The completion time of some of Oracle's data center projects for artificial intelligence model developer OpenAI has been postponed from 2027 to 2028, mainly due to shortages of labor and building materials, according to people familiar with the matter who requested anonymity to discuss internal schedules.
Still, the timeline for these U.S.-based projects is seen as radical, with plans to build one of the largest data center clusters in the world.
Oracle is moving forward with a long-term contract totaling approximately $300 billion signed this summer to provide OpenAI with the large-scale computing power needed to train and run its models. Affected by the delay news, Oracle's stock price fell as much as 6.5%. As of 11:03 a.m. local time in New York on Friday, the stock was trading at $188.26, a decrease of about 5.3%.
Oracle and OpenAI both declined to comment on the project delay. Oracle co-CEO Clay Magouyrk said on an earnings call this week that the company has set "ambitious but achievable" capacity delivery goals globally. He also pointed out that the construction of OpenAI's first data center in Abilene, Texas, is progressing smoothly, and more than 96,000 Nvidia chips have been delivered on site to support related artificial intelligence computing power needs.
Related articles:
Oracle CEO: "Of course" OpenAI can afford $60 billion in annual costs