SpaceX Just Made the Internet 95% Cheaper: What the Market Is Missing

One number decides whether SpaceX is the cheapest IPO in history or the most insane one. And most people are going to get it wrong.

In June 2026, SpaceX went public. Ticker SPCX on the Nasdaq. Price: $135 a share. Company value: about $1.77 trillion. They raised roughly $85 billion — more than two and a half times Saudi Aramco's record IPO. BlackRock reportedly wanted at least $5 billion of the deal. Index funds are lining up to own a sliver whether you personally click buy or not.

Then Morningstar ran a discounted cash flow — the boring textbook method — and landed at about $780 billion. Half the IPO price. At a $2 trillion valuation you are paying something like 107 times annual revenue. A hot tech name might trade at 30 times sales. Normal healthy growth sits closer to 10 or 15. The bear case writes itself: hype machine, money grab, bubble with a rocket painted on the side.

So why did Oppenheimer slap an outperform rating with a $190 target — call it $2.5 trillion? Why is BlackRock elbowing in?

I do not think they are just drunk on Elon. What Morningstar's DCF cannot see is a cost that has not fully hit the cash flow statement yet: the cost of putting internet bandwidth into orbit, measured in dollars per megabit per second.

That sounds nerdy. It is also the unit cost of SpaceX's product. Megabits per second is the firehose your ISP sells you. Dollars per Mbps to orbit asks a simpler question: how much does it cost to launch one unit of internet speed into space so it can beam back down to a customer? The number that decides everything is the cost to deploy.

SpaceX is a telecom company with rockets

People still call SpaceX the rocket company. At this point it is a telecom company that owns the cheapest rockets on Earth, with an AI bet stacked on top. In Q1 2026, SpaceX did about $4.69 billion in revenue. Roughly $3.26 billion of that — about 69% — came from connectivity. Starlink is the cash cow. The rockets exist largely to feed the satellites.

At the IPO filing they had about 10 million subscribers, a number that doubled in a year, with a path toward something like 17 million by the end of 2026. People pay roughly $80 a month. In May 2026 they raised prices by up to $10 in some markets. You do not raise prices when customers have easy alternatives. You raise them when you are the only game in town for a lot of those customers.

The old number: Falcon 9

Falcon 9 already crushed historical launch costs by something like 85%. A Falcon 9 can haul about 17 tons to orbit. At SpaceX's internal cost of roughly a thousand dollars per kilogram, that is about $17 million to fill the rocket and fly it. For that money you deploy a batch of Starlink satellites that add about 2,600 gigabits per second of capacity.

Do the division. $17 million divided by 2,600 Gbps works out to about $6.55 per megabit per second. That was already the cheapest number in the industry. A lot of people looked at $6.55 and treated it like a floor.

Then Starship showed up.

The new number: Starship + V3

Starship is not a slightly bigger Falcon. It is a fully reusable two-stage monster built to carry about 100 tons to orbit and eventually get caught and flown again the same day. Where Falcon 9 carries 17 tons, Starship carries almost six times that. Cost per kilogram drops toward under $185 versus Falcon's roughly $1,000. A single Starship launch costs about $18.5 million — barely more than Falcon 9 — but the payload is a different world.

You are not just launching more of the old birds. You launch Starlink V3. V3 carries about four times more bandwidth per kilogram than the V2 minis flying today. Rocket mass up almost 6x. Bandwidth per kilogram up about 4x. Those multipliers stack.

One Starship launch at $18.5 million deploys about 60 V3 satellites. Those 60 birds add up to about 61,000 gigabits per second. Falcon 9 at $17 million buys you 2,600 Gbps. Starship at $18.5 million — basically the same launch bill — buys you 61,000 Gbps. More than 20 times the internet for the same money. One Starship replaces about 23 Falcon 9 launches worth of capacity.

Run the new division. $18.5 million divided by 61,000 Gbps lands near 30 cents per megabit per second.

Thirty cents.

We went from $6.55 down to about $0.30. That is a 95% reduction. SpaceX's single biggest cost — putting the product in orbit — just fell by ninety-five percent in one generation of hardware. Elon saw the math Rob Maurer posted and replied "interesting analysis." That is the tell.

What a 95% cost collapse actually does

Imagine your company's cost of goods fell 95% overnight. Two things happen, and both are enormous.

First, margins explode. You are still charging customers something like $80 a month while it costs a fraction of what it used to serve them. Second — and this is the bigger one — things that were impossible at the old cost become trivial at the new one. Markets that made no sense at $6.55 a unit make perfect sense at 30 cents.

This is Wright's Law in public. Model T. Solar panels. Shipping containers. In 1956 Malcolm McLean put 58 metal boxes on a converted oil tanker and loading cost fell from $5.86 a ton to about 16 cents. That collapse did not just make shipping cheaper. It made globalization economical. SpaceX did not make satellite internet a little better. It made a whole category of stuff economical that simply was not before.

Three conditions for the bull case

That cost drop is what bridges Morningstar's $780 billion and the $1.77 trillion IPO. Three things have to be true.

Condition one: Starlink becomes the default where fiber cannot go. When deploy cost falls 95%, you can cut into markets nobody else can afford and flood capacity into markets you already own so service does not melt as you add millions of users. Amazon's Kuiper was still barely off the ground at IPO time. Fiber wins on latency — roughly 11 to 14 milliseconds versus Starlink's 25 to 60 — but fiber cannot reach a cargo ship in the Pacific, a farm in Montana, a village in the Congo, or an airplane at 30,000 feet. Two-thirds of Earth's surface is never getting a cable. Direct-to-cell makes that even more lopsided. The price hike already happened. About 10.3 million subscribers and a 63% connectivity margin are already on the S-1.

Condition two: those margins fund the science fiction. Quilty has forecast Starlink around $20 billion in revenue in 2026, roughly $14 billion in EBITDA, and about $8.1 billion in free cash flow. That cash pays for Starship, the next satellites, and the crazier bets. The cash cow funds the moonshot. Also already visible in the numbers.

Condition three: orbital AI compute has to become real. This is the piece that bridges $780 billion to $1.7 trillion and beyond. Right before the IPO, SpaceX unveiled AI One — a satellite flying the equivalent of a full AI server rack at about 150 kilowatts peak, with a 70-meter wingspan wider than a 747. They have filed toward eventually building up to a million of them out of a new factory in Bastrop, Texas, with volume targeted by the end of 2027. In orbit you get continuous solar, dump heat into the cold of space, and skip terrestrial power permits and cooling towers. Power and cooling — the two things strangling Earthside AI data centers — get a lot freer upstairs. AI One rides the same Starlink V3 platform. Rockets feed satellites. Satellites feed internet. The internet platform feeds data centers in space.

The bears are not wrong about today's math. AI chips obsolete every two to three years. At old launch costs, refreshing orbital compute forever is brutal. But the break-even for orbital data centers sits somewhere around $500 to $1,000 per kilogram to launch. Starship is targeting under $185 to $250. The same 95% collapse that took satellite internet from $6.55 to $0.30 is what drags orbital compute from insane to inevitable. Starlink is the proof the cost curve pays out. Orbital AI is the trillion-dollar payoff the same curve unlocks next. Same bet. Same master variable.

The only question that matters

Two of the three conditions are already demonstrably true. The third is a physics and engineering bet on the most capable launch team on Earth.

If you are even thinking about the stock, ask yourself one question: do you believe Starship gets under $250 a kilogram and stays there? If yes, the rest of the bull case follows. If no, treat $1.7 trillion exactly the way the bears describe it and size accordingly.

And even if you never buy a share, you live downstream of that curve. Cheaper space bandwidth means cheaper internet for the two-thirds of the planet that never had it. Cheaper launch means the space economy stops being a government-sized project and starts being a place where real businesses get built.

The market is arguing about a DCF. The real fight is about 30 cents.

Check the video here.