TLDR: A year ago SpaceX lost $1.1 billion in a quarter. This quarter it made $4.4 billion, and its largest business is no longer rockets. It’s AI. I believe SpaceX is assembling something no company has ever owned: rockets, a planetary communications network, gigawatts of compute, and the energy to run it, all under one roof. That’s the path to $10 trillion, and then $100 trillion.
This morning I joined CNBC’s Squawk on the Street to discuss SpaceX’s third-quarter results. Disclosure: I’ve known Elon for 27 years and first invested in SpaceX in 2013. Here’s the case, because I think most people are still looking at SpaceX through the wrong lens.
THE QUARTER THAT CHANGED THE STORY
Here are the numbers SpaceX reported on October 20:
Revenue: $14.6 billion, up 161% year over year.
Net income: $4.4 billion, versus a $1.1 billion loss a year ago.
Adjusted EBITDA: $8.8 billion, up 306%. Profit grew roughly twice as fast as revenue.
Guidance: Q1 2027 revenue of $33 to $37 billion, against analyst estimates in the low $20 billions.
And the headline: SpaceX’s Super Intelligence business, which rents compute and sells Grok and Cursor, generated $7.6 billion, up about 830%. It’s now bigger than Starlink ($5.25 billion) and more than four times the size of the launch business ($1.8 billion).
SpaceX quarterly revenue by segment, Q3 2025 vs. Q3 2026. Source: SpaceX Q3 2026 shareholder deck.
“SpaceX has stopped being a rocket company. It’s a superintelligence company, a global communications company and the dominant infrastructure company of the Intelligence Revolution.”
THE CASE IN FIVE PARTS
Each piece is strong on its own. Together, they form a flywheel no competitor can copy.
1. COMPUTE: THE BOTTLENECK MOVED, AND ELON IS STANDING ON IT
For three years, AI’s constraint was chips. Now it’s energy, cooling and physical infrastructure: how fast you can turn a warehouse of GPUs into running compute.
That’s exactly where SpaceX excels. Elon said on the earnings call that SpaceX aims to deploy “power, cooling infrastructure at a rate that is vastly faster than anyone else.”
The results: compute capacity went from 1.4 to 1.7 gigawatts in one quarter, with 2.5 to 3 gigawatts targeted by year-end. SpaceX is bringing roughly 200,000 Nvidia GB300s online every month and plans to keep that pace next year with Vera Rubin chips. Elon talks about 13 gigawatts by the end of 2027, with 20 or more possible.
Management says new compute deployments pay back in under a year, and Elon floated six months next year. Google and Anthropic rent the capacity, and SpaceX is its own biggest customer, training Grok for the enterprise.
2. STARLINK: BECOMING THE NERVOUS SYSTEM OF THE PLANET
Starlink now has 13.6 million subscribers, adding 1.6 million in the quarter, across 174 countries with about 11,100 satellites. It’s active on more than 3,800 aircraft, and connectivity revenue grew 69% to $5.25 billion.
But this is still early. On the call, Elon said that “probably within three years” Starlink will have “about 100 times the bandwidth of the current Starlink constellation.” A single Starship flight, only about half loaded, just added 26 terabits per second, equal to 10 Falcon 9 launches.
Then comes your phone. SpaceX is buying 800 MHz spectrum for Starlink Mobile, and the FCC has approved a 15,000-satellite constellation for it. The dead zone is going extinct.
Here’s what most analysts miss: soon, the dominant users of the internet will be AI agents talking to other agents, by the billions. That traffic needs a network, and Starlink is the only one that covers the entire planet.
3. STARSHIP: THE BACKBONE EVERYTHING RUNS ON
Launch grew only 6% and ran a small loss this quarter. But it’s the foundation the other two businesses are built on.
Flight 14 was Starship’s first time reaching orbit, carrying 26 operational Starlink V3 satellites, with about 10 times the downlink capacity of a Falcon 9 Starlink launch. Next month, Flight 15 will attempt to catch both the booster and the ship. If it works, it will be the first full recovery of an orbital rocket in history.
The breakthrough here is frequency, not size. SpaceX is targeting weekly Starship launches in the first half of 2027 and twice weekly by year’s end. That’s the airline model applied to orbit.
4. INTEGRATION: OWNING THE WHOLE STACK
This is the real moat. SpaceX launches its own satellites, runs its own network, builds its own data centers and powers them. With Terafab, it plans to make its own chips. Across SpaceX and Tesla, Elon has committed to 100 gigawatts of solar, which may be the fastest energy ramp on the planet before fusion arrives.
That’s why I told CNBC I’m in favor of a SpaceX-Tesla merger. Combine digital intelligence (Grok and the compute behind it) with physical intelligence (Optimus robots and Cybercab), and you get a company integrated up and down the entire superintelligence stack. Special committees have formed, but no terms are agreed, and a deal needs 50% of all outstanding Tesla shares. That’s a high bar.
5. ELON: THE MAN WHO BUILDS THE MACHINES THAT BUILD THE MACHINES
I’ve watched people underestimate Elon for more than 20 years. He’s sometimes late. He’s rarely wrong about the direction.
He isn’t driven by wealth. He’s driven by problems that need solving. I guarantee there are countries that wish they had someone like Elon driving their economy.
“I don’t think we’ve seen a fraction of what SpaceX is going to deliver over the decade ahead.”
THE MATH TO $100 TRILLION
At about $148 a share, SpaceX is worth roughly $2 trillion today. Getting to $10 trillion takes 5x. Getting to $100 trillion takes 50x. So let’s look at the path.
SpaceX expects to reach $100 billion of annual recurring revenue by December. Elon has put end-of-2027 ARR between $500 billion and $1 trillion. On CNBC, I was asked whether SpaceX could hit a trillion dollars in annual revenue by 2028 or 2029. My answer: “100%, yes.”
A trillion dollars of revenue at a 10x multiple is a $10 trillion company. The $100 trillion case rests on a bigger idea: Elon predicts AI and robots will make the global economy 10 times larger within a decade. If SpaceX supplies the launch, network, compute and energy for that economy, it captures a share of a far larger pie. Elon has gone further, saying he expects SpaceX to be worth multiples of Earth’s entire GDP… And that would make it a quadrillion-dollar company plus.
That’s the bull case. It’s ambitious, and I believe it.
WHAT COULD GO WRONG
The spending is enormous. SpaceX spent $19.6 billion on capex in Q3, $16.4 billion of it on compute, and expects $35 to $40 billion in Q4, more than twice Q3’s revenue. Watch the return on that capital, and the contracted number: a $63 billion backlog, about $42 billion of it expected within a year.
An AI credit cycle. In the segment before mine, investor Bob Elliott warned that corporate AI spending has flattened for four months and that overbuilt capacity could trigger credit problems within 12 to 18 months. It’s a serious argument. My counter: SpaceX is profitable, investment-grade and building against visible demand.
Political and key-man risk. Washington’s support could change after the midterms, and no company has ever depended this much on one person. My view: the world depends too much on SpaceX infrastructure for anyone to throttle it.
WHAT THIS MEANS FOR YOU
If you’re an entrepreneur: The bottleneck has moved from chips to energy, cooling and infrastructure. Build there, or build on top of the networks being created.
If you’re an executive: Plan for your AI agents talking to your customers’ agents, over bandwidth and compute that keep getting cheaper.
If you’re an investor: Watch three numbers each quarter: compute payback period, Starship launch cadence and contracted backlog. They tell you whether the flywheel is spinning.
If you’re a student: The most valuable skills of the next decade sit at the intersection of AI, energy and hardware. Learn to build physical things.
If you’re a parent: Teach your kids to save and own equity in the future. Elon’s pledge of a SpaceX share to every American child is a great lesson in compounding.
In 2004, when we awarded the $10M Ansari XPRIZE, most people thought private spaceflight was a hobby for billionaires. Today, a private space company is on track to become the infrastructure layer of the Intelligence Revolution. The question for the next five years isn’t whether SpaceX keeps growing. It’s how much of the economy of the future ends up running on it.
To a future of abundance,
Peter
