Trending
Time to buy SpaceX? Back in June, just before the biggest IPO in history, I warned you not to buy SpaceX (SPCX).
I showed Jolt readers how large IPOs tend to do poorly out the gate.
Six weeks ago SpaceX peaked at $225. Today it trades near $110.
It’s lost roughly $1.4 trillion in market cap since its peak, which is more than a whole Berkshire Hathaway.
It's now almost 20% below the $135 IPO price, meaning virtually everyone who bought this stock after shares began publicly trading is underwater.
My problem with SpaceX was…
It just wasn’t worth the $2 trillion valuation it went public at.
I’m a huge SpaceX fan and think it’s the single most important company in America today. Heck, they catch skyscraper-sized rockets with chopsticks! And having met many present and former SpaceX employees I can tell you they are the most brilliant minds on the planet.
But when I see a company with $18.7 billion of 2025 revenue trading at 100X sales, I wince.
Now that the post-IPO selloff knocked $500 billion+ off its market cap, I think we’re getting close to a buying opportunity.
$1 trillion…
That’s how much I think SpaceX’s core businesses, launch and Starlink, are worth.
SpaceX essentially owns the gateway to orbit. It flew 165 Falcon missions in 2025 and carries 80%+ of all mass to orbit.
Every space company needs a cheap, reliable ride. That includes satellite makers. Defense firms. Sensor companies. And lunar lander builders. And most of the time, that means paying SpaceX for a spot on one of its rockets.
Now couple this with Starlink, the largest satellite constellation ever built. More than 10,000 satellites beam high-speed internet to over 10 million subscribers in more than 100 countries.
Starlink’s subscriber base has almost doubled in the past 15 months. It took the old telecom giants decades to reach comparable scale.
In 2025, Starlink’s sales jumped about 50% to $11.4 billion. It now brings in two-thirds of SpaceX’s revenues.
Starlink’s subscriber base is still growing fast. Its revenue could double this year.
It’s also rolling out its direct-to-cell service which will deliver full contiguous global coverage by mid-to-late 2027.
Building a competing network would require tens of billions of dollars. Thousands of satellites. A reliable rocket fleet. And years of regulatory approvals. Very few companies have the money or technical ability to pull that off.
Both Starlink and the launch business are near-monopolies with canyon-sized moats.
xAI is easily worth another $300 billion.
SpaceX absorbed xAI back in February. It owns:
--xAI’s Grok models.
--X (formerly Twitter)
--And most importantly, two of the world’s largest AI data centers.
In 2025, this segment brought in some $3.2 billion in revenue. It’ll rake in a lot more in 2026 and 2027.
SpaceX recently inked a deal to rent out some of its computing power to AI powerhouse Anthropic. It will reportedly pay $1.25 billion per month.
It also struck a similar agreement with Google. The tech giant will pay $920 million per month to use Colossus from October 2026 through June 2029.
These two deals alone are set to bring in $26 billion per year.
SpaceX is worth $1.5 trillion at today’s prices. If xAI, Starlink, and the launch business are worth about $1.3 trillion altogether, then…
At today’s market cap of $1.5 trillion…
You’re paying $200 billion for data centers in space.
I want to clear something first...
The phrase “data center in space” creates the wrong image in most people’s minds. You probably picture a giant floating warehouse packed with endless rows of Nvidia chips.
Instead, a space data center may use hundreds or thousands of small satellites connected with lasers. Each satellite would carry chips. Solar panels. Communications equipment. And large panels that release heat. Working together, they would act like one giant computer.
And because they’re operating in orbit, they avoid many of Earth’s biggest problems, namely NIMBY (“not in my backyard”) resistance.
The big issue keeping them back is launch costs.
Elon Musk argues that he’ll bring those down enough so running AI compute in orbit will become cheaper than doing it on the ground within two to three years.
Starship is key to this. SpaceX’s massive next-gen rocket promises to make sending cargo to space 4X-9X cheaper than it is today.
Last week, Starship completed its 13th test flight. It was the first time it deployed commercial payloads into space. This time, Starship carried 20 Starlink V3 satellites.
Earlier flights carried dummy payloads that only copied their size and weight. Also important is that V3s are in the same general design class and size scale as AI data center satellites.
One last thing…
August 6 doesn’t scare me.
You’ll see this date repeated by the pundits in financial media.
On that date, employees and early investors can start selling 911.5 million shares. That is about 1.5 times the number trading today. If they end up selling en masse, it could tank SPCX stock.
Here’s what no one mentions... SpaceX employees could already sell shares through regular tender offers. Those offers came about twice a year. Yet few did. Quite the opposite... When insiders had a chance to buy more stock at IPO, most did.
I’m not dismissing the risk. But it’s overblown.
A simple way to mitigate risk in case the stock does indeed go lower is to buy in tranches.
A third at today’s prices, roughly 50% off the high.
A third at $100.
And the last tranche at $85. At that price you pay less than the parts are worth. The space data centers come free, and you get the rest at a discount.
If it never hits $85, you own two-thirds of a position in the best hard-tech company on earth at a fair price.
If you liked this, consider joining me inside the Jolt. Every week, I write about disruptive megatrends like AI and what to do as an investor. Join here for free.