日本語
한국어
Tiếng Việt
简体中文
The tech world’s biggest mystery of the past month has been unraveled: OpenAI is the unnamed giant behind Oracle’s colossal $30 billion per year data center services agreement. This groundbreaking deal, initially disclosed by Oracle in an SEC filing last month, has now been officially confirmed by OpenAI CEO Sam Altman, as reported by The Wall Street Journal and validated by Altman himself via an X post and a company blog entry.
Oracle first sent ripples through the market on June 30 when it announced a cloud computing contract of unprecedented scale, projected to generate an astonishing $30 billion in annual revenue. The identity of this client remained a closely guarded secret, fueling intense speculation across the industry. This news alone propelled Oracle’s stock to an all-time high, cementing co-founder and CTO Larry Ellison’s position as the world’s second-richest individual, according to Bloomberg.
The sheer magnitude of the deal immediately raised questions about which enterprise could possibly require such immense data center resources. For context, Oracle’s entire cloud services revenue for its fiscal year 2025, from all customers combined, totaled $24.5 billion. The new single contract with OpenAI dwarfs this figure, underscoring the extraordinary demands of advanced artificial intelligence development.
Powering Stargate: A $500 Billion Vision
OpenAI has clarified that this monumental commitment to Oracle is a cornerstone of the ambitious Stargate project. Announced in January, Stargate is a proposed $500 billion initiative by OpenAI, Oracle, and SoftBank aimed at building the next generation of AI supercomputing infrastructure. The $30 billion annual Oracle deal specifically covers 4.5 gigawatts of data center capacity for Stargate, though it appears SoftBank is not directly involved in this particular financial agreement.
To put 4.5 gigawatts into perspective, The Wall Street Journal notes this is equivalent to the power output of two Hoover Dams, sufficient energy to supply approximately four million homes. This immense power is crucial for training and running sophisticated large language models (LLMs) and other advanced AI applications.
A Massive Undertaking with Significant Investment
While a landmark win for Oracle, this partnership is far from a straightforward transaction. Both OpenAI and Oracle face the significant challenge and considerable expense of constructing these hyperscale AI data centers. The initial Stargate I site is slated for Abilene, Texas, requiring immense capital expenditure and energy resources.
Oracle’s financial commitments reflect this endeavor. CEO Safra Catz reported that Oracle spent $21.2 billion on capital expenditures in its last fiscal year and anticipates investing another $25 billion this year. This nearly $50 billion outlay over two years is primarily directed towards expanding its data center footprint, a necessity to meet not only OpenAI’s demands but also the needs of Oracle’s existing and growing customer base.
OpenAI’s Strategic Bet on Infrastructure
An intriguing aspect of this deal lies in its financial implications for OpenAI. Just last month, Sam Altman stated that OpenAI recently achieved $10 billion in annual recurring revenue, a substantial increase from roughly $5.5 billion the previous year. This $30 billion annual commitment to Oracle represents triple OpenAI’s current revenue, and it doesn’t account for the company’s other significant operational expenses, including existing data center obligations.
This massive investment underscores OpenAI’s strategic imperative to secure the vast computing power required to advance artificial intelligence. It highlights a critical phase in the AI race, where access to state-of-the-art digital infrastructure is as crucial as algorithmic innovation. The partnership with Oracle positions both companies at the forefront of building the fundamental layers for the future of AI.