Backblaze and WEKA Integrated Storage Platforms
The partnership combines NeuralMesh with B2 cloud storage to optimize AI infrastructure workflows.
Updated on Sept. 22, 2026 in Data Centers
Live Poll
Do you trust integrated storage solutions to effectively reduce the complexity of your company's AI operations?
Backblaze and WEKA have announced a partnership to integrate NeuralMesh with B2 cloud object storage. This announced integration allows teams to balance high-performance GPU demands with lower-cost capacity storage.
Why it matters
The collaboration aims to reduce engineering overhead for AI infrastructure teams by providing a tiered system for managing massive datasets. It addresses the rising economic challenge of maintaining fast data access alongside long-term storage for AI assets.
Backblaze brings two decades of capacity storage history to the partnership, while WEKA provides NeuralMesh for high-speed AI workloads. Certification of B2 Cloud Storage for NeuralMesh is currently in progress.
The players
Backblaze
A provider of cloud object storage and data backup services with two decades of industry experience.
WEKA
A developer of specialized data platforms for AI and high-performance computing workloads.
The details
The integration utilizes a workflow where customers maintain raw data in B2 and shift it to NeuralMesh for active processing. Completed workloads and checkpoints are then moved back to B2, an object-based storage platform designed for scalability, for long-term retention. A feature called Snap-to-Object—a mechanism for synchronizing and tiering data between high-performance filesystems and cloud buckets—has already been tested with this architecture.
Timeline
September 22, 2026: Backblaze and WEKA officially announced their integrated storage partnership.
The Tech Race
This integration follows a broader industry trend toward tiered data management for AI, positioning WEKA against traditional high-performance parallel file systems. It marks a push to standardize how cloud storage handles the checkpointing requirements of large language model training.
Infrastructure teams can expect to reduce manual engineering work by automating the movement of data between performance tiers. The full benefit will be realized once the certification of B2 Cloud Storage for NeuralMesh is finalized.
The takeaway
The partnership highlights a critical shift toward hybrid storage architectures that separate active GPU-bound data from passive capacity assets. Watch for the completion of B2 certification and further benchmarks comparing total cost of ownership for these integrated workflows.
Further reading
Find more analysis on infrastructure scalability in our Data Centers section.
Source note: This article includes information reported by ChannelLife US.
Live Poll
Do you trust integrated storage solutions to effectively reduce the complexity of your company's AI operations?
