The reasons for the delays are mainly attributed to shortages in labor and materials, according to these sources, who requested anonymity while discussing confidential schedules.
Since signing a $300 billion contract this summer to provide the computational resources needed to train and operate OpenAI’s models, Oracle has been striving to fulfill its commitments. Despite the setbacks, the timelines for projects in the US are still ambitious, with sites anticipated to rank among the largest globally.
An Oracle spokesperson stated that the company is still confident in fulfilling its obligations and future growth plans. “There have been no delays on any sites crucial for meeting our contractual commitments, and all milestones are on track,” the spokesperson noted, adding that “the timelines for site selection and delivery were coordinated closely with OpenAI.” OpenAI chose not to comment.
As of 1:46 p.m. on Friday, Oracle’s shares were down approximately 3.4% in New York.
“We have ambitious and achievable goals for capacity delivery worldwide,” Oracle Co-Chief Executive Officer Clay Magouyrk mentioned during an earnings call this week. He added that the first data centre being developed for OpenAI in Abilene, Texas, is progressing well, with over 96,000 Nvidia Corp. chips already delivered.
