Oracle (ORCL, Financial) plans to invest approximately $400 billion in NVIDIA's (NVDA) new AI chip, GB200, to support the construction of a large-scale AI data center for OpenAI in Abilene, Texas. This project is the first phase of the Stargate initiative led by OpenAI and SoftBank, aiming to create one of the world's largest AI data centers with a power requirement of 1.2GW.
Oracle intends to purchase around 400,000 superchips from NVIDIA and lease the computing power to OpenAI. The data center, owned by Crusoe and Blue Owl Capital, involves a total investment of $15 billion, with funding from JPMorgan (JPM) and shareholders. Oracle has entered a 15-year lease agreement, and the center is expected to be fully operational by mid-2026.
This collaboration marks OpenAI's shift from reliance on Microsoft (MSFT), as it transitions to Oracle's support due to Microsoft Azure's capacity constraints. The data center will serve as a crucial platform for OpenAI's AI model training and deployment, potentially rivaling Elon Musk's Colossus AI supercluster in Tennessee.
The Stargate project has secured $1.8 billion each from OpenAI and SoftBank, with Oracle and Abu Dhabi's MGX contributing $7 billion each. While Stargate hasn't directly invested in data centers yet, it plans global expansion, including a partnership with Abu Dhabi's G42 to establish a 5GW data center.
