CoreWeave's argument that Nvidia A100 GPUs can remain economically useful through 2029 gained fresh backing from Nvidia CEO Jensen Huang, who said the 2020-era chips should still be able to handle compute workloads at the end of the decade after the cloud provider extended an A100 leasing contract to 2029. Huang said Nvidia's value rests not only on chips but also on CUDA, which lets developers and Nvidia engineers keep improving software performance across Ampere, Hopper and Blackwell systems, making compute more interchangeable across generations and helping raise utilization and asset life. The comments strengthen CoreWeave's broader case that older Nvidia hardware still commands demand: CFO Nitin Agrawal has said the older-generation fleet is essentially fully leased and that average selling prices are at or above year-earlier levels, while CEO Mike Intrator said some H100 capacity was rebooked at 95% of its original price. CoreWeave reported Q2 2026 revenue of $2.575 billion, up 112% year over year, with a revenue backlog of about $104 billion as of June 30, as it argues that legacy AI systems remain valuable partly because many existing data centers cannot yet support newer liquid-cooled, high-power racks.