Oracle is set to purchase $40 billion worth of Nvidia’s cutting-edge chips to support the first Stargate data center, a massive AI-focused facility being built in Abilene, Texas. The company will acquire 400,000 of Nvidia’s GB200 “superchips” to power advanced artificial intelligence systems, and plans to lease this computing power to OpenAI, according to a report by the Financial Times.
The Abilene data center, expected to be fully operational by mid-2026, will be one of the largest in the world, capable of delivering 1.2 gigawatts of power. This large-scale project is part of the broader Stargate initiative, first introduced by former President Donald Trump in January 2025. The program aims to develop up to $500 billion worth of AI data centers across the U.S., with the first ten located in Texas.
Oracle joins a group of equity partners in the Stargate venture, which includes OpenAI, SoftBank, and MGX. On the technology side, the initial partners are Oracle, Arm, Microsoft, Nvidia, and OpenAI. These companies will provide the hardware and software necessary to drive the project forward.
Recent updates show the Abilene site has secured $11.6 billion in funding commitments, bringing the total raised so far to $15 billion. These funds will support the expansion of the data center from two to eight buildings, signaling the project’s rapid scaling.
AI-specific data centers like Stargate differ from traditional facilities, as they require specialized infrastructure to manage high computational loads. Experts say that traditional data centers aren’t equipped to meet the demands of AI, which relies on dense configurations of GPUs and accelerators like Nvidia’s H100 and GB200 chips to handle large-scale matrix calculations.
This announcement follows earlier news that Elon Musk’s xAI and Nvidia joined a $30 billion AI Infrastructure Fund supported by BlackRock, Microsoft, and MGX. That fund aims to raise $100 billion to build and power next-generation AI facilities.
