Qumulo Brings High Performance to Azure Blob Data for AI, Without Moving It
New Blob Accelerator speeds AI pipelines on existing Azure Blob data, as Cloud Native Qumulo adds lower-cost Azure
Press Release Disclaimer: This is a press release distributed through the XPR Media network. It has not been independently verified by our newsroom.
![]()
At Microsoft Ignite today, Qumulo launched the Qumulo Blob Accelerator, which connects AI pipelines and file-based applications directly to data already stored in Azure Blob, even across Azure regions, without migrating or copying it. Qumulo also announced that Cloud Native Qumulo (CNQ) now runs on Azure Cobalt, Microsoft’s ARM-based processor, giving customers compute for their data layer that is easier to obtain and less expensive to run as AI absorbs cloud capacity.
Qumulo Blob Accelerators take object latency, between 10-150 ms, and reduce it to 1-5 ms while simultaneously ensuring that there are no additional transaction costs after the first read of a given object, eliminating API charges.
For most enterprises, the data that will power AI already exists, and a growing share of it sits in cloud object storage. Putting it to work has meant tolerating high latency and repeating API charges or managing high cost scratch space storage with migration for each data set. Meanwhile, AI is absorbing cloud capacity so quickly that even the compute to run the data layer is hard to secure. Qumulo’s approach is to leave data where it lives, bring performance to the data, and run on compute customers can actually get, regardless of location.
“Enterprises have spent years moving data into Azure because it is durable and economical. Now they want to use that data for AI, and two things keep getting in the way: getting the data to the workload without copying it, and getting the compute to run it,” said Brandon Whitelaw, senior vice president and head of product at Qumulo. “The Blob Accelerator puts the data customers already have in Azure to work in place, faster and at a lower cost to read. CNQ on Cobalt gives them compute they can count on when their workloads need it.”
The Blob Accelerator attaches directly to an Azure Blob container a customer already owns, which remains the single, authoritative copy of the data. Applications that expect a file system can reach the data over NFS and SMB, while applications using the Blob APIs continue to work against the same container unchanged. Qumulo NeuralCache serves more than 90% of repeat reads from NVMe and DRAM, returning cached data in under a millisecond and reducing the read-request API charges that build up when AI pipelines and analytics scans read the same data again and again.
- Cloud Data Fabric: The same container can also be read from any Qumulo system in another Azure region, another cloud, an on-premises data center or at the edge, with strict consistency and no copies, so organizations can put their data to work wherever GPU or CPU capacity becomes available.
- Flexible deployment: Existing Azure Native Qumulo customers can add the Blob Accelerator to the cluster they already run, connecting existing Azure Storage Accounts, and customers new to Qumulo can deploy it on their own Azure instances.
- Elastic scaling: The Blob Accelerator adds cache and performance when a data transformation, inference, analytics, or general file workload needs it and scales down when the job is done.
- Early use cases: AI training and inference, lakehouse and observability analytics, media delivery, genomics, and subsurface and seismic imaging simulations.
Cloud Native Qumulo on Azure Cobalt
CNQ on Azure Cobalt solves the other half of the problem: getting compute. As AI workloads eat up capacity, customers face shortages, delayed deployments and pressure to over-commit to reservations just to get x86 instances. Every CNQ tier, plus Qumulo AI Accelerators and the Blob Accelerator, now runs on Cobalt-based instances. These are easier to get and cost 9% to 48% less than comparable x86 instances. The software, features and single namespace are unchanged, so nothing needs to be refactored. And because CNQ scales by adding compute on demand, better instance availability means customers can scale further when they need it most.
Availability
The Qumulo Blob Accelerator is available in controlled preview beginning today, with read/write support and general orderability expected in the first quarter of 2027. Organizations interested in the preview can contact their Qumulo account team. CNQ on Azure Cobalt is available today through the Azure Marketplace and directly from Qumulo. Visit Qumulo at Microsoft Ignite at booth 2126R or learn more at qumulo.com.
About Qumulo
Qumulo is the only seven-time Leader in the Gartner Magic Quadrant for Distributed File and Object Storage and the foremost provider of cloud data platforms. With exabytes under management and more than 1,000 production customers, Qumulo is trusted by Fortune 500 companies and global enterprises to manage, store, curate, and protect their data, unlocking new possibilities and driving innovation across diverse industries. For more information, visit www.qumulo.com.
View source version on businesswire.com: https://www.businesswire.com/news/home/20261117643397/en/
Media gallery

