SpaceX Slashes the Cost of Orbital Internet by 95 Percent as Its $1.77 Trillion IPO Prices

By pairing Starship's freight-elevator payload with the higher-bandwidth Starlink V3 satellite, SpaceX drove the cost of deploying capacity to orbit from $6.55 to about 30 cents per megabit per second, the single number that decides whether the largest IPO in history is cheap or insane.

The whole $1.77 trillion argument sitting on top of SpaceX right now rests on one boring, unsexy figure: the cost of putting a unit of internet speed into orbit, measured in dollars per megabit per second. Track that number and you can stop arguing about the stock price. It just fell 95 percent in a single generation of hardware, and once you see how, you understand why some of the most careful money on the planet is willing to pay double what any spreadsheet currently supports.

Key Takeaways

• SpaceX priced its IPO in June 2026 at $135 a share, roughly $1.77 trillion, and raised about $85 billion. That's more than two and a half times Saudi Aramco's $29 billion record from 2019, making it the largest IPO in the history of human capital markets.
• Morningstar's discounted cash flow analysis pegs the company at about $780 billion, roughly half the IPO valuation. At $2 trillion you're paying about 107 times annual revenue, when a hot tech name normally trades at 30.
• Deploying bandwidth to orbit fell from $6.55 per megabit per second on Falcon 9 to about 30 cents on Starship. Same launch cost, more than 20 times the internet.
• One Starship launch delivers 61,000 gigabits per second of capacity for $18.5 million, matching what it took 23 separate Falcon 9 flights to put up.
• Starlink is already the engine: $3.26 billion of SpaceX's $4.69 billion first-quarter 2026 revenue, about 69 percent, came from connectivity, running a 63 percent adjusted EBITDA margin.
• With 10.3 million subscribers and a price increase of up to $10 a month pushed through in May 2026, SpaceX is behaving like the only game in town, because for most of its customers it is.
• SpaceX unveiled the AI-1 satellite, a 150-kilowatt orbital compute rack with a 70-meter wingspan, and has filed to build up to one million of them in Bastrop, Texas, with volume by the end of 2027.
• Orbital data centers break even somewhere around $500 to $1,000 per kilogram to launch. Starship is targeting under $185 to $250. The same curve that already worked for internet is what makes compute in space go from insane to inevitable.

---

The One Number the Whole Valuation Rests On

Most people still think of SpaceX as the rocket company. At this point that's backwards. It's a telecom company that happens to own the cheapest rockets on Earth, with an AI company bolted onto the side. In the first quarter of 2026 it did about $4.69 billion in revenue, and roughly $3.26 billion of that, call it 69 percent, came from Starlink beaming internet down from space. The rockets exist to feed the satellites now.

So the cost that actually governs this business isn't fuel or factories in the abstract. It's dollars per megabit per second to orbit: how much it costs to launch one unit of internet speed up to where it can be sold back down as an $80-a-month subscription. That is SpaceX's single biggest cost of goods. Forget the subscriber count for a second. Forget even the revenue. The cost to deploy is the master variable, and it just collapsed.

The Bear Case Is Real, and You Should Hear It First

Let me steel-man the skeptics properly, because their strongest version is genuinely good. Morningstar, a buttoned-up research shop, ran a discounted cash flow on SpaceX. A DCF is the textbook way to value a business: project the cash it throws off for a decade or two, then discount it back to what it's worth today. It is the most conservative, most defensible method there is. Their number came out to about $780 billion.

The IPO priced the company at $1.77 trillion. So the careful analysis says the market is paying roughly double what the numbers justify. At a $2 trillion valuation you're paying about 107 times sales. A healthy, fast-growing tech company trades at 10 or 15 times revenue. A genuinely hot one, maybe 30. SpaceX is at 107. If you bought purely on the financials as they exist today, you are paying a price no spreadsheet on Earth currently supports. That case is legitimate, and anyone waving it away is selling you something.

The Falcon 9 Floor Everyone Thought Was Permanent

Here's the old world. For a decade the workhorse was Falcon 9, a reusable rocket that lands itself on a drone ship and already cut the cost of reaching orbit by something like 85 percent versus everything before it. It carries about 17 tons of payload. At SpaceX's internal cost of roughly $1,000 per kilogram, that's about $17 million to fill it up and launch.

For that $17 million you deploy a batch of Starlink satellites adding up to about 2,600 gigabits per second of bandwidth. Do the division and you land at about $6.55 per megabit per second. And $6.55 was already the cheapest number in the history of the world. The whole industry looked at it and decided that was the floor, about as good as physics would ever allow. Nobody thought space internet could get meaningfully cheaper.

How Two Multipliers Stacked Into a 95 Percent Collapse

Then Starship. It's the biggest flying object humans have ever built, a fully reusable two-stage machine designed to carry about 100 tons and eventually be caught out of the sky and flown again the same day, like an airplane. Where Falcon 9 hauls 17 tons, Starship hauls almost six times that. Because it's rapidly reusable, cost per kilogram drops to under $185, versus Falcon 9's roughly $1,000. Six times the mass for a fifth of the price per kilo.

Now stack a second multiplier on top. You don't launch the old satellites, you launch Starlink V3, which carries about four times more bandwidth per kilogram than the V2 minis flying today. Six times the mass, four times the internet per unit of that mass. Those two numbers multiply. One Starship launch at $18.5 million, barely more than a Falcon 9, deploys about 60 V3 satellites totaling 61,000 gigabits per second. That's more than 20 times the internet for essentially the same launch cost. One Starship replaces 23 Falcon 9 flights. Run the division and you're at about 30 cents per megabit per second, down from $6.55. A 95 percent cut in SpaceX's biggest single cost, in one generation of hardware.

Why a Cost Collapse Rewrites the Whole Map, Not Just the Margin

Picture running a business where your core input, the raw material of your entire product, suddenly dropped 95 percent. An airline where jet fuel fell 95 percent overnight. A steel mill where iron ore did the same. Two things happen, and the second one matters more. Your margin explodes, because you still charge the customer $80 a month while it costs a fraction to serve them. And markets that made no economic sense at the old cost suddenly make perfect sense at the new one.

This is the oldest pattern in technology, sometimes called Wright's law: every time you double cumulative production, cost falls a predictable amount. The Model T went from a rich man's toy to something a factory worker could afford, and it reshaped American cities. Solar fell something like 99 percent and became the cheapest electricity ever. The cleanest rhyme is the shipping container. In 1956 Malcom McLean put 58 metal boxes on a converted tanker and the cost of loading cargo fell from $5.86 a ton to about 16 cents. That collapse didn't make shipping a bit cheaper. It created globalization. Going from $6.55 to 30 cents does the same thing here. It doesn't improve satellite internet. It makes a whole category of things economical that simply weren't.

Two of the Three Conditions Are Already True

The bull case needs three things, and I want to be honest about which ones are real versus which are still a bet. Condition one: Starlink becomes the default winner of the internet. When your deploy cost drops 95 percent you can flood capacity into markets you own and drop prices into markets nobody else can afford to serve. Amazon's Kuiper is in enterprise beta only. AST SpaceMobile has around 45 satellites targeting intermittent US coverage. Fiber beats Starlink on latency, 11 to 14 milliseconds against 25 to 60, but you cannot run a cable to a cargo ship in the Pacific, a farm in rural Montana, a village in the Congo, or a plane at 30,000 feet. Two-thirds of the planet's surface is unfiberable. With direct-to-cell beaming to the phone already in your pocket, Starlink is the only company that can light all of it up. And notice: they raised prices up to $10 a month in May 2026. You don't do that when customers have options.

Condition two: the margins have to be defensible and fund everything else. Starlink runs a 63 percent adjusted EBITDA margin, a software-company number on what looks like a hardware business. The research firm Quilty forecasts about $20 billion in revenue for 2026, roughly $14 billion in EBITDA, and $8.1 billion in free cash flow. That's spendable money left over after building new satellites. I'm not asking you to believe a prediction here. The 10.3 million subscribers are real, the margin is on the S1, the price increase already happened. Two of three conditions are demonstrably true today.

The Third Condition: Data Centers in Orbit

Condition three is the one the bears can't model, and it's the only thing that bridges $780 billion to $1.77 trillion and beyond. Right before the IPO, SpaceX unveiled the AI-1 satellite, which flies the equivalent of a full rack of AI computers at 150 kilowatts of peak compute, about one of NVIDIA's top-end server racks, with a 70-meter wingspan wider than a 747. SpaceX has filed to eventually build up to a million of them in a new factory in Bastrop, Texas, with volume by the end of 2027.

Why orbit? Because the sun never sets up there. You get about 1,361 watts per square meter of continuous solar with no clouds, no nighttime, no grid to wait years to plug into. You dump waste heat straight into the cold of space by radiating it, so no water, no cooling towers, no reservoir getting drained to keep a data center cold. The two things strangling AI compute on Earth right now, power and cooling, are nearly free in orbit. And AI-1 is derived from the Starlink V3 platform: same laser links, same solar tech, same factory muscle. That's the convergence almost nobody zooms out far enough to see. Rockets feed satellites, satellites feed the internet, the internet platform feeds the data centers.

It All Comes Back to One Sentence

The skeptics have a sharp objection here, and it's correct at today's prices. AI chips go obsolete every two or three years, so orbital compute means launching hundreds of thousands of tons of new chips forever, which at current launch cost runs to hundreds of billions a year just in shipping. Brutal. A non-starter. At today's cost.

But the orbital data center breaks even somewhere around $500 to $1,000 per kilogram to launch, and Starship targets under $185 to $250. The exact same cost collapse that dragged satellite internet from $6.55 to 30 cents is what drags compute in space from economically insane to economically inevitable. It is the identical problem governed by the identical curve. Starlink is the proof the curve is real and already paying out. The AI data center is the trillion-dollar payoff it unlocks next. They're the same bet.

So compress the whole thesis into one question and ask yourself honestly. Do you believe Starship gets under $250 per kilogram and stays there? If yes, the rest of the case follows logically, and the two conditions already true carry a lot of the weight. If no, or if you're not sure, treat the $1.77 trillion exactly as the bears describe it and size accordingly. Either way, you live downstream of that curve. Cheaper bandwidth brings billions of new people online into markets that didn't exist. Cheaper launch turns space from a government-sized project into a place where actual businesses get built. Whether or not you ever buy a share, that's where a lot of the next decade's wealth gets created.