Oracle Pours $10B Into Data Center Expansion
Tech giant will build 100 new data hubs, expand capacity at 66 existing data centers.
Oracle is investing $10B to expand its data center footprint as the database giant embraces cloud-based infrastructure and software services that have become its dominant business segment.
In an earnings report for the three-month period that ended on Feb. 29, Oracle CEO Safra Catz disclosed that the Austin, TX-based company plans to spend $10B to build out its data center capacity in the next year to service burgeoning demand driven by the explosive growth of GenAI.
“This quarter marks the first time our total cloud revenue is more than our total license support revenue,” Catz said during an earnings call. “We have crossed over.”
The company reported $5.1B in cloud revenue, a 24% year over year increase, including a 49% jump in infrastructure cloud services revenue. Oracle’s license support revenue was $4.9B.
Oracle, a top player in the ERP and database sector, is aiming for the top tier of hyperscale operations with its cloud data center buildout. The company operates 68 customer-facing cloud regions, two of which became operational this year, and 47 public cloud regions with another eight in the building phase, Catz said.
Oracle is planning to build 100 new data centers, including 20 that will service an expanding alliance with Microsoft involving 12 public cloud regions that interconnect with Microsoft’s Azure platform.
Oracle also is planning to expand capacity at 66 existing data centers to accommodate exponentially expanding demand for AI workloads and enterprise migrations. In an earnings call at the end of December, Larry Ellison, the Oracle chair and CTO, said the company had “billions of dollars more in contracted demand than we currently can supply.”
In addition to the partner regions with Microsoft involving 20 data centers, Oracle is building public regions based on direct customer demand that will need the capacity of up to 80 new hyperscale data processing facilities, Ellison said.
Oracle also is supplying Elon Musk’s xAI startup with Nvidia processors to train its Grok model. During the December call, Ellison said “there are other multicloud agreements that are being signed.”
The scale of the facilities that are under development to support GenAI will be massive, Oracle’s CTO said. “We’re building an AI data center in the United States where you can park eight Boeing 747s nose to tail,” Ellison said, referring to a data center campus under development in Salt Lake City.
Demand for cloud infrastructure and generative AI services is exploding, with total expenditures for public cloud buildouts expected to hit $700B this year, according to a Gartner forecast.