← Back to BillCut Daily
Oracle Just Spent $28 Billion to Win a Race Nobody Can Define
Persona #3 · Vol: 10000
Oracle stock has been on a run that would make a crypto bro blush. Up roughly 60% over the past year, the 47-year-old database company suddenly trades like a hot AI startup. The reason? Cloud infrastructure. Specifically, the kind of cloud infrastructure that powers artificial intelligence models.
Here's the pitch: Oracle Cloud Infrastructure is now the preferred home for AI workloads, the company says. Massive compute clusters. Nvidia GPUs. Deals with OpenAI, xAI, and others. Chairman Larry Ellison has gone from legacy software baron to born-again AI visionary in about eighteen months.
The numbers, though, deserve a closer look.
Oracle recently reported cloud infrastructure revenue of about $3 billion for the quarter — growing fast, yes, but still a rounding error next to Amazon Web Services, Microsoft Azure, and Google Cloud. Combined, those three control roughly two-thirds of the global cloud market. Oracle's slice is somewhere around 2% to 3%, depending on whose estimate you trust.
So why is the stock acting like Oracle is about to eat everyone's lunch?
Because of the backlog. Oracle's remaining performance obligations — essentially contracted future revenue — have ballooned to over $130 billion. That's a staggering figure and the single biggest reason the stock has re-rated. Wall Street sees that number and extrapolates a decade of hypergrowth.
But a backlog is a promise, not a deposit. Contracts can be renegotiated. Customers can slow-walk deployments. And the AI spending boom, for all its hype, has not yet proven it generates durable profits for most of the companies doing the spending. OpenAI reportedly loses money at scale. xAI is burning cash. If the AI capex cycle cools — and every capex cycle eventually does — Oracle's backlog becomes a lot less impressive.
Then there's the capital problem. Oracle is spending tens of billions on data centers and GPUs. That money has to come from somewhere. The company's debt load has climbed past $80 billion. Free cash flow, once a hallmark of mature Oracle, is getting squeezed by the spending. You're not buying a stable dividend machine anymore. You're buying a leveraged bet on AI demand continuing to accelerate.
Who benefits from the hype? Ellison, obviously — his stake in Oracle has made him one of the richest people on earth. Nvidia benefits, because Oracle buys its chips. Investment banks benefit, because AI-themed equity stories generate trading volume and underwriting fees. And every cloud provider benefits from the narrative that AI requires infinite infrastructure, because it justifies their own spending.
What about you, the retail investor buying ORCL at 40-plus times forward earnings? You're paying a growth-stock multiple for a company that still gets most of its revenue from legacy database licenses and support contracts. That's not a crime. But it's not the same thing as owning Nvidia in 2016, and it's certainly not the same as owning a cheap, boring software company collecting maintenance fees.
The bull case isn't crazy. Oracle has real technology, real enterprise relationships, and a CEO in Safra Catz who knows how to run a tight ship. If AI infrastructure demand compounds for another five years, today's price will look cheap. But "if" is doing a lot of work in that sentence.
There's also a simpler read: Oracle spent years as an underdog in cloud, and the AI wave gave it a narrative it could never earn on merits alone. Markets love a comeback story. They love an underdog even more. And they absolutely love attaching a tired old company to whatever technology is currently printing headlines.
That doesn't mean the stock goes down tomorrow. Momentum can run longer than anyone expects. But the gap between Oracle's actual cloud position and its market valuation is wide enough to drive a data center through. When you buy ORCL here, you're not buying the database giant of 2010. You're buying a promise about 2030 — one backed by $80 billion in debt and a customer list that hasn't yet proven it can pay the bill.
Worth remembering the next time someone tells you Oracle is "the next AWS."
**The bottom line:** Oracle's AI story is real but wildly overpriced relative to its actual market position. The people selling you this narrative — Ellison, the banks, the chipmakers — get