Listen to the audio version of this article (generated by AI).
Google sends AI chips into space… Earth’s data center backlash… did the “plumbing” lift SPCX?… Jonathan Rose on what comes after the landing
It’s been a whipsaw session on Wall Street.
This morning, investors greeted blockbuster earnings from Micron (MU) by selling the stock, and the broader market followed.
The memory giant crushed estimates last night, with data-center revenue jumping 11-fold from a year earlier. But Wall Street worries that the highly cyclical memory trade is nearing its peak. Micron plans a massive increase in capital spending, heavily driven by new fab and clean room construction. Some investors see that as a sign of coming overcapacity.
Adding to the pressure, the 10-year Treasury yield hit 5.34% earlier in the session – its highest level since 2002.
But as I write in the early afternoon, buyers have stepped in. Treasury yields are dramatically off their highs, now down to 5.21%, and stocks – including Micron – have clawed back their losses and are in the green.
Whether that rebound holds into the close remains to be seen. Either way, the questions behind this morning’s selloff aren’t going away. So, we’ll dig deeper into both Micron and the Treasury market in a coming Digest.
In this issue, I want to focus on a different event happening today…
Last month, inside a lab in San Francisco, a team of technicians tried to shake a satellite apart to answer one question…
Could a handful of AI chips survive the bumpy trip into space?
Tucked inside that satellite were four of Alphabet Inc.’s (GOOGL) AI chips – the same kind that help power Google’s Gemini chatbot.
So, these technicians bolted the satellite to a table and shook it violently, simulating the pounding it would take on its way to space. According to The New York Times, which got an inside look at the project, nothing came loose.
Today, that satellite left Earth. At 2:32 p.m. ET, it launched aboard a SpaceX Technologies Corp. (SPCX) Falcon 9 rocket from Vandenberg Space Force Base in California.
Google calls the satellite “MVP.” And it’s the first real test of a project with a big goal: moving AI data centers off the planet and into orbit.
To be clear, MVP isn’t a data center in space. It has roughly the computing power of a single server, running on about as much electricity as a hair dryer. Google simply wants to learn whether its chips can handle the harsh conditions of orbit.
The reason why is simple: Earth is becoming a very difficult place to build AI.
Why Big Tech wants its data centers off the planet
AI requires enormous amounts of electricity. And the chips run so hot that data centers need huge cooling systems, often using millions of gallons of water, to keep them from frying.
In theory, space solves both problems. In the right orbit, there are no clouds and almost no nighttime, so solar panels collect far more energy. According to Google’s figures, we’re talking up to eight times more than solar panels on the ground. Meanwhile, the chips can shed their heat into the cold of space.
Space also avoids another problem here on Earth: angry neighbors.
Communities across the U.S. are fighting new data centers over rising power bills, water use, and noise – and they’re winning.
According to Data Center Watch, which tracks local opposition, at least 75 projects worth about $130 billion were blocked or delayed in the first three months of this year alone. That roughly matches the total for all of 2025. Another 45 projects worth $68 billion followed in the second quarter.
There are no community pushback or zoning issues 400 miles up, which is a point Elon Musk has been making for months. Here he is on X in August:
Orbital compute will be the only way to scale AI probably sometime in 2029 due to power availability & permitting problems on land.
And last week, reacting to Google’s launch announcement, he went further:
The amount of compute in space will obviously round up to 100% of all compute.
Now, Musk is likely getting ahead of himself, as he often does. Google’s research suggests space-based AI won’t make economic sense until around 2035. And even then, launch costs would need to fall to about $200 per kilogram, a fraction of today’s prices.
But Musk’s company SpaceX isn’t waiting around. It has filed with the Federal Communications Commission for a constellation of up to 1 million computing satellites, targeting its first orbital AI deployments for late 2027.
And so, today’s blastoff is a significant step in the space race. It marks the moment orbital computing moved from a slide in an investor presentation to actual (admittedly still experimental) hardware circling the planet.
Now, speaking of SPCX, let’s check in on a call we made here in the Digest earlier this summer…
Did the “plumbing” drive SPCX higher?
On July 13, I wrote that SPCX was probably going higher – but not because of fundamentals. The driver would be “plumbing.”
As restrictions on when insiders and early investors can sell their shares – lockups – expired and more shares became tradable, SpaceX’s weighting in the Nasdaq-100 would climb, forcing index funds to buy whether they wanted to or not.
That call proved early, and SPCX fell roughly 22% from where it traded when I made the call. But from there, the plumbing did its job. At the September rebalance, SpaceX’s Nasdaq-100 weight more than doubled, from 1.28% to 2.82%. That triggered an estimated $15.5 billion to $22 billion in forced index-fund buying.
As I write on Thursday, SPCX trades around $152, which means anyone buying in after our Digest call would be up about 10%, even with that 22% haircut.
So, should investors keep buying? Let’s weigh both sides…
Starting with the bull case, the fundamentals are improving faster than many expected.
When SpaceX merged with Musk’s AI company, xAI, earlier this year, it took over xAI’s Colossus data centers in Memphis. Those buildings are packed with hundreds of thousands of Nvidia Corp. (NVDA) chips. Nvidia sold the chips. SpaceX owns the buildings and rents out the horsepower – think of a landlord leasing a fully furnished office tower. Its tenants include Anthropic and, yes, Google.
That landlord business is changing SpaceX’s numbers. In its first quarterly report as a public company, SpaceX posted revenue of $7.8 billion, up 92% from a year earlier. Its net loss narrowed from $1 billion to $541 million.
Meanwhile, the plumbing tailwind isn’t finished. SpaceX’s float – the shares available to trade – is approaching one-third of the company. Once it crosses that line, Nasdaq’s rules stop discounting SpaceX’s size. That could mean another sizable weight increase – more forced index-fund buying – at the December rebalance.
On the bearish side, the stock trades at roughly 106 times sales. For context, the current price-to-sales ratio for the S&P 500 as a whole is roughly 3.7, and its long-term average is about 1.8. Excuse the pun, but this 106X price-to-sales valuation is stratospheric.
Plus, those same unlocks cut both ways. More are due October 9 and October 24, followed by a larger tranche after third-quarter earnings, and the full lockup ends December 8. Each new batch of tradable shares – after the initial forced buying – is stock that early SpaceX investors and employees are free to sell, and some of them will.
Overall, in the shorter term, I still expect index buying to help SPCX grind higher, with bumps along the way. Longer term, it’s earnings that will decide whether this is a temporary forced-buying story or a sustained winner.
But I will say that, right now, SPCX is richly priced as the toll-booth winner of the space economy. The question is whether it actually grows into that title.
Now, if you’d like a different way to play the growing space economy, our trading expert Jonathan Rose, founder of , just named three ideas.
What comes after the landing
While Google and SpaceX focus on data centers in orbit, a second space race is underway – the U.S. and China are competing to put a working base on the Moon.
SpaceX is part of that story, too. Its Starship is NASA’s pick to land astronauts. But Jonathan is watching what happens after the rockets touch down.
His reasoning is simple – NASA can’t build a Moon base alone, so it’s hiring private companies to handle launch, communications, power, and logistics.
Here’s Jonathan:
So, the question I’m asking isn’t “Who builds the rocket?” That answer is pretty easy: SpaceX, Blue Origin, and a few other big players.
The better question is “Who builds and operates everything a Moon base needs once someone’s actually living there?”
His first answer is Intuitive Machines Inc. (LUNR). It won a $180.4 million NASA mission to the lunar south pole region in March and carries a $1.8 billion backlog. But Jonathan sees a bigger role ahead:
I’m more interested in what it’s trying to become — the utility company for the Moon…
A lander gets you to the destination. A company that can also handle communications and navigation once you’re there takes on a role that doesn’t end after a single landing.
Second is Firefly Aerospace Inc. (FLY). That Austin-based company landed its Blue Ghost spacecraft on the Moon in March and now has six contracted lunar missions. Here’s Jonathan:
Landing once is an accomplishment. Building a dependable service around it is a business…
Jonathan notes that FLY has pulled back hard after the initial SpaceX enthusiasm wave – and that patient traders sometimes get a better entry that way.
Third is the lower-volatility option, L3Harris Technologies Inc. (LHX), which delivered the four RS-25 engines for NASA’s planned Artemis III mission. Back to Jonathan:
[It] wins regardless of which rocket startup succeeds. If you want exposure to the lunar buildout with less volatility than the smaller names, LHX is how you get there.
One caution: The smaller space stocks tend to rise and fall as a pack. Jonathan is fine with that. As he puts it, “you’re not really picking the winner of the Moon race. You’re buying the industry.”
By the way, Jonathan walks through setups like these live every market day at 11 a.m. ET in his free videos. He profiles trading ideas, walks through entries and exits, and hands out plenty of tickers in real time.
And just a heads-up…
Jonathan will be hosting a free live event later this month, the $10K to $100K Challenge, where he’ll show attendees the exact signals he uses to find trades – and how he sizes his risk. More details soon.
Zooming out
It’s easy to dismiss orbital data centers and Moon bases as science fiction. But look at where we are…
Google now has AI chips in orbit. SpaceX has filed for up to a million computing satellites. And NASA is writing checks for Moon-base infrastructure.
None of this means we’ll be taking weekend trips into space anytime soon, but the money is already moving.
For investors, the key is knowing what’s actually driving the stock you own: index mechanics, real earnings, or government contracts. Each can make you money. But they carry very different risks and reward very different levels of patience.
We’ll keep you updated on all of it here in the Digest.
Have a good evening,
Jeff Remsburg
(Disclosure: I own MU and GOOGL)
