summary：
It’s easy to get distracted by the headline numbers. Larry Ellison, the 81-year-old co-fou... It’s easy to get distracted by the headline numbers. Larry Ellison, the 81-year-old co-founder of Oracle, saw his net worth drop by about $38 billion in a few days this October. For most mortals, that’s an unimaginable sum. For Ellison, it was a rounding error on a fortune that had surged by over $140 billion in the preceding year, briefly crowning him the world’s richest person.
The financial media loves this kind of palace intrigue—billionaires swapping places on a leaderboard. But the real story isn’t about Ellison’s wallet. It’s about the market’s sudden, violent shudder of doubt. That $38 billion evaporation wasn't a random market fluctuation; it was a tremor originating from the colossal weight of promises Oracle has made to Wall Street. The recent stock pullback, from a peak of around $345 to the $280s, is the first sign that investors are beginning to question whether the company can actually deliver on one of the most audacious growth narratives in modern corporate history.
The Half-Trillion Dollar IOU
To understand the current volatility, you have to look at the numbers that fueled the ascent. Oracle’s stock ran up roughly 75% year-to-date by its September peak, a rally that left even the Magnificent Seven in the dust. This wasn’t driven by nostalgia for databases. It was fueled by a string of breathtaking AI and cloud announcements.
The figures are, frankly, staggering. In its last quarterly report, Oracle announced its backlog of Remaining Performance Obligations (RPO)—essentially, contracts signed but not yet fulfilled—had exploded to $455 billion. That’s a 359% year-over-year increase. CEO Safra Catz then casually mentioned on the earnings call that the RPO would likely exceed half a trillion dollars in the coming months. I’ve analyzed countless earnings reports, and a backlog growth of that magnitude for a company of this scale is a genuine anomaly. It’s the kind of number that forces you to re-read the filing to make sure you didn’t misplace a decimal.
The centerpiece of this backlog is a deal with OpenAI, reportedly worth around $300 billion over five years for Oracle’s cloud infrastructure. This, along with other massive contracts, prompted Oracle to raise its fiscal year 2030 revenue guidance from an already ambitious $144 billion to $166 billion. Analysts at Jefferies and Citigroup slapped price targets of $400 and $410 on the stock, calling the performance "truly historic."
This is the bull case in its purest form: a legacy tech giant that has successfully strapped itself to the AI rocket. The market, for most of 2025, bought it completely. But a backlog is just an IOU. The question is, can they actually build the machine to service it?
A Valuation Built on Faith
The recent stock correction suggests the market is shifting its focus from the promise to the execution. The drop on October 17th wiped out $24.1 billion of Ellison’s wealth; the subsequent slide on October 20th erased another $14 billion. To be precise, a combined paper loss of over $38 billion in two key trading sessions. This isn't just profit-taking; it’s the introduction of risk into the valuation equation.
Skeptics are finally finding their voice. Redburn Partners initiated coverage with a Sell rating and a $175 price target, calling the OpenAI deal "very risky." How, they ask, can a startup that has raised only $60-70 billion and runs at a loss guarantee a $300 billion contract? JPMorgan’s Mark Murphy pointed out that the implied long-term growth rate of 75% annually is "virtually unheard-of" for a company Oracle’s size. Morningstar’s analyst gave the forecast "late-90s… vibes," a polite way of referencing the dot-com bubble’s penchant for fantasy math.
The valuation metrics support this caution. At its peak, Oracle was trading at nearly 50 times forward earnings. That’s a multiple you expect from a hyper-growth startup, not a five-decade-old behemoth. The company's plan requires a capital expenditure cycle of unprecedented scale. During his keynote at Oracle AI World, Ellison detailed plans for just one new datacenter in Abilene, Texas: a 1,000-acre campus with eight buildings, consuming 1.2 gigawatts of power to run half a million Nvidia GPUs. He even spoke of building their own power plants.
This is where the narrative becomes a physical, logistical nightmare. Where does the money for this buildout come from? Oracle has been vague on the financing details. And who, exactly, is going to construct and power these digital cathedrals at the speed required to meet those contractual obligations? This is no longer a software problem; it’s a brutal industrial challenge involving concrete, power grids, and supply chains.
The market’s recent jitters reflect this dawning reality. A half-trillion-dollar backlog is like a giant engine. It has immense power, but it also requires an immense amount of fuel and an incredibly robust chassis to keep from tearing itself apart.
Vectorizing the Customer
So, how does Oracle plan to fund this Herculean effort? Part of the answer seems to lie in a strategy Ellison proudly detailed at his conference: vectorizing customer data. He explained that Oracle has taken all of its customer data and "vectorized it" to make it accessible for AI models to reason with.
The goal? To predict what products customers are likely to buy in the next six months, allowing Oracle’s sales teams to target them with precision. As one report, Larry Ellison's latest craze: Vectorizing all the customers, dryly noted, "Whether customers fancied being vectorized is apparently immaterial."
This isn't just a sales tool; it looks like the financial engine intended to underwrite the AI expansion. By using AI to systematically extract maximum revenue from its massive, entrenched customer base—perhaps identifying who might pay up after a "friendly software audit"—Oracle could be trying to generate the cash flow needed for its datacenter ambitions. It’s a closed loop: use AI to make money from existing customers to build the infrastructure to service new AI customers.
But what happens if that loop breaks? What if customers resist this new, more aggressive monetization? Or what if the predictions are simply not accurate enough to generate the needed windfall? It adds another layer of execution risk to an already teetering tower of promises. Standing on stage in Las Vegas, with an image of an Oracle-branded, AI-powered ambulance on the screen behind him, Ellison seemed to inhabit a world where vision alone could bend reality. He quipped that if someone told him years ago he’d be building billion-watt power plants, he’d have told them to get more rest. The audience laughed, but the question lingers: is this brilliant foresight or a fantastical fever dream?
A Bet on Execution, Not Vision
Ultimately, the story of Oracle’s stock is no longer about Larry Ellison’s vision for an AI-powered future. Wall Street has already bought the vision and priced it to perfection. The next chapter is about something far less glamorous: execution. It's about pouring concrete in Texas, securing transformers from a strained supply chain, and managing the single largest industrial buildout in the company's history. The half-trillion-dollar backlog is a liability until the infrastructure to service it exists. Every project delay, every financing gap, every power grid negotiation is now a direct threat to the stock’s lofty valuation. The market isn't questioning the dream anymore; it's starting to question the plumbing.