Backblaze has signed a five-year, $335 million multi-exabyte storage agreement with CoreWeave to provide HDD-based storage tiers within CoreWeave's AI Object Storage, supporting AI workloads with cost-efficient capacity while preserving high-performance resources.
The five-year, multi-exabyte deal places Backblaze's HDD-based storage directly inside CoreWeave's AI Object Storage as a cost-efficient capacity tier, preserving higher-performance resources for active AI workloads. CoreWeave's customer base — nine of the top ten AI model providers — means this storage architecture will be tested at the largest scale of AI training and inference. Watch whether existing CoreWeave customers migrate meaningful data volumes to the new HDD tiers, and whether Backblaze replicates this model with other neoclouds.
Backblaze has been building its position as a capacity-tier provider for AI infrastructure beyond its traditional media and backup markets. In its Q1 2026 earnings report, Backblaze reported 76% year-over-year growth in AI customer count and 24% B2 Cloud Storage revenue growth to $22.4 million, with CEO Gleb Budman citing wins including an AI training data company and a generative AI video creation company that together contributed roughly $1.5 million in annual contract value. The company raised its full-year 2026 revenue outlook to between $161.5 million and $163.5 million, up from a prior range of $156.5 million to $158.5 million, signaling that the CoreWeave agreement is landing on top of an already accelerating AI storage business.
The CoreWeave deal fits within a broader neocloud storage strategy that Backblaze's management outlined in its Q1 2026 investor presentation. Backblaze estimated a $14 billion AI infrastructure and tools market opportunity by 2030 and positioned itself as the performant hard-drive tier behind flash in neocloud architectures, noting that flash costs roughly 10 times more per terabyte than hard drives. The company described active engagement with multiple leading neocloud providers, suggesting the CoreWeave agreement is the first of several planned deployments of its B2 Neo offering.
The infrastructure implications of Backblaze's neocloud strategy are becoming measurable. Backblaze reported that neocloud traffic is driving sustained transfers between storage systems and compute clusters at observed rates ranging from 100 Gbps to 1 Tbps, according to a May 2026 Data Center Knowledge report on how AI workloads are reshaping data center traffic patterns. These sustained, high-bandwidth flows between HDD-based storage and GPU compute clusters represent a fundamentally different traffic profile from the bursty, request-response patterns typical of web and streaming workloads, and they validate the architectural premise behind the CoreWeave deal: that AI training pipelines need a large, inexpensive capacity tier sitting directly adjacent to flash-backed compute.
Add StreamingMeme as a preferred source on Google to see more of our streaming news at the top of your Search results.
Add as preferred source