Databricks Exhausts Nvidia GPU Capacity in Multiple Regions, Driving Demand to Nebius Compute
Databricks nearly exhausted Nvidia GPU capacity in multiple regions as demand for hosted AI models outpaced hardware availability. The compute shortage is forcing cloud providers to ration supply and pushing startups to reroute workloads to alternative platforms.
Nebius revenue grew 622% in the first quarter to $399 million, prompting the company to raise its full-year annual recurring revenue guidance to $7 billion to $9 billion. CEO Arkady Volozh said demand from enterprise clients continues to exceed chip deliveries, leaving Nebius to sell out capacity after recent price hikes while hosting major accounts including Meta and Microsoft.
From the sources (6 posts)
@theinformationThe Nvidia GPU shortage is forcing cloud providers to ration capacity and startups to move workloads wherever chips are available. Even Nebius, which has benefited from AWS constraints, is selling out after raising prices. Read more: ht
@theinformationThe Nvidia GPU shortage is forcing cloud providers to ration capacity and startups to move workloads wherever chips are available. Even Nebius, which has benefited from AWS constraints, is selling out after raising prices. Read more: ht
@melvininvestsRT @MelvinInvests: Databricks just admitted it's running out of GPUs, and that shortage is exactly why neoclouds like Nebius are about to p…
@milkroadaiRT @MelvinInvests: Databricks just admitted it's running out of GPUs, and that shortage is exactly why neoclouds like Nebius are about to p…
@melvininvestsDatabricks just admitted it's running out of GPUs, and that shortage is exactly why neoclouds like Nebius are about to print money (Save this). Databricks CEO Ali Ghodsi said the company hosts open source models like Kimi for customers, an
@shanumathew93Probably nothing… 😉