WTF Is Happening at SpaceX: Why Launch Monopoly Meets Orbital Internet

SpaceX is confidentially filing to go public at roughly $2 trillion. That would be the largest IPO in history — bigger than Saudi Aramco's 2019 raise — with talk of pulling in $50 to $75 billion in one offering. The financial media will fight over the number. Bears will call it insane. Bulls will call it cheap. Both miss what actually matters.

You are not buying a rocket company at $2 trillion. You are buying a near-monopoly on orbital launch plus a rapidly scaling global internet business with about 9 million paying subscribers, more than $15 billion in government contract backlog through 2030, and a next-gen vehicle that rewrites what space costs. Once you see that stack, $2 trillion starts looking more like a floor than a ceiling.

Nothing here is financial advice. It is an explainer of what SpaceX is today, what Starship changes, and the convergence almost nobody prices correctly.

From near-death to 82% of commercial launch

In 2008, SpaceX almost died. Three Falcon 1 flights failed in a row. Elon had already poured most of his PayPal fortune into the company. A fourth failure would have ended Falcon 9, Starlink, Starship, and any IPO talk. The fourth launch worked. That was 18 years ago.

Today the same company is filing at a valuation larger than the GDP of most countries on Earth. In 2025, Falcon 9 flew about 165 orbital missions — roughly one every two days. SpaceX now controls roughly 82% of the global commercial launch market. Imagine one airline running 82% of every flight on the planet. That is SpaceX for getting mass into orbit.

Launches are the spectacle. They are not where most of the money comes from.

Starlink is the cash cow

Most people still file SpaceX under "rockets." Five years ago that was fair. Now it is better described as a telecom company that also builds rockets. Starlink is why.

More than 10,000 satellites are on orbit, serving rural homes, ships, planes, military units, and plenty of city customers who just want something that works. The constellation threw off something like $10–12 billion in revenue in 2025. Total company revenue landed around $15–16 billion with roughly $8 billion in profit. Elon has said NASA will be only about 5% of SpaceX revenue in 2026. The vast majority is commercial Starlink.

So when someone says "$2 trillion rocket company," translate it. Buyers get launch dominance and a global internet network growing at a brutal clip, with a government backlog that stretches to 2030. Rockets open the door. Connectivity pays the rent.

Keep one more detail: in February, SpaceX absorbed xAI in an all-stock deal. In March 2026, Tesla, SpaceX, and xAI jointly announced Terafab — a $25 billion chip fab complex in Austin targeting 100,000 wafer starts per month, scaling toward a million. Custom AI and memory chips for cars, humanoid robots, and AI satellites. About 80% of the compute output is aimed at space. That is shared infrastructure across Elon's companies, not a side quest.

Why Starship is the real valuation story

Starship is why this valuation can even be discussed. Not because of what it flies today, but because of what it makes cheap enough to try.

Target payload to low Earth orbit: 100–150 metric tons. Expendable mode pushes toward 250 tons. Falcon 9 carries about 23 tons. Starship is five to seven times that per flight.

Payload is not the crazy part. Cost is. Falcon 9 already runs customers roughly $2,700–$3,000 per kilogram — cheapest in the world by a wide margin. ULA's Vulcan sits near $4,000/kg. Europe's Ariane 6 runs $4,000–$6,000/kg. A decade or two ago you were talking $10,000 to $50,000 per kilogram. Starship's aspirational target is $10–$100 per kilogram. That is a 30x–300x cut versus Falcon 9, and more than a thousand-fold versus where the industry sat ten years ago.

How? Full reusability. Booster and upper stage both come back, get refueled, and fly again — like an airliner, not a disposable missile. SpaceX already proved partial reuse with Falcon 9. Starship takes it further with tower catches and upper-stage recovery in test. Full routine reuse is not proven yet. The engineering is moving fast.

Propellant might run $0.5–2 million per launch. A vehicle might cost ~$90 million to build. Fly it 100 times and vehicle cost per flight is $900,000. Add propellant and ops and you are in the $2–5 million per launch band for 100–150 tons — math that lands near tens of dollars per kilo.

When space gets cheap, you do new things

Cut the cost of orbit by ~100x and you do not just do the old missions cheaper. You unlock missions that never made sense. Wright's law says costs fall as cumulative production doubles — the same curve that took solar from $76/watt in 1977 to about $0.20 today. Jevons paradox says when something gets cheap, volume explodes. Cheap smartphones after 2007 created Uber, Instagram, mobile banking, and creator economies. Starship is the iPhone bet for space: a platform, not a ticket to LEO.

At $10–$100/kg you can talk seriously about AI satellite swarms, orbital data centers with continuous solar and radiative cooling, microgravity manufacturing, and point-to-point Earth transport — New York to Shanghai in about half an hour, or 100 tons of cargo anywhere on the planet faster than a Netflix episode. DoD is already poking at contracts for that kind of logistics.

Energy is the AI bottleneck. Microsoft, Google, Amazon, xAI, Meta — all scrambling for power. Microsoft is restarting Three Mile Island because the grid cannot feed the clusters. An orbital data center with constant sunlight and vacuum heat rejection sounds absurd until Starship's cost curve shows up in the spreadsheet. Some analysts sketch 5,000–10,000 Starship flights to build large-scale orbital AI. At $2–5 million per launch, that is $10–50 billion for a shot at effectively unlimited compute — less than what Big Tech already spends on terrestrial campuses in a busy year.

Laughing at a "$2 trillion rocket company" is valuing the wrong object. You are pricing a platform that could unlock a multi-trillion-dollar space economy that does not exist yet.

People, competitors, and who finally gets shares

SpaceX employs roughly 14,000 people. Many hold options granted when the company was worth a few billion. A $2 trillion IPO is a generational wealth event for early engineers. For the rest of the industry, it is existential. Boeing's Starliner has burned more than a billion in overruns. ULA's Vulcan is more expensive and less capable than Falcon 9, let alone Starship. Pour $50–75 billion of IPO cash into Starship and Starlink and the gap widens. The short-term silver lining: when IPO news hit, Rocket Lab, Planet, and Intuitive Machines all jumped. A SpaceX IPO tells pensions and sovereign funds that space is a real sector.

Skeptics will hammer the multiple: ~$15 billion revenue into a $2 trillion price tag is roughly 80x sales. The counter is platform optionality. 2026 revenue talk already sits around $22–24 billion, and Arc Invest has modeled ~$2.5 trillion enterprise value by 2030 with a bull case near $3.1 trillion.

For retail, this is the first clean on-ramp after years of secondaries and accredited-investor walls. Morgan Stanley is the frontrunner to lead; E*TRADE talk includes retail allocation up to ~30%. Twenty-one banks on the deal.

National security and the convergence

Starshield is the military flavor of Starlink. SpaceX flies classified payloads and builds the rockets that loft America's sensitive birds. China is building constellations, reusable rockets, a station, and moon plans. This race is about who owns communications, launch, orbital compute, and military infrastructure for the rest of the century. A public SpaceX with tens of billions of fresh capital accelerates the U.S. side.

Then comes the bigger frame. Chamath Palihapitiya argued on All-In in January 2026 that SpaceX might reverse-merge into Tesla so Elon consolidates cars, robots, autonomy, AI training, Starlink, Starship, and custom semiconductors onto one cap table — the Berkshire Hathaway of this century. Terafab is already the physical tell: shared chip infrastructure, with Optimus expected to help build and run the fab. Layer Tesla AI, SpaceX launch, xAI models, Starship lofting robots and chips, orbital data centers, and Starlink as the backbone, and you get one vertically integrated machine spanning Earth, orbit, and eventually Mars. Polymarket has put Tesla–SpaceX merger odds in a rough 6–18% band depending on horizon.

Gwynne Shotwell, SpaceX president and COO for two decades, told Time she is not supposed to talk about the IPO — and that she is looking forward to it anyway. The last time one entity held this much capability across this many domains, it was a nation-state, not a company. Whether SpaceX floats alone at $2 trillion or folds into a larger Tesla stack, the infrastructure being built will shape jobs, energy, communications, transport, and intelligence systems for decades. That is what is actually happening at SpaceX. The launch footage is the trailer. The monopoly, the internet, the cost curve, and the chip-AI-orbit convergence are the movie.

Check the video here.