Uploaded September 2026 | Updated November 2026, 5 days ago
AI infrastructure is pushing storage along two dimensions at once: extreme performance near the GPU and massive capacity for rapidly expanding datasets.
In this Backblaze Investor Day conversation, Backblaze CTO Dan Spraggins and WEKA Chief Strategy Officer Nilesh Patel discuss how high-performance and capacity storage work together across modern AI infrastructure.
WEKA provides software-defined storage designed to deliver memory-like latency for demanding GPU workloads, while Backblaze provides the capacity layer needed to scale large datasets cost efficiently. Together, the companies see an architecture that delivers speed where it matters and scale wherever it is needed.
The discussion covers GPU utilization, training and inference, checkpoint data, distributed storage, global namespaces, hardware lead times, and the growing infrastructure demands of AI-native companies, enterprises, neoclouds, and sovereign clouds.
AI infrastructure is pushing storage along two dimensions at once: extreme performance near the GPU and massive capacity for rapidly expanding datasets.
In this Backblaze Investor Day conversation, Backblaze CTO Dan Spraggins and WEKA Chief Strategy Officer Nilesh Patel discuss how high-performance and capacity storage work together across modern AI infrastructure.
WEKA provides software-defined storage designed to deliver memory-like latency for demanding GPU workloads, while Backblaze provides the capacity layer needed to scale large datasets cost efficiently. Together, the companies see an architecture that delivers speed where it matters and scale wherever it is needed.
The discussion covers GPU utilization, training and inference, checkpoint data, distributed storage, global namespaces, hardware lead times, and the growing infrastructure demands of AI-native companies, enterprises, neoclouds, and sovereign clouds.










