Oracle CEO Magouyrk: 'Of course' OpenAI can pay $60 billion per year for cloud infrastructure (www.cnbc.com)

Oracle co-CEO Clay Magouyrk told CNBC he’s confident OpenAI can shoulder massive cloud bills—saying “of course” the company could pay roughly $60 billion for a year’s worth of infrastructure—after the two firms agreed in July to a five‑year partnership worth more than $300 billion. His comments come as OpenAI reports explosive user growth (about 800 million weekly active users) but also a $5 billion net loss in 2024, highlighting a tension between rapid scale and hefty operating costs. Oracle paired CEO Mike Sicilia noted Oracle is already embedding OpenAI models into vertical products, such as a patient portal for Cerner electronic health records, signaling stronger enterprise integration. For the AI/ML community this underscores two trends: first, commercial cloud and enterprise deals are becoming a critical revenue mechanism to fund model scale; second, infrastructure is diversifying beyond public GPU rentals. OpenAI currently runs models on Nvidia GPUs via Oracle and other cloud partners (CoreWeave, Google, Microsoft) while also developing custom Broadcom-built AI chips, with Broadcom/OpenAI targeting a 10‑gigawatt deployment. That magnitude raises energy and supply-chain implications even as it accelerates specialized hardware, multi‑cloud consumption, and verticalized AI deployments—factors that will shape model economics, latency, and enterprise adoption going forward.