"After a busy day at the Goldman Sachs tech conference earlier this week, I sat down with the firm's internet analyst Eric Sheridan to take stock. His main takeaway: "No one's slowing down." Despite spending massively on AI infrastructure, almost every tech exec told him AI demand is outstripping their ability to supply intelligence, he said. This was summed up by executives at CoreWeave, which builds and runs AI data centers."
"The room was packed for her talk. Even the overflow room was full, with many analysts and investors sitting on the floor. I've never seen so many loafers and crossed legs at the same time. OpenAI is on course to generate $13 billion in revenue this year, but the company is "still massively compute constrained," she said. That leads to tough decisions such as holding back new products, running some services intentionally slower, and having to choose which research projects get resources."
AI demand is surging across the industry and is increasingly outpacing providers' ability to supply compute and intelligence. Executives report unrelenting, recent upward inflections in demand and are investing heavily in AI data centers and cloud capacity. Current infrastructure growth is driven by paying customers and real revenue for AI services rather than just eyeballs. Large AI firms are generating substantial revenue yet remain compute constrained, prompting difficult choices to hold back products, slow services, and prioritize research. Rival companies are forming pragmatic partnerships to secure capacity. The compute shortage also raises the risk of AI disrupting software and SaaS businesses.
Read at Business Insider
Unable to calculate read time
Collection
[
|
...
]