Qumulo launched Cloud AI Accelerator, a data platform that enables enterprises to access distributed datasets across cloud, edge, and on-premises environments in real time.

The offering combines Qumulo's Cloud Data Fabric, Cloud Native Qumulo, and NeuralCache technologies to enable AI workloads to access data across clouds, regions, and hybrid environments without replication or staging delays.

The company said the platform addresses a key challenge in enterprise AI deployments, where expensive graphic processing unit (GPU) resources remain idle while data is moved, copied, and prepared for training and inference workloads. Qumulo cited industry analysis suggesting enterprise GPU utilization averages around 5%, leaving significant compute resources underused.

According to Qumulo, Cloud AI Accelerator creates what it describes as “GPU liquidity” by allowing enterprises to run AI workloads wherever GPU capacity is available rather than where data is stored. The platform combines Cloud Native Qumulo (CNQ), Qumulo Cloud Data Fabric, and Qumulo NeuralCache technologies to provide real-time access to enterprise data across regions, clouds, and hybrid environments.

“Every enterprise we talk to is focused on GPU availability, but availability is only half the problem. The deeper issue is utilization, and the culprit is data gravity,” Qumulo CEO Douglas Gourlay noted in a statement.

Brandon Whitelaw, Qumulo's SVP and head of product, added that many existing approaches rely on tightly coupled GPU-attached storage systems that optimize active compute time but do little to address the data-staging delays that precede AI workloads.

Whitelaw said joint testing with cloud AI and machine learning teams showed less than 0.05% performance degradation when accessing data across distances of up to 2,000 miles compared with locally staged datasets.

The platform is designed to connect on-premises and cloud-native Qumulo environments to AI services, including Microsoft AI Foundry, Amazon Bedrock, and Google Vertex AI without requiring data copies. Qumulo said customers can use the technology to eliminate data staging delays, reduce idle GPU time, and avoid maintaining separate storage silos across multiple environments.

The launch builds on Qumulo's Cloud Data Fabric technology and follows the company's November 2025 preview of Cloud AI Accelerator as part of a broader AI infrastructure strategy focused on connecting enterprise data with cloud-based AI compute resources. The company positioned the technology as a way to accelerate data access for AI training, inference, and analytics workloads using its NeuralCache technology.

Cisco networking, security, and Unified Computing System (UCS) infrastructure form part of the architecture for hybrid deployments. Qumulo said the combination enables enterprises to build AI infrastructure that can adapt to changing GPU availability across clouds and regions.

Cloud AI Accelerator is available immediately across Amazon Web Services (AWS), Microsoft Azure, Google Cloud, and Oracle Cloud Infrastructure (OCI), with support for hybrid deployments using Cisco UCS environments.