Key Takeaways
- The head of data center energy at Alphabet’s Google says data centers in space can cost up to 100 times as much as ones on Earth, and that US power is sufficient while permits are slow.
- Jim Chanos says that gap should give pause to anyone modeling SpaceX, and that the profits from AI compute will go to GPU makers like NVIDIA.
- Analysts expect SpaceX’s revenue to reach $336 billion in 2028, and even bull Shaun Maguire doesn’t see orbital compute as its dominant revenue line until 2030 or 2031.
- Orbital compute deserves no credit in a SpaceX valuation until Starship shows what it costs.
Bulls on SpaceX (SPCX) spent the weekend arguing over how fast orbital compute could take over its business. Then Google weighed in with a number.
On Sunday, The Information reported that Amanda Peterson Corio, head of data center energy at Alphabet (GOOG) unit Google, puts the cost of data centers in space at up to 100 times that of ones on Earth. She also reportedly said US power supply is sufficient, and the slow part is permits.
“Let’s explore all the terrestrial options before we go extraterrestrial,” she said.
Short seller Jim Chanos read that as $5 trillion per gigawatt of compute, and added in a post on X: “The fact that the head of data center energy at a hyperscaler is this far off from estimates from $SPCX management/bulls should give pause to anyone trying to model this business out.”
That’s a lot of zeros to model around.
TIKR tracks 39 analysts on SpaceX. How many say “Buy”?
See the full breakdown of Buy, Hold, and Sell ratings on SpaceX, along with Wall Street’s revenue and earnings estimates, all in one place. More than 500,000 investors use TIKR to research stocks like this one. See SpaceX’s Analyst Ratings
Orbit’s revenue is years out
Analysts expect SpaceX’s revenue to climb from $48.9 billion this year to $336 billion in 2028:
Elon Musk says on X that orbital compute deployed per month “probably exceeds terrestrial around the end of 2028.” But Sequoia Capital partner Shaun Maguire, whose SpaceX model started the weekend’s fight, says orbital compute is “just not the dominant revenue line item until 2030 or 2031.”
For orbit, SpaceX needs Starship. Maguire says each launch carries about 10 megawatts of compute, and he’s forecasting “about 280 orbital compute launches by end of 2028.” If not then, he says, “it will be in 2029.”
On SpaceX’s third-quarter call, Musk guided to weekly Starship flights in the first half of 2027 and twice-weekly flights by the end of that year. Twice a week for all of 2028 comes to about 100 flights, for everything Starship carries, Starlink satellites included.
Multiply the payback by 100
SpaceX’s capital spending more than quadrupled from 2023 to 2025:
And SpaceX spent $48.1 billion in just the first nine months of 2026, with $35 billion to $40 billion of fourth-quarter capex guided on top.
SpaceX’s CFO said on its third-quarter call that new compute pays for itself in less than a year at current rental prices. Build that compute in orbit at up to 100 times the cost, rent it at the same price, and a payback of under a year becomes one of up to a century. (Even at a tenth of Google’s high end, you’re waiting about a decade.)
And if US power is sufficient and permits are the holdup, orbit’s edge over the ground comes down to skipping the permit line, at up to 100 times the price.
NVIDIA gets paid first
Chanos made a second point, replying to another X user: “Over time the economic rents will accrue to the GPU providers, not the equipment lessors/landlords. They are financial conduits.”
SpaceX fits that description today. It’s raising $40 billion of debt that Musk calls “primarily callable debt for GPU purchases,” and the chips come from NVIDIA (NVDA).
NVIDIA gets paid when those chips ship, and SpaceX only as long as customers keep renting them. Anthropic, which agreed in its IPO filing to spend up to $120.5 billion with SpaceX through 2029, can walk on 90 days’ notice once its first noncancelable periods end early next year.
Give orbit zero for now
Of course, Google is reportedly testing AI server chips in orbit itself, so it’s hardly a neutral party, and Starship is built to cut exactly the launch costs that make space so expensive.
But Chanos is right that a gap this wide should give pause. At 45x forward earnings as of Friday’s close, investors are already paying up for Starlink and the compute SpaceX rents out on the ground. I’d value the stock on those two and give orbital compute nothing until Starship shows what a gigawatt in space actually costs.
The first clue comes soon: Flight 15, which Musk expects in November, will try to catch both the booster and the ship. A cadence anywhere near Maguire’s would be the first real sign orbit can pencil out.
So what is SpaceX stock actually worth?
TIKR lets you forecast the future price of any stock in less than a minute. Just enter a few assumptions into TIKR’s valuation model and see what SpaceX could be worth. Start from Wall Street consensus estimates, or adjust the inputs to reflect your own view of the business. It’s free to use.
Disclaimer:
Please note that the articles on TIKR are not intended to serve as investment or financial advice from TIKR or our content team, nor are they recommendations to buy or sell any stocks. We create our content based on TIKR Terminal’s investment data and analysts’ estimates. Our analysis might not include recent company news or important updates. TIKR has no position in any of the stocks mentioned. Thank you for reading, and happy investing!
