AUSTIN, Texas, November 2, 2026 — Oracle plans to raise $20 billion through a combination of stock and bond sales as it continues expanding the cloud infrastructure used for artificial-intelligence workloads. Regulatory filings show that the proposed financing includes a bond offering split into six tranches, with maturities from 2029 to 2046, alongside an equity sale whose final size and pricing have not yet been set.

The plan arrives as investors scrutinise how quickly spending on AI data centres will translate into revenue and cash flow. Oracle shares fell nearly 6% on Monday, extending a decline of about 33% through Friday since the start of the year. The share-price reaction does not by itself establish that the company’s strategy will fail, but it reflects concern about the amount of capital required to build infrastructure before the resulting services generate cash.

Oracle said the proceeds may be used for general corporate purposes, including capital expenditure, debt repayment, acquisitions, dividends and share buybacks. The company has not fixed the number of shares to be sold or the final pricing terms for the bonds. The offering therefore represents a financing plan rather than a completed $20 billion raise, and the amount ultimately secured through each part could depend on market conditions and the final documents.

The bond component includes six tranches maturing between 2029 and 2046, according to Reuters. One tranche carries a floating rate, while five have fixed rates. Pricing, coupons and the number of shares had not been set in the reports reviewed. Investors will assess the final terms against Oracle’s existing borrowing costs, expected cash generation and the risks attached to its growing data-centre business.

Oracle has been expanding its cloud infrastructure to serve customers seeking computing capacity for AI training, inference and enterprise software. Large AI workloads require more than processors: they also depend on data-centre buildings, networking, storage, cooling systems and reliable electricity. Those assets must often be funded and installed before all of the capacity is generating revenue, creating a gap between capital spending and the cash returns a company hopes to earn.

Also read: Banks Arrange $63 Billion AI Compute Financing for Anthropic

The scale of Oracle’s borrowing makes that timing important. The company reported $125.34 billion in notes payable and other borrowings as of August 31, while revenue for its 2026 fiscal year was $67.36 billion. The figures measure different things and should not be compared as if revenue were cash available to repay debt, but they show why investors are paying close attention to leverage, financing costs and the pace at which new infrastructure becomes productive.

Oracle has said that capital spending should peak in the current fiscal year and the next, with free cash flow expected to turn positive in the second half of fiscal 2028. That is management guidance, not a guaranteed outcome. The forecast depends on the timing of customer deployments, the cost of building and operating data centres, the mix of contracts and the company’s ability to convert booked demand into cash receipts.

The company has also said it intends to limit additional capital-market financing for AI through fiscal 2030. The latest plan places that commitment in focus: investors will want to know how the new $20 billion raise fits within the stated limit and what portion is intended to support near-term spending rather than other corporate purposes. The public filings and company statements reviewed do not establish that every dollar raised will be earmarked exclusively for AI infrastructure.

Oracle’s strategy is linked to major customers that need large amounts of computing capacity, including AI developers and cloud users. Long-term contracts can provide visibility into future demand, but they can also require Oracle to spend heavily before the services are delivered. The economics depend on the price paid for chips and equipment, the cost of power and facilities, the time taken to bring capacity online and the margins the company can earn once customers use it.

That creates a difference between contracted revenue and cash generation. A contract may support a strong backlog, but the company still needs to finance construction, install equipment and meet service obligations. If a project is delayed, the cost of capital can accumulate before revenue is recognised. If customer demand grows faster than expected, Oracle may need to accelerate spending to provide capacity. Either scenario can put pressure on near-term cash flow.

Oracle’s borrowing plans are part of a wider shift across the technology sector. Large cloud companies are using debt and equity markets to finance AI infrastructure rather than relying entirely on operating cash flow. Alphabet launched a €6 billion euro-denominated bond sale on Monday, while other companies have disclosed large spending plans for data centres and advanced computing. The transactions are not identical, but together they show how financing costs and investor appetite have become central to the AI expansion.

The immediate question is how the offering is priced. Fixed-rate bond coupons will determine the cost of borrowing over the relevant maturities, while the floating-rate tranche will expose Oracle to changes in benchmark rates. The equity component would raise capital without requiring repayment of principal, but it would dilute existing shareholders. The balance between the two can affect both the company’s future interest expense and the ownership of its earnings.

Also read: Persistent Systems Plans ₹8,000 Crore Bond Sale for Nagarro Deal

Investors will also consider how the new financing interacts with Oracle’s existing obligations. Debt is only one part of the funding picture; leases, construction commitments and equipment purchases can create substantial future cash requirements even when they do not appear as current borrowing. The company’s disclosures on data-centre leases and infrastructure spending will therefore remain important alongside the headline bond amount.

The quality and timing of customer demand are equally important. AI infrastructure is being built to support workloads that are expected to grow, but the market is still developing. Customers may change suppliers, adjust the size of their deployments or seek lower prices as hardware and model efficiency improve. If demand is delayed, Oracle could have capacity that is expensive to maintain before it reaches planned utilisation. If demand holds up, the company could benefit from long-term contracts and a larger recurring cloud business.

Oracle’s exposure to OpenAI and other AI customers has become a focus for investors because it links infrastructure spending to the financing and growth prospects of the companies using that capacity. The relationship can be commercially attractive if customers pay for the resources they reserve and use. It can also concentrate risk if a small number of customers account for a large share of expected demand or if their own funding needs change. The latest financing announcement does not disclose a new breakdown of customer-level exposure.

The company’s guidance for positive free cash flow in the second half of fiscal 2028 gives investors a timetable against which to measure execution. Progress will depend on whether operating cash flow grows enough to cover capital spending and other needs. Investors will also look for signs that spending is beginning to moderate as planned, that new facilities are coming online on schedule and that cloud revenue and margins are keeping pace with the investment.

Credit investors will assess the transaction differently from equity investors. Bondholders focus on the company’s capacity to service debt and maintain access to funding across economic and market cycles. Shareholders may be more willing to accept near-term dilution or capital spending if they believe the resulting business will grow substantially. A financing structure that looks manageable under one set of assumptions can become more difficult if interest costs rise, projects are delayed or the returns from AI services take longer to arrive.

The offering’s final documents will clarify the maturities, interest rates, covenants and equity terms. Until then, the headline $20 billion figure should be treated as the amount Oracle plans to raise, not money already received. The company may use proceeds for several corporate purposes, and the allocation could differ from the broad categories described in its filings.

The next milestones are the completion of the securities offering, disclosure of final pricing and share terms, and Oracle’s subsequent reporting on capital spending and cash flow. Any change to its financing outlook or expected timing for positive free cash flow will be closely watched. For now, Oracle is seeking capital to keep expanding its AI and cloud business while investors demand clearer evidence that the infrastructure build-out can generate sustainable returns.