Trending
Are you ready? SpaceX is set to IPO today under the ticker SPCX.
It’ll be the largest IPO in history, and should instantly slot in as the eighth-largest company in the world by market cap on Day 1, right behind Broadcom (AVGO).
I’ve been writing about why I’m bullish on the company long term. If you missed my recent pieces, catch up here and here.
Today, I want to clearly break down the businesses that make up SpaceX, so you know what you’re buying should you choose.
You see, when most folks hear “SpaceX,” they think rockets. But if you buy stock in SpaceX, you’re not buying a rocket company. You’re buying a vertically integrated infrastructure platform with three layers: space, connectivity, and AI.
Let’s take a look. First up…
Space: the foundation. This is the launch business. Falcon 9, Falcon Heavy, the Dragon capsule, Starship, NASA missions, defense launches, commercial payloads, and, someday, lunar and Mars logistics. In 2025, the Space segment generated revenue of about $4.1 billion.
More important than the revenue is the dominance. SpaceX essentially owns the gateway to orbit. Everybody else operating in space (companies building satellites, sensors, defense systems, lunar landers) still needs an affordable, reliable ride. And much more often than not, SpaceX sells the ticket.
It’s important to note, however, that the Space segment is not SpaceX’s profit engine today.
Remember, it’s building the biggest, most powerful rockets in history. That requires an enormous amount of investment. So while this segment’s strategic value is huge, its near-term profitability is not.
Connectivity: the cash engine. This is Starlink. It’s like the cash machine built on top of the Space segment.
Starlink is the largest satellite constellation ever built—more than 10,000 satellites beaming high-speed internet to over 10 million subscribers across more than 100 counties, from mountaintops to open ocean to aircraft in flight. Its subscriber base almost doubled in the past 15 months. It took the old telecom giants decades to reach comparable scale.
In 2025, Starlink’s revenue grew roughly 50% year over year to about $11.4 billion and accounted for 61% of SpaceX’s total revenue. And it’s profitable, producing around $4.4 billion in operating income at a 38% margin.
Starlink’s flywheel effects are compounding. Each new subscriber makes the constellation more valuable for everyone else over time by contributing recurring revenue that directly funds more satellite launches, a denser constellation, upgrades to new generations of satellites, global expansion, lower latency, and new features.
Analysts expect Starlink revenue to reach $16 billion to $24 billion this year as the subscriber base continues to grow rapidly. It already controls about 90% of the satellite broadband market and its margins are improving as the constellation scales.
Put simply, Starlink gives SpaceX a high-margin, predictable, subscription-based revenue stream with massive long-term flywheel effects and enormous barriers to entry. You can think of it like Amazon’s AWS cloud computing business—a recurring revenue machine quietly humming away inside what most folks still see as a rocket company.
Falcon 9 made Starlink possible. Starlink made SpaceX financially strong. And the next-generation rocket, Starship, is expected to make Starlink much larger. Bigger and better satellites, more capacity per launch, better service, lower costs, and faster expansion into direct-to-cell service (beaming signals directly to your handheld device instead of a dish/terminal on your roof, boat, plane, etc.).
AI: the growth accelerator. In February 2026, SpaceX acquired xAI, the company behind the Grok chatbot, which also owns the X platform (formerly Twitter). So if you buy SpaceX stock, you’re also buying one of the world’s largest AI companies.
The company’s AI stack includes xAI’s Grok models, the X platform as a real-time data and distribution firehose, two of the world’s largest AI clusters called Colossus and Colossus II, contracts renting out some of that computing power, and a long-term plan to put AI data centers into orbit.
In 2025, this segment generated about $3.2 billion in revenue. The company spent a lot more than that building it out, but there’s real traction underneath the losses.
SpaceX recently inked a deal to rent out some of its computing power at its Colossus data center to AI powerhouse Anthropic (the company behind Claude) for a reported $1.25 billion per month.
Meanwhile, according to its S-1—the official document companies must file with regulators before going public in the US—SpaceX estimates its total addressable market at $28.5 trillion.
The company calls it “the largest actionable total addressable market (TAM) in human history.” Of that $28.5 trillion total, a whopping $26.5 trillion (93% of it) is AI. And the biggest single piece, estimated at $22.7 trillion, is the market for AI that businesses everywhere might buy someday.
Now, TAM is not revenue or profit. But it highlights how much of SpaceX’s future is tied to AI, by management’s own estimates.
The next frontier: Terafab and orbital AI data centers
Here’s where things get really interesting…
On March 21, Musk officially unveiled Terafab, a joint venture between SpaceX, Tesla (TSLA), and Intel Corp. (INTC) to build the largest semiconductor fabrication plant in history. The facility would consolidate every stage of chip production under one roof. That’s never been done before.
The goal: Produce 1 terawatt worth of AI computing power per year, about 50X the current global AI compute capacity.
The kicker: Musk expects 80% of that output to go to space, where SpaceX will do the AI computing hyperscalers like Amazon and Microsoft currently do on Earth.
SpaceX has filed an application with the FCC to launch up to one million AI data center satellites into low-Earth orbit.
These won’t beam internet like Starlink satellites. They’ll be distributed orbital data centers running AI workloads, powered by solar energy—where sunlight is continuous in orbit and about 1.4X stronger than on the ground, and there’s no need for real estate, bulky and expensive cooling systems, or grid connections.
Musk’s argument is that as launch costs continue to decline, running AI compute in orbit will become cheaper than doing it on the ground within two to three years.
Whether you fully buy that timeline or not, the vision is extraordinary. SpaceX wouldn’t just be the company that launches things into space and provides satellite internet; it would own the compute layer of low-Earth orbit.
Here’s the takeaway many folks miss: SpaceX is becoming a vertically integrated space infrastructure platform—rockets + satellite broadband + AI compute + its own chip supply chain. And the Terafab and orbital AI data center efforts tie the ecosystem together into one unified bet on the future of AI infrastructure.
More soon… Let me know if you’re planning on buying SpaceX on Day 1. And please like and share to help us continue to grow.