WTF Is Happening with Elon: What Most Narratives Miss

Three days after SpaceX went public in the largest IPO in human history, Elon Musk replied to a financial reporter on X with two sentences. SpaceX might reach about $1 trillion in revenue in 2030. He would be surprised if revenue is not greater than $1 trillion in 2031.

Elon says crazy things. He is late on basically everything. Take those two facts as given. Then take the claim seriously anyway. The SpaceX that just promised a trillion dollars is not primarily a rocket company anymore. Once you see what it became in early 2026, the number stops sounding like pure theater and starts sounding like a bet on a different business.

The math that should make you flinch

In May 2026, SpaceX filed an S-1 — the official paperwork before a company sells shares to the public. SpaceX's total revenue in 2025 was $18.67 billion. Call it $19 billion. Getting from roughly $19 billion to $1 trillion in five years means multiplying revenue by about 54 times. That works out to growing at about 123% every year for five years straight. Not a spike. Five years in a row.

The largest annual revenue any company has ever produced was Amazon at around $716 billion. Anthropic more recently printed upwards of $40 billion in about four years. Amazon took three decades, a global retail empire, and the biggest cloud business on Earth. Elon is saying SpaceX beats the all-time records.

Wall Street is not buying that timeline. The most bullish 2030 forecast on the Street comes from Goldman Sachs at around $470 billion. Morgan Stanley sits near $330 billion. The single most optimistic professional analyst on the planet has SpaceX at less than half of what Musk just claimed. On the surface this looks like classic Elon: richest human alive, fresh off the biggest IPO ever, throwing a number into the feed. It is not that simple. It has everything to do with artificial intelligence.

The company inside the company

When people hear SpaceX, they picture a rocket landing on a drone ship or getting caught by tower arms. Fair. That S-1 tells a different story. It breaks the company into three businesses, and the rocket part is the smallest.

Segment one is space: Falcon 9, satellite launches, government and defense contracts. In 2025 that brought in about $4 billion. It lost money on an operating basis — roughly $650 million — because the company is pouring about $3 billion a year into Starship. The famous part of SpaceX is the smallest segment, and it is slightly in the red.

Segment two is connectivity: Starlink. More than 10,000 satellites in orbit, more than 10 million subscribers in 164 countries. Starlink did about $11.4 billion in revenue in 2025 with $4.4 billion in operating profit. Almost three times the rocket revenue. The only consistently profitable piece of the whole operation. Financially, this is a satellite internet company that also launches rockets.

Segment three is AI. Hold that. Picture the whole thing first: a roughly $19 billion company whose profit engine is internet satellites, whose rockets lose money, and whose boss just promised 54x the whole machine in five years. If you are trying to decide whether any of these Elon numbers deserve airtime, this is the moment skepticism should be loud. Keep that skepticism. Then look at what happened three months before the IPO.

The absorption nobody priced correctly

In February 2026, SpaceX absorbed xAI. That is the single most important fact in this story. xAI builds Grok, the assistant that lives on X and inside Tesla cars. SpaceX swallowed it whole and renamed the piece SpaceX AI. The combined entity was valued around $1.25 trillion — roughly a trillion for SpaceX and roughly $250 billion for xAI. After going public, SpaceX is worth over $2 trillion.

When the S-1 dropped, the accountants restated the books back to 2023 as if these had always been one company. That is why the third segment exists. xAI plus the X platform did about $3.2 billion in 2025 revenue. In that same filing, SpaceX put its total addressable market at around $28.5 trillion. Roughly 90% of that number was attributed not to rockets, not even to Starlink, but to the AI business.

The trillion-dollar revenue prediction has almost nothing to do with rockets. Almost everything to do with AI.

The steel-man, then the trapdoor

The argument runs like this. The economy is being rebuilt around AI. AI runs on compute — data centers full of specialized chips that train and run models. Whoever controls the most compute at the lowest cost captures an enormous slice of the build-out. There is already real evidence SpaceX is a serious compute player.

Anthropic is paying around $1.25 billion a month for compute through 2029. Google is paying around $920 million a month starting in late 2026 for capacity it openly called bridge capacity for Gemini. Those are signed deals. Billions a month. Important detail: those deals are for compute on the ground, in places like Memphis packed with hundreds of thousands of Nvidia chips. They are not compute in orbit. The money you can see today is earthbound. They are renting buildings full of chips.

That matters for the trillion-dollar number. AI revenue is real and growing fast, but it is growing where SpaceX has to compete head-to-head with Microsoft, Amazon, Google's own data centers, and OpenAI's Stargate. Brutal, capital-soaked fight. New ground data centers are getting harder too — NIMBYs, grid fights, political backlash.

Break down where a trillion would have to come from and you hit something uncomfortable. The math only closes if SpaceX builds a brand-new economy in space that does not exist yet. Ask whether that economy is physically possible this decade and you are asking the right question.

Four things that all have to work

To hit a trillion in 2030, four things need to happen at once. You need a Starship that is rapidly reusable and built at scale — the only rocket that can put enough satellites up for that kind of revenue. You need chips for those satellites, bought or built. You need to solve radiating heat away from massive orbital clusters. And you need those satellites connected so they work as one coherent cluster for AI workloads.

All of that is extremely hard. The pieces are starting to show up anyway. In June 2026, days before the IPO, Musk unveiled AI-1, an orbital data center. The pitch is simple: unlimited solar with no night or clouds in the way, and the cold of deep space to dump heat into. Instead of fighting communities for electricity and water on Earth, put the computers in orbit, point the panels at the sun, and let them run. Elon has said AI satellites are simpler than Starlink satellites and use technology they have mostly already built.

There is also Terafab, a chip factory joint venture between SpaceX, Tesla, xAI, and Intel, estimated at $55 to $120 billion. His reasoning in one sentence: build a Terafab or do not have the chips — and they need the chips. Vertically integrate whatever is blocking you. Manufacture the chips, launch them on your own rockets, power them with the sun, rent the computing power to the AI industry. Chip to orbit. One flywheel.

The heat problem and the deadline problem

When a chip computes, it produces heat. On Earth you blow air or run water over it. In vacuum there is no air and no water. The only way to dump heat is to radiate it slowly off giant panels. Analysis in IEEE Spectrum is brutal: a serious orbital data center would need so many radiators that the radiators themselves could outweigh the computers by something like 10 to 1. Ten pounds of cooling fins for every pound of computer.

Space is also full of radiation. Radiation degrades normal high-end AI chips in weeks to months unless they are hardened, which makes them slower and far more expensive. In February 2026, Sam Altman looked at orbital data centers mattering at scale this decade and called the idea ridiculous — that it will not matter at scale this decade. He is a competitor with every incentive to trash the idea. Factor that in. Also factor in that Elon is the king of missed deadlines and easy to dismiss because of it.

He is almost never wrong about whether the thing happens. He is often wrong about when. Reusable rockets that land themselves. EVs at massive scale. A global satellite internet constellation that actually works. Brain chips helping paralyzed people. Self-driving cars going coast to coast. They all looked impossible. They all arrived late. They all arrived.

Right as his own company was going public, Starship — the rocket the entire physical vision depends on — had flown exactly one time in 2026. The trillion-dollar story rests on a vehicle that has not delivered a single commercial payload to orbit yet. Plenty of people will call 2030 or 2031 revenue of a trillion a fantasy. They are not crazy for saying that.

What actually matters

We spent a hundred years valuing companies by what they make: cars, phones, barrels of oil, boxes on ships. The richest man alive just told the market his rocket company will out-earn the largest company in history not by building bigger rockets, but by becoming the place the entire economy goes to think.

If he is even half right — not on the date, on the direction — then the most valuable thing a company can own in the 2030s is the power and the silicon to run intelligence itself. The ability to think at scale becomes the product. One person is trying to own it from the chip to the orbit. That is what is happening with Elon. The rocket narrative is the distraction. The AI stack is the story.

Check the video here.