NetApp has launched NetApp Novus, the fastest storage architecture designed for large-scale AI factories operating with GPUs. The product supports a zettabyte-scale file system and offers up to 100TB/s of performance.
Key points
•NetApp Novus: fastest storage architecture for AI factories
•Supports up to 100TB/s performance, zettabyte-scale file systems
•Maximizes GPU utilization and investment efficiency
NetApp, an intelligent data infrastructure company, introduced NetApp Novus, tailored for AI infrastructure providers managing extensive GPU environments. This next-generation architecture enables the efficient scaling of metadata and bandwidth independently within a single namespace, supporting zettabyte-scale file systems.
Syam Nair, Chief Product Officer at NetApp, stated that 'NetApp Novus is crucial for ensuring AI factories can generate desired outcomes and achieve returns on investment by providing sufficient data to GPUs.' The architecture supports up to 100TB/s of performance, addressing the limitations of traditional storage architectures in handling large-scale AI environments.
Designed to eliminate bottlenecks and maximize GPU utilization, NetApp Novus separates metadata from the data path. This allows for independent scaling of performance, capacity, and concurrency while maintaining enterprise-level data services through ONTAP. The solution is intended to meet the demands of multi-tenant cloud-scale AI factories.
