Do these 3 things before closing this tab:
1Scan for outdated or missing drivers - takes under a minute2Clear out junk files and repair common Windows errors3Fix the driver behind crashes, sound loss and screen glitchesSome links on this page are affiliate links: if you buy through them we may earn a commission, at no extra cost to you.
OpenAI and Oracle’s much-discussed “$30 billion deal” is not a $30 billion acquisition or equity investment. It is a reported, multiyear commitment for Oracle cloud and data-center capacity tied to OpenAI’s Stargate infrastructure program. Oracle disclosed that an unnamed cloud-services agreement was expected to generate more than $30 billion in annual revenue beginning in Oracle’s fiscal 2028. OpenAI subsequently confirmed a related agreement with Oracle to develop 4.5 gigawatts (GW) of additional U.S. capacity.
If the reported annual figure continued for five years, the arithmetic would be roughly $150 billion—not the often-repeated $300 billion—unless “more than $30 billion” is interpreted as a $30 billion-plus annual run rate over a longer or differently defined period. The companies have not published a complete contract value, term, minimum commitment or payment schedule.
The deal in one minute
| Item | What is publicly established |
|---|---|
| Customer | OpenAI, according to subsequent reporting identifying the customer behind Oracle’s disclosure |
| Provider | Oracle Cloud Infrastructure and related data-center capacity |
| Infrastructure | About 4.5 GW of additional U.S. Stargate capacity |
| Financial headline | Oracle said one cloud-services agreement should produce more than $30 billion in annual revenue from fiscal 2028 |
| Program | OpenAI’s Stargate AI-infrastructure initiative |
| Status | Multi-site development program; individual plans and sites have changed |
The safest description is: OpenAI agreed to obtain a very large amount of Oracle infrastructure, with the expected economics reported at more than $30 billion of Oracle revenue per year from fiscal 2028. The precise commercial contract remains private.
Quick wins for a faster PC:
Scan for outdated or missing drivers - takes under a minuteDriver Scan →Repair Windows errors before they cause bigger problemsFix Now →Fix the driver behind crashes, sound loss and screen glitchesFind Drivers →What Oracle actually disclosed
On June 30, 2025, Oracle filed an SEC Form 8-K saying that multiple cloud-services agreements included one expected to contribute more than $30 billion in annual revenue beginning in fiscal 2028. The filing did not name the customer and described the statement as forward-looking, subject to risks involving capacity, financing, hardware supply, construction and customer performance.
#1 Best Overall
That wording matters. It describes expected annual revenue, not a cash payment received on signing, a one-time investment or a guaranteed $30 billion transfer. Revenue would depend on when capacity is built and accepted, how the services are billed, and whether the agreement performs as expected. Read Oracle’s SEC filing.
How OpenAI connected the disclosure to Stargate
On July 22, 2025, OpenAI announced that it and Oracle would develop 4.5 GW of additional Stargate data-center capacity in the United States. OpenAI said the new capacity, combined with the initial Abilene, Texas site, put more than 5 GW under development and could support more than two million chips. OpenAI’s announcement did not repeat Oracle’s $30 billion figure, but reporting identified OpenAI as the customer behind Oracle’s previously unnamed agreement.
A gigawatt measures power capacity, not a fixed number of operational GPUs. The compute ultimately delivered depends on accelerator types, rack density, networking, cooling, power-utilization efficiency, grid availability and the mix of training and inference workloads. “More than two million chips” was a capacity-oriented projection, not evidence that two million chips were already installed and running.
Free tools Windows power users keep installed
One-click scans. No signup required.
Rank #2
Where Stargate fits
OpenAI, SoftBank, Oracle and MGX announced Stargate on January 21, 2025, as a company intended to invest up to $500 billion over four years in U.S. AI infrastructure. NVIDIA was identified as a technology partner. The $500 billion figure is Stargate’s project ambition, not the value of Oracle’s contract and not proof that the full amount had already been funded.
The arrangements should be separated:
- Stargate: the broad infrastructure initiative and its investment targets.
- Oracle’s cloud agreement: the commercial services relationship that Oracle described in its SEC filing.
- The 4.5-GW plan: the additional U.S. capacity OpenAI publicly linked to Oracle.
- Individual sites: campuses, developers, power connections and construction projects that can be revised independently.
- OpenAI’s broader cloud strategy: a continuing use of multiple providers, including Microsoft.
Oracle’s March 2025 earnings release had already listed OpenAI among Oracle Cloud customers and referred to a planned Stargate contract. Oracle’s release.
Why OpenAI wants Oracle capacity
Frontier-model training, post-training, evaluation and inference all require large accelerator clusters. OpenAI also needs capacity for ChatGPT traffic, API customers, coding products and future enterprise and agentic applications. A second major infrastructure provider can improve availability, geographic flexibility and negotiating leverage while reducing dependence on a single supplier.
Rank #3
This is not a clean break with Microsoft. OpenAI explicitly said Microsoft would continue providing cloud services, including through Stargate. Oracle is therefore best understood as an additional large-scale infrastructure partner, not a replacement for Azure.
Why Oracle wants the contract
A customer capable of generating more than $30 billion in annual cloud revenue would materially increase Oracle’s AI-infrastructure scale and provide a high-profile reference account. Oracle is competing with Microsoft Azure, Amazon Web Services, Google Cloud, specialized GPU clouds and data-center operators.
Long-term demand can help justify the capital required for land, buildings, electrical interconnection, generation, GPUs, networking and cooling. But it also creates exposure: Oracle may need to spend heavily before revenue arrives, and a change in OpenAI’s requirements could affect utilization, margins, financing and debt.
What the headline does—and does not—tell you
- It does not establish a $30 billion lump-sum payment. The disclosed figure is annual expected revenue.
- It does not prove a $300 billion contract. That number is an extrapolation sometimes made by coverage; the public filing does not state it.
- It does not prove all 4.5 GW is operational. OpenAI described capacity under development.
- It does not identify every site, owner or financier. Locations, funding arrangements and hardware allocations were not fully disclosed.
- It does not show that Oracle finances everything. The public announcements do not provide a complete project-finance model.
- It does not show an unconditional minimum purchase. Minimums, cancellation rights, pricing and delivery milestones remain private.
The physical and financial risks
Power and construction
A 4.5-GW build-out is an energy and civil-infrastructure project as much as a cloud sale. Interconnection queues, transmission limits, generation availability, permitting, water and cooling requirements, local opposition and construction delays can all change the schedule or cost.
Hardware depreciation
Accelerators can become economically outdated before a facility’s financing or service term ends. The project must keep useful performance as GPU generations, networking standards and cooling designs change.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
Demand and financing
OpenAI’s compute demand may grow rapidly, but efficiency improvements, model changes, competition or product decisions could alter the requirement. OpenAI also has to fund or finance the resulting commitments; the public material does not show that the full annual amount would come from current operating cash flow.
Best Value
Customer concentration
If Oracle’s AI growth becomes dependent on a small number of customers, a delay or renegotiation could affect revenue timing and data-center utilization. Oracle’s SEC filing expressly warns about execution and non-performance risks.
What changed after the 2025 announcement?
The infrastructure plan has evolved. In March 2026, Reuters reported that Oracle and OpenAI dropped plans to expand the flagship Abilene, Texas site after financing negotiations and changing requirements. That report did not establish cancellation of the Oracle relationship or Stargate as a whole; the existing Abilene campus remained a major Stargate site. Separate reporting said Microsoft planned additional facilities nearby, illustrating how a regional data-center build-out can continue even when a particular expansion changes. Reuters’ report.
OpenAI and partners also announced five additional U.S. Stargate sites in October 2025, later identifying the Midwest project as a Wisconsin site developed by Oracle with Vantage Data Centers. Announced capacity should therefore be read as a changing portfolio of projects, not a single fixed facility.
The Tool Desk
Outbyte PC Repair FREERepair Windows errors before they cause bigger problemsFix Now →Outbyte Driver Updater FREEScan for outdated or missing drivers - takes under a minuteDriver Scan →What it means for enterprise buyers
The Oracle–OpenAI agreement is not a product that ordinary customers can purchase directly. It signals more infrastructure behind OpenAI services and expands enterprise deployment options: in 2026, OpenAI announced that customers could access its models and Codex through Oracle Cloud commitments. OpenAI’s Oracle Cloud announcement.
Organizations choosing a platform should compare model access, deployment type, data-retention and training policies, regional controls, identity integration, quotas or reserved capacity, latency, egress, support and portability. Oracle may be particularly attractive to enterprises already standardized on Oracle databases and applications. Azure OpenAI suits Microsoft-centered identity, security and compliance environments; AWS Bedrock and Google Vertex AI suit buyers seeking broad multi-model platforms. Current prices and service terms should be checked on the vendors’ official pages.
Bottom line
The important fact is the structure, not the slogan. OpenAI’s Oracle relationship is a large, multiyear cloud and data-center capacity commitment linked to Stargate. Oracle disclosed expected annual revenue of more than $30 billion from fiscal 2028, while OpenAI confirmed 4.5 GW of additional U.S. capacity. The contract’s full value, term, financing, sites, delivery schedule and minimum obligations remain undisclosed, and later site revisions show why the headline should not be treated as a guaranteed $30 billion payment—or as a completed $300 billion transaction.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
