obvious. SpaceX’s own IPO prospectus sizes the Starlink Mobile opportunity at $740 billion. Mobile is the biggest adjacent pool of cash on the planet, and SpaceX needs cash.

The architecture answer is the one nobody is writing about.

A Mars network with an Earth address

Read SpaceX’s own description of what it plans to build with the 800 MHz band: one network that blends satellites with ground cells and reaches ordinary phones indoors, outdoors, and in dead zones. The point is propagation. Low-band gets through concrete and energy-efficient glass. That’s why SpaceX called it one of the last technical gaps for Starlink Mobile.

Now think about Mars.

Everything is a dead zone. Everyone lives inside, because the outside is lethal. There are no towers, and there will never be a buildout crew. A settlement gets connectivity exactly one way: satellites overhead, plus small cells the settlers deploy themselves, running on radio software the operator controls end to end.

That is the hodge-podge. The architecture Wall Street calls inefficient on Earth is the only architecture that exists on Mars. The decision to own the radio software, run the network rather than rent one, and make satellite-to-ground-cell handoffs seamless looks like a stubborn capital allocation choice if you think the goal is beating Verizon. It looks like the minimum viable design if you think the goal is a second planet.

Rural America is the test range

Having the right design is one thing. Knowing it works is another, and that’s what Earth customers are for.

You cannot fund a Mars network by selling it to Mars. There are no customers there yet. And SpaceX has no sponsor to fall back on. The company pitched Marslink to NASA: satellites in Mars orbit, laser crosslinks, 4-plus megabits between the planets, and surface coverage. But on September 1, Jeff Bezos’ Blue Origin got NASA’s Mars Telecommunications Network contract, worth up to roughly $700 million. If SpaceX wants its own network on Mars, it has to pay for it. Rural Texans with bad coverage are the way.

Those same subscribers are also the test market. Everybody SpaceX signs up on Earth is a live trial of the hard problems: handoffs between orbit and ground, the radio access software, and the operational discipline of actually running a mobile network instead of partnering for one. SpaceX is hiring engineers specifically to test and deploy radio-network software on the satellite constellation. That is a company building an operator, with paying customers absorbing the learning curve.

This is the Starlink playbook, repeated. Earth customers funded the constellation. The constellation became the template for everything after. This foray into mobile is the same move, one rung up.

And SpaceX told us so a couple days before the deal. On October 6, SpaceXAI President Michael Nicolls said the company aims to form an “interplanetary mesh of connectivity.” Musk followed by saying SpaceX would connect the whole solar system.

Coverage treated these as separate headlines. I read them as one story.

Where this interpretation stops

I want to be honest about the limits, because the honest version of this argument is stronger than the sweeping one.

The 800 MHz licenses themselves do not go to Mars. Their value comes from FCC rules and US spectrum scarcity, not from physics that transfers to another planet. The hardest Mars problem is the link back to Earth: three to 22 minutes of latency, plus total outages when the Sun sits between the planets during conjunction. That takes lasers and deep-space work, and a phone network does not solve it.

So the claim is narrower, and better for it: Earth mobile funds and tests the local network on Mars. The interplanetary link is its own program. To be clear, this is my reading of the moves, not anything SpaceX has stated as strategy.

Why the “bad economics” critique misses

Even the narrow version answers the strongest objection. Zhu’s sharpest argument is that wireless cannot match the one-year payback SpaceX could get by putting the same capital into AI compute. For a carrier, that’s the right test. But it’s wrong for a company with a second objective.

If the Starlink network pays for itself on Earth and also teaches SpaceX how to run a phone network on another planet before the first settlers land, the return isn’t captured by subscriber margin alone. The return includes not having to do the hardest engineering of the century under a deadline with no fallback.

There is a general rule hiding in here, and it applies well beyond SpaceX. When a company enters your industry with a goal that lives outside it, your industry’s payback math will misprice every move that company makes. We watched this with Apple and the handset business. We’re watching it now with spectrum. The spreadsheet is measuring the wrong objective.

The villain is the hero

What I believe most has nothing to do with spreadsheets.

Our industry has spent the past month casting Musk as the threat. He’s the guy with the spectrum, the guy who wiped billions off carrier market caps in a morning. I understand the instinct, and I think it is exactly backwards.

Mars is a hundred-year problem, and Musk is treating it that way. Look at the roadblocks that have shown up in the past six weeks alone. The carriers hold 15 times his spectrum. The FCC approval process is out of his hands. The NASA contract went to his rival. Each one would be a reason for a normal company to pause. His solution has been to buy more spectrum.

That is what first-principles thinking looks like when it has a deadline measured in generations. He does not ask whether the industry structure allows for a fourth carrier. He asks what a settlement on another planet needs, works backward to what has to exist on this one, and then does not sleep until it does. You can find that exhausting or inspiring, and I know plenty of people in our industry who find it both. I’ll tell you where I land: this attitude is exactly what is going to give our planet options, and I do not see anything that stops him.

Every other player in this story is protecting something, and they’ve said so out loud. In September, AT&T’s CEO told investors that direct-to-cell satellite is solving the last 2% of coverage, and T-Mobile’s outgoing CFO said he wasn’t worried about Musk’s cellular ambitions. The week before the deal, the three major US carriers announced a joint venture to expand satellite coverage in underserved areas, and the largest satellite operator on Earth was not invited. Wall Street is protecting a valuation model. Musk is the only one trying to build something that does not exist yet, and he’s doing it with his own capital and his own rockets. In any story I know, that’s the hero.

I think the industry should ask itself a question. Every analyst note I’ve read treats Musk as a threat to be modeled. No one has asked the question I’d be asking if I ran a carrier: how do we help?

Musk needs backhaul, towers, fiber, roaming, and operational expertise at a scale nobody has run across two radio domains before. Those are assets you already own. Whether telco figures out how to be part of the Mars story is genuinely open. The man building the network has already decided what his goals are. The rest of us still get to choose what role we’ll play in this story.

Recent Posts

Get my FREE insider newsletter, delivered every two weeks, with curated content to help telco execs across the globe move to the public cloud.