🚀 SpaceX Beats. Stock Falls 8%

Ready to transform artificial intelligence from a buzzword into your personal revenue generator?

HubSpot’s groundbreaking guide "200+ AI-Powered Income Ideas" is your gateway to financial innovation in the digital age.

Inside you'll discover:

Download your guide today and unlock a future where artificial intelligence powers your success. Your next income stream is waiting.

Get The Free Guide

SpaceX beat on revenue, beat on earnings, and the stock fell nearly 8% anyway.

Why? $18.4 billion of capex in a single quarter.

Up from $10.1 billion last quarter. More than six times the $2.8 billion it spent a year ago. Roughly $200 million a day, weekends included.

$15.8 billion of that went straight into AI, expanding compute capacity and building infrastructure for new cloud deals. The rest went to Starlink satellites and Starship.

Which is a shame, because the quarter itself was excellent:

Connectivity is still the engine. Revenue up 66% to $4.29 billion, Starlink subscribers doubled to 12 million, operating income up 79% to $1.66 billion. Enterprise and government more than doubled on airline deals, carrier partnerships and over $6 billion of multi-year Starshield contracts.

AI grew fastest. Revenue up 247% to $2.56 billion on $14.1 billion of signed cloud services agreements plus Grok and X subscriptions. Positive EBITDA of $1.15 billion, alongside a $1.26 billion operating loss.

Space is the expensive child. Revenue of $962 million (up 29%) and a $542 million operating loss as Starship R&D climbed. Starship V3 flew its first suborbital mission in the quarter, and Flight 13 deployed production satellites just after it closed.

More than $15 billion has gone into Starship so far, on a bet it eventually cuts launch costs by up to 99%.

CFO Bret Johnsen's read: growth accelerated across every segment, with margins expanding on the new AI compute deals.

On the same call, Gwynne Shotwell pointed at America's wireless industry. And made an offer nobody asked for.

AT&T, Verizon and T-Mobile are roughly a $600 billion a year market, she said. She expects SpaceX to take "quite a few" of their customers, because she thinks the service will be better.

All three slipped in after-hours trading. The market clearly believes her.

Starlink already does direct-to-cell, which fixes dead zones. SpaceX now intends to build the terrestrial component too, real ground hardware paired with the satellites to run a proper mobile service.

It has the airwaves for it. The FCC approved the transfer of EchoStar's spectrum licences to SpaceX during the quarter. Add carrier tie-ups with SoftBank, NTT Docomo and Spark NZ, and inflight deals with American, Southwest, Virgin Atlantic, Iberia and Aer Lingus.

This was the first look inside the books since June's IPO, which raised about $75 billion at a roughly $1.1 trillion valuation, the largest listing in history.

At that price, "good" gets marked harshly. A beat isn't enough when the price demands perfection."

One date to circle: August 6. The first tranche of locked-up shares becomes sellable, which can mean fresh selling pressure now and a deeper, more liquid market later.

The cushion is enormous. $100 billion in cash and a $47.5 billion backlog. SpaceX can afford to keep spending like this for years, whether or not shareholders enjoy watching it.

Pricing on these AI IPOs gets announced once. Hear about it late and the entry window has already moved — along with the easy gains. The free briefing covers the timeline and optional text alerts the moment pricing drops, so you're not the last to know.

Get the Free Briefing

Apollo's chief economist just called time on the 60/40 portfolio.

That's the split your pension uses. Your dad's IRA uses it. Basically every "balanced" fund on the planet uses it.

Torsten Sløk says the thing that made it work has stopped working. Here's his case 👇

The old deal was simple. Economy good, stocks up. Economy bad, stocks down. Predictable enough to build a retirement around.

Sløk argues equity performance is no longer tied to the broader business cycle. Returns have piled into a narrow set of artificial intelligence leaders.

So you think you own "the market." You mostly own one trade wearing a market costume.

And that trade is showing signs of losing momentum.

Bonds were meant to be the airbag. Stocks crash, bonds rally, you survive to invest another day.

Sløk says bond returns are now being dictated by fiscal pressures and rising government debt instead of normal economic fluctuations.

The number doing the damage: U.S. government debt is projected to climb toward 175% of GDP.

In simple terms: for every dollar the American economy produces in a year, there's about $1.75 of debt sitting on top of it.

Bond investors aren't reading jobs reports for that. They're reading the government's credit card statement.

Here's the part Sløk actually loses sleep over. A reversal in the AI trade, or the market suddenly deciding it cares about fiscal deficits.

Either one hits stocks and bonds at the same time. Your hedge and the thing it was hedging go down together.

Which kills the whole point, because catching you when the other side fell was the 60/40's only job.

So it's worth looking under the hood of anything labelled "balanced." If both halves now answer to the same two questions, the pie chart is selling you a diversification the market can't honour.

The 60/40 didn't get a funeral. It just stopped showing up for work.

Wall Street now thinks Oracle is a bigger default risk than it was in 2008.

Oracle's $ORCL ( ▲ 2.74% ) five-year credit default swap spread just hit the highest level in company history. It blew straight past the peak of the global financial crisis.

It's insurance on debt. You pay a premium, and if the borrower stops paying, you get covered.

The "spread" is what that insurance costs. When it climbs, investors are charging more to take the risk that Oracle doesn't pay them back.

Right now they're charging more than they did when the entire banking system was on fire.

So what changed?

Oracle went all in on AI.

Multibillion-dollar spending on cloud capacity and data centres, built to catch the surge in demand for AI computing. Nobody's arguing with the demand.

The argument is over the funding. Oracle has leaned increasingly on debt to pay for the build-out.

Capex climbs. Borrowing climbs. And bondholders start doing maths they'd rather not do.

Every new data centre is a bet that the AI revenue shows up before the interest payments get uncomfortable. The CDS market ha decided that bet is less of a sure thing than it looked.

ORCL is down 27% year to date and nearly 44% over the past 12 months.

$10k in Oracle a year ago is worth roughly $5,600 now. The AI trade giveth, then it posts you an invoice.

What's striking is the direction of travel. Investors spent 2024 rewarding anyone who so much as whispered "AI capex." Now they're pricing the same spending as a liability.

The credit market and the stock market are telling the same story here, which doesn't happen that often. Usually one of them is asleep.

The real question this raises: what happens to every other company funding AI infrastructure with borrowed money? Oracle is the first big name where the bond crowd has openly not liked it. It's unlikely to be the last one they look at.

Worth watching from here: whether the CDS spread keeps widening or settles, and whether Oracle's next capex update leans harder on debt or starts showing restraint. Those two data points will tell you more about the AI build-out than any earnings call soundbite.

The bond market rarely shouts. When it does, it's usually worth turning the volume up.