In the normal course of things, I wouldn’t dedicate special coverage to Oracle’s earnings release.

But the company’s a bellwether of sorts these days — a real-time indicator of the market’s patience (or lack thereof) for the most extreme version of the hyper-scaler business model, which entails sacrificing free cash flow and leveraging the balance sheet in pursuit of AI compute.

To say Oracle’s had a rough go of things since September’s outlandish post-earnings rally — which briefly made Larry Ellison the richest person on the planet — would be an understatement.

As the figure above reminds you, the shares bled out, falling 54% over five months from the September peak, erasing more than half a trillion in market cap.

Early last month, the company attempted to address the locus of investor concerns by saying it’d meet half of this year’s $50 billion fund-raising target with equity issuance. Thanks to that announcement, the company’s subsequent $25 billion debt sale was massively-oversubscribed (because suddenly, Oracle IG supply expectations were cut in half).

But, as I wrote here, 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.

With that in mind, I have to think it’ll come as a relief to markets that Oracle on Tuesday i) said it doesn’t expect to raise any incremental funds to support its backlog (RPO) and ii) reiterated its $50 billion capex plan (i.e., didn’t raise it but didn’t cut it either, with both alternatives being potentially perilous). The press release also noted that Oracle hasn’t initiated the follow-on equity offering from the above-mentioned financing plan.

Tuesday’s earnings update came on the heels of reports that Oracle broke off talks with OpenAI for data center expansion at a Stargate site in Texas, and separate reporting which tipped thousands of layoffs to help offset cash burn.

Addressing the job cuts, Oracle on Tuesday said “AI models for generating computer code have become so efficient that we have been restructuring our product development teams into smaller, more agile and productive groups.”

If that’s a little too euphemistic for you, they were more direct in the very next sentence: “This new AI Code Generation technology is enabling us to build more software in less time with fewer people.”

As to the actual numbers, they were pretty good. Sales of $17.2 billion counted as a beat, and rose 22%. Cloud infrastructure revenue — i.e., what everyone’s watching — rose a brisker-than-expected 84% to $4.9 billion.

As the chart shows, we’re still on the upswing where it counts. Analysts were looking for 78% growth from the infrastructure business.

The RPO metric mentioned above came in at $553 billion, up $30 billion from the prior quarter. Remember: The success or failure of this whole AI endeavor for Oracle depends very heavily on OpenAI, which accounts for around half of the company’s backlog. Some view that as problematic considering Sam Altman doesn’t yet have the revenue to support all of OpenAI’s commitments.

On Tuesday, Oracle said “some of the largest consumers of AI Cloud capacity have recently strengthened their financial positions quite substantially, enabl[ing] Oracle to comfortably meet and likely exceed our revenue growth rate forecast for fiscal year 2027 and beyond.” That sounds encouraging. The company raised its full-year sales guide for 2027 to $90 billion, ahead of analysts’ expectations.

All in all, this release looked “right down the middle” to me, which is to say Oracle might’ve threaded the needle. Capex for last quarter was a mile above the official consensus and the trailing four-quarter FCF metric Oracle includes obviously fell deeper into negative territory. But if you’re surprised by either of those outcomes, you haven’t been paying attention.