Oracle has moved to raise $20 billion to keep paying for its artificial intelligence buildout, splitting the plan evenly between $10 billion of new shares and $10 billion of bonds. The company filed both offerings on Monday, 2 November, just days after telling investors it would cap further fundraising through fiscal 2030. The market answered at once: Oracle shares slipped in early trading, a reminder that even a planned raise can sting when investors are already nervous about AI spending.
What Oracle filed
The two filings, made with the United States Securities and Exchange Commission, describe a $10 billion common stock sale and a $10 billion sale of senior unsecured notes. Oracle said the proceeds would go to general corporate purposes, covering capital spending, debt repayment, acquisitions, investments, dividends and share buybacks. The filings do not tie the money to one project.
The debt side comes in six series of notes, maturing from 2029 to 2046. One series carries a floating rate tied to a benchmark plus an unset spread, while the fixed-rate series fall due in 2029, 2031, 2033, 2037 and 2046. Early pricing talks on the longest bond pointed to a spread of about 3.6 points above United States Treasuries.
Neither offering depends on the other, and the final terms are still open. The share count, sale price and interest rates have not been fixed. Oracle also agreed not to issue additional common stock for 60 days after filing without lead underwriters’ consent, subject to exceptions.
At Friday’s closing price of $130.67, the stock sale alone would mean roughly 76.5 million new shares, lifting the share count by about 2.5%, from about 3.02 billion to about 3.10 billion. That is a modest dilution on paper, but it arrives after a long slide in the shares, which is why holders noticed it.
The cap that came first
Monday’s filings did not arrive out of nowhere. They follow Oracle’s Investor Day in Las Vegas on 28 October, where management promised to limit additional fundraising to $20 billion through fiscal 2030, split evenly between equity and gross debt. This raise therefore looks like that capped plan being put into motion, not a new amount piled on top of it.
At the same event, Oracle said capital spending would peak in fiscal 2027 and fiscal 2028 as it builds more cloud capacity, then step down in fiscal 2029. Management framed that peak as the expensive middle of the buildout, not a permanent run rate.
Management also projected free cash flow turning positive in the second half of fiscal 2028 and reaching at least $30 billion in fiscal 2029, with total revenue targeted at $225 billion by fiscal 2030. Those targets are the company’s case that the spending eventually pays.
Why the market flinched
Investors still sold first. Oracle shares fell as much as 4% as markets opened on Monday, after closing down 6.37% on Friday at $130.67. The stock has already fallen by more than 30% this year, so a fresh equity sale lands on holders who have watched a long slide.
Dilution is part of the reaction, but so is timing. Earlier this year Oracle completed a $25 billion debt deal and a $20 billion stock offering, and had said at the time that it did not expect to issue more bonds in 2026.
The debt load behind the story is large. Oracle carried more than $125 billion in borrowings as of 31 August, against fiscal 2026 revenue of about $67 billion. That ratio is why every new financing gets read as a test of confidence, not just a routine funding step, and why a recent debt warning about AI financing frames the debate.
Our read: a planned raise, priced like a warning
The interesting tension is that Oracle did roughly what it said it would do, and the market still treated it as bad news. A cap announced on 28 October should have removed surprise, yet the first use of that cap still cost shareholders in early trading.
That gap says less about Oracle’s arithmetic than about the mood around AI infrastructure. Investors now want proof that data centres earn their keep, not just a schedule for building more of them. A promise of future cash flow no longer calms a market that has seen the bill grow.
Oracle is not alone in paying that price. Amazon recently put its AI hardware up for an $80 billion loan facility of its own, another sign that AI builders are funding hardware with borrowed money rather than current cash flow alone.
Microsoft ended a nearly three-month pause in big technology bond sales with a $25 billion offering of its own. The pattern is clear: the industry’s biggest builders are all going back to lenders and shareholders at the same time, so each new deal competes for patience.
There is a kinder reading, and Oracle will lean on it. Raising half the money as equity, rather than loading it all onto the balance sheet, keeps the debt cap intact and shares the cost with new shareholders, at least on paper.
The harsher reading is that issuing stock after a fall of more than 30% this year is costly funding, and it signals that cash flow cannot carry the buildout alone. Both readings can be true, which is why the next test is not the pricing of these notes but the fiscal 2029 cash flow Oracle promised.
Why It Matters
This raise matters beyond Oracle because it sets a template other AI builders may have to follow: a stated cap, a split between debt and equity, and a public cash-flow deadline. If Oracle hits its fiscal 2029 target, the cap will look like discipline.
If it misses, the cap will look like the first instalment of a much larger bill. The wider squeeze is already visible in buybacks, where the biggest technology companies returned far less cash to shareholders this year as AI spending swallowed their surpluses.
For customers and rivals, the signal is that Oracle intends to keep building through the downturn in its share price rather than throttling back. That keeps pressure on cloud rivals to match its capacity, and it keeps Oracle’s execution under a spotlight, with every quarter judged against the peak-spending path.
Background
Oracle’s cloud push has already reshaped its business, including opening its cloud to rival chips, led by Google’s tensor processors, as it chases AI workloads that once ran elsewhere. The $20 billion raise is the financing side of that same bet, and its success will be measured in data centres delivered, not in filings made.
Source: TipRanks report and TradingView report
0 Comments