BackBlaze Announces Five-Year Deal with CoreWeave

Backblaze (Nasdaq: BLZE) announced a multi-exabyte, $335 million storage agreement with CoreWeave (Nasdaq: CRWV), with order forms covering initial terms of five and seven years, per the Business Wire press release issued June 22, 2026. Under the deal, Backblaze provides cost-efficient HDD-based storage capacity supporting portions of CoreWeave's managed storage infrastructure - specifically the tiered storage layer within CoreWeave AI Object Storage. Customers already using CoreWeave AI Object Storage will gain access to new service tiers without code modifications, per the announcement.

Gleb Budman, co-founder and CEO of Backblaze, said: "Storage is the foundation every AI workflow is built on -- without it, even the world's most powerful compute sits idle. We're pleased to work with CoreWeave on elements of their storage environment."

Nick Hoover, Vice President at CoreWeave, said: "Backblaze has built a reputation for making complex, HDD-based storage infrastructure reliable and easy-to-consume at scale. We're pleased to work with them as we continue expanding our platform and managed service offerings to support AI workloads at scale."

CoreWeave serves 9 of the top 10 AI model providers, per the announcement, and has built a GPU cloud platform focused on AI training and inference. The storage deal addresses the tiered-storage problem in AI workloads: high-performance NVMe storage is expensive, and offloading checkpointing, data preparation, and archival layers to cost-efficient HDD-based storage reduces cost without sacrificing throughput for hot-path inference. Backblaze serves more than 500,000 customers in 175 countries, per the company.

Backblaze issued equity warrants covering up to 4.19 million common shares at $7.60 per share to CoreWeave, with vesting linked to time under the agreement and contracted storage capacity, per the SEC 8-K filing. This aligns the two companies' incentives around growing storage utilization under the contract.

For AI/ML practitioners and infrastructure teams, this deal illustrates the storage-tiering architecture emerging at hyperscale AI clouds: GPU compute at the top tier, high-performance object storage in the middle, and cost-efficient bulk storage (like Backblaze B2) at the base. The $335M committed value and multi-exabyte scale signal that AI data volumes are large enough to justify long-term infrastructure contracts with independent storage providers, rather than relying solely on first-party cloud storage.