The opening bell rang on the Nasdaq on Friday, June 12, and a ticker that didn't exist a week earlier — SPCX — started printing numbers. By the close, SpaceX was worth $1.75 trillion. The stock jumped 19% on day one to $161, and for a few minutes the company's paper value pushed past $2 trillion.
It was the largest IPO in the history of public markets. SpaceX raised $75 billion. The previous record, Saudi Aramco in 2019, raised $29.4 billion. This wasn't a new record. It was the old record set on fire.
Retail investors lost their minds. Through Robinhood, Fidelity, and SoFi, ordinary people placed roughly $100 billion in orders — more than the entire offering. Everyone wanted a piece of Elon Musk's rocket company. Most of them won't get much: only 3% to 4% of shares are actually trading. More than 75% of the offering went to insiders and existing investors before the public ever saw a share.
The part nobody put on the confetti
SpaceX makes money from exactly one thing: Starlink. Nine million subscribers beaming internet from orbit. That's the profitable business, and it's a good one. The rockets were profitable as recently as 2024.
Then there's xAI, the artificial intelligence company that merged into SpaceX in February. xAI lost approximately $6.4 billion in 2025. Its chatbot, Grok, has captured almost no market share against ChatGPT, Claude, and Gemini. The combined company posted a net loss of $4.27 billion in the first quarter of 2026 alone — up from $528 million in the same quarter a year earlier. The accumulated deficit sits at $41.3 billion.
So here's the trade investors just made. They paid $1.75 trillion — about 94 times the company's $18.7 billion in 2025 revenue — for a satellite business duct-taped to an AI money pit. Nvidia, the most valuable chipmaker on earth, trades at less than a quarter of that multiple. Morningstar values SpaceX at $780 billion and calls it "significantly overvalued." That's not a short-seller. That's a mainstream research desk saying the stock is worth less than half its price.
Why the AI part matters more than the rockets
The bull case isn't really about launches anymore. It's about compute. xAI's data centers are now renting their capacity out, and the contracts are eye-watering. Google agreed to pay $920 million per month for GPU access through mid-2029. Anthropic — yes, an xAI competitor — pays $1.25 billion a month to rent the entire output of the Colossus 1 data center. Combined, that's around $26 billion in annualized compute revenue, signed in under four months.
Impressive. Until you read the fine print: both deals carry 90-day termination clauses after December 2026, and Google itself called its arrangement "bridge capacity," not a commitment. The whole AI thesis rests on revenue that either party can walk away from in a quarter.
My Opinion
I'll be blunt: this IPO is a bet on a man, not a business. Musk controls roughly 85% of the voting power through super-voting shares. Outside investors who just handed over $75 billion have almost no say in anything. A Danish pension fund called the governance structure "catastrophic" and blacklisted the stock. Senator Elizabeth Warren asked the SEC to scrutinize the listing because index funds will soon be forced to buy SPCX, dragging millions of passive savers into Musk's orbit whether they chose it or not.
Here's what bugs me most. The market wrapped a $6.4 billion AI loss inside a satellite company and a rocket legend, and called the package the safest mega-bet of the decade. xAI couldn't raise this on its own merits — Grok is an also-ran. So it got stapled to Starlink and floated on the strongest brand in tech. That's not a vote of confidence in AI. It's a workaround.
The tiny float will probably keep the price high for a while. Scarcity does that. But when the lock-ups expire and real supply hits the market, we'll find out whether people paid $1.75 trillion for the future of intelligence — or for the privilege of holding Elon Musk's losses. My money says the second number gets a lot smaller before it gets bigger.
Author: Yahor Kamarou (Mark) / www.humai.blog / 14 Jun 2026