NetApp and Nvidia redesign storage for AI factories

NetApp and Nvidia are co-engineering storage infrastructure to meet the demands of AI factories. NetApp's Novus architecture separates metadata and data functions so each can scale independently, preventing small transactional operations from blocking large data transfers and keeping GPUs fully utilized. As AI agents proliferate, thousands may access data simultaneously with temporary permissions, sharply raising concurrency requirements alongside throughput demands. NetApp is also building an API-driven control plane so AI teams can provision storage without deep storage expertise, with agents potentially handling that work in the future.