Oracle’s down 50% from last year’s vertiginous highs. Did anyone else notice that?
I’m not a buyer here. Frankly, I’d sooner buy into PayPal trading at a nine-year low (and down 85% from the 2021 bubble highs) on the off chance they can turn things around with a new CEO than I would Oracle down by half on the notion they can borrow their way to outcompeting the deepest-pocketed companies on Earth in the AI spending race.
On February 1, Oracle indicated it’ll raise $25 billion this year through equity issuance. The announcement was meant to placate investors concerned about the company’s escalatory borrowing and falling free cash flow, which is set to turn negative and stay there for quite a while.
Prior to Oracle’s press release on its 2026 financing plan, the market suspected (read: worried) the company might sell as much as $60 billion in bonds this year. By tipping an intention to meet half of this year’s $50 billion fund-raising target with equity issuance, the company changed the optics overnight: Oracle engineered a scarcity premium at least compared to expectations, hence a record $129 billion in orders for a $25 billion issue. (Never mind they just sold $18 billion in bonds not five months ago.)
Some credit analysts quickly sounded the all-clear. And yet, Oracle’s shares ultimately closed lower on Monday. They were still falling on Tuesday morning, and as the figure below shows, they’re down every week but one this year.
The problem’s straightforward: There’s no way to fund new cloud infrastructure without irritating somebody. If you burn more cash, you’re playing into the market’s biggest fear. If you issue more debt, you’re playing into the market’s second-biggest fear. And while you can get creative with equity-linked issuance, selling shares ultimately dilutes stockholders.
I speak from personal experience when I say this isn’t conceptually different from someone who’s well-off and otherwise eminently creditworthy getting in over his skis trying to compete with people who are actually rich. I have to check myself on that front every, single day.
Generally speaking — and with the caveat that I’m excluding things like yachts, mansions and supercars, which I’m resigned to never owning — there isn’t anything I can’t go out and buy if I really want it during a given week. However, I have to be cognizant that if I insist on pushing the limits in order to impress people with whom I have no business competing, I’m out on a shaky limb. (A couple of weeks back, for example, I set about shopping for a new bag. I wanted the Hermès Allback messenger, but I bought the Goyard Coursier instead. Painful an admission as this surely was, the Allback’s out of my league.)
Although nothing’s out of Larry Ellison’s league personally (for laughs, note that his net worth, at an estimated $215 billion, is on par with Hermès’s entire market cap), there are things that are out of Oracle’s league as a company. In that regard, some investors worry he has Clay Magouyrk and Mike Sicilia in over their heads trying to keep up appearances in a spending competition with the biggest of the big guns.
I don’t like that. And neither does anybody else by the looks of things. There’s a lot of hedging going on with Oracle. I talked about that at length last year when the company’s CDS was front-page news seemingly every other session. Below, find Michael Hartnett’s favorite chart. (I don’t know if it’s actually his favorite, but he sure does roll it out a lot.)
Although Oracle’s CDS retreated sharply early this week following the press release on the funding mix (green box in the figure), it’s still quite wide, to put it politely.
“Big tech’s AI flip from asset-light to asset-heavy business models means bonds are the new regulators of AI capex,” Hartnett said in his latest, adding that bonds (which here means protection costs against an Oracle credit event) are “saying ‘slow it down.'”
But that’s just it: Big tech isn’t going to “slow it down.” The growth rate of AI capex will slow, but overall spending won’t. Already in 2026, bottom-up consensus estimates for hyper-scaler full-year capex are up more than $20 billion. They’ll surely rise further.
As the figure above, from Goldman’s Ben Snider, shows, consensus now expects $561 billion in capex from the hyper-scalers this year, or more than $110 billion each. That’s up nearly 40% YoY.
Is this a race Oracle belongs in? I don’t know, but on some days I doubt it. So do investors, apparently. The market cap loss over the last five months sums to nearly half a trillion.
It’s too late now, though. Oracle’s pot-committed. There’s no turning back. And remember: The success or failure of this endeavor depends heavily on OpenAI, which accounts for something like 55 cents out of every dollar of Oracle’s backlog.
That’d be a lot of concentration risk even if OpenAI had the revenue to make good on its commitments. Currently, it doesn’t.
Musk moving to merge SpaceX and XAi is an interesting development too. Seems like he is trying to go public before the mid-terms and any additional political risk, since he’s the ultimate government welfare Queen. I’d bet he plans to use part of the funds to pull Tesla into the new company, after the robot initiative tanks it.
I don’t think Oracle belongs, but I used to think that about msft, and they just started buying high quality companies and now they belong
Or 30 ounces of silver for the Goyard Coursier.
With a Timbuc2 from Amazon you’d have your bag and 28 ounces of silver to carry it in.
This is missing from the beginning of my previous reply:
$7,300-$8,100 for a Herme’s Allback? You can buy an ounce and a half of gold for that much.
In 1984, I bought my first new car for $4,600, and the dealer threw in a free pair of Raiders season tickets. The Raiders had just won the Super Bowl and moved to Los Angeles, and gold was selling for about $300 an ounce. Now a decent new car runs $27,000, gold is well north of $4K, and the a single Raiders season ticket runs about $3K (and the Raiders stink!) Man, times have really changed.
Season tickets on top of a new car? Now that’s a sale. Also, ya, that Allback is sexy.
I like the looks of the Coursier a lot more, but I suppose when it comes to such accessories, it’s not the look that matter most.
Yeah, it’s hard to beat the cool factor on a Goyard. It’s its own little club and the admission price is basically Louis + 10%, which isn’t bad considering you’re getting something that’s kinda neat / quirky. Hermès says something entirely different about you, or at least it does once you get beyond the trinkets and wallets and up into the real bags and clothes. That’s where you find the Audemars folks. If you don’t know what you’re getting into, it’s easy to embarrass yourself at Hermès. I strolled into the Hermès in lower Manhattan in 2014 not really understanding the difference between it and Louis / Dior / etc. and it was a humbling experience. You can screw up and embarrass yourself in Chrome Hearts too, but it’s a little easier to get out of there without buying anything, because you can always say “Well, it’s just not my aesthetic.” That won’t work at Hermès, because they look at you like, “Oh, so your aesthetic isn’t hand-crafted, understated beauty? Got it.”
I hate when you mention watches, because I’m guaranteed to spend the next hour online window shopping.
The Audemars catalogue always blows me away. There’s not a single one I don’t like. Any other watchmaker, there’s inevitably some pieces you can dismiss as gaudy or tacky or just a plain “old man watch” (a Rolex specialty). But not Audemars. Take this for example: https://www.audemarspiguet.com/com/en/watch-collection/code-1159/77410OR.OO.A344CR.01.html
Powder blue alligator band with pink gold case just should not work, but somehow it does.
I see you in a DeWitt, something in their Academia line. I like the Twenty-8-eight line myself. It’s a watch, Michael, what can it cost? Ten dollars?
You have an evident fondness for apparel so I gotta ask, favorite running shoes?
be extremely cautious longing paypal.
Oh, I didn’t actually buy it. Ha. I was just making a point about how skeptical I am of Oracle.
LOL. And before someone comments that I “don’t get it”, as in the difference between an Hermes/Goyard and my choice – I absolutely, 100%, do get it.
My backpack of choice is Osprey. I own 3: a 22L, a 26L (my most recent purchase, which I plan to carry for 10 days on the TMB this summer), and a 32L (I barely use this anymore because when I need a backpack for day to day hiking trips, I stay in hotels).
🙂
The Goyard looks great, for what it’s worth. I like the Navy one the best I think. Hope you customized it…