Nvidia GTC 2026: HPE Alletra Storage MP X10000 Becomes First Nvidia-Certified Storage Object-Based Platform for Enterprise AI
Why object storage is emerging as the critical data foundation for production AI
This is a Press Release edited by StorageNewsletter.com on March 18, 2026 at 2:02 pm- HPE Alletra Storage MP X10000 becomes the first object storage platform to achieve Nvidia-Certified Storage validation for object-based systems at the Foundation level
- Certification validates that X10000 can efficiently feed data to accelerated computing environments powering enterprise AI workloads
- The milestone reinforces the role of object storage as the foundational data layer for modern AI infrastructure elit.
As artificial intelligence moves into production, organizations are discovering that the next frontier of AI infrastructure performance is the data platform itself.
GPUs and accelerated computing may power AI models, but the ability to move massive volumes of data quickly and reliably is what ultimately determines how efficiently those systems operate.
Training, fine-tuning, and inference workloads require rapid access to enormous volumes of unstructured data. As a result, object storage is increasingly emerging as the foundational data layer for modern AI architectures.
This milestone reflects a broader shift in enterprise AI infrastructure, where object storage is rapidly becoming the primary data foundation for large-scale AI pipelines and GPU-accelerated computing environments. The certification is achieved under the Foundation level, which validates storage platforms against AI workloads scaling to environments of up to 128 GPUs.
HPE today announced that HPE Alletra Storage MP X10000 has achieved Nvidia-Certified Storage validation for object-based systems at the Foundation level, becoming the first object storage system to reach this milestone.
“AI infrastructure performance ultimately depends on how efficiently data can reach accelerated computing environments,” said Jim O’Dorisio, SVP and GM, Storage, HPE. “Becoming the first object storage platform to achieve Nvidia-Certified Storage validation underscores the critical role object storage now plays in AI architectures and validates the design of HPE Alletra Storage MP X10000 as a foundation for enterprise AI data platforms.”
Validation for AI infrastructure
The Nvidia-Certified Storage program validates that storage platforms can meet the demanding performance, reliability, and scalability requirements needed to support AI workloads and integrate with NVIDIA accelerated computing environments. The certification is achieved under the Foundation level, which validates storage platforms against AI workloads scaling to environments of up to 128 GPUs.
With Nvidia-Certified Storage Foundation certification, HPE Alletra Storage MP X10000 has been validated by Nvidia to deliver the performance required for these AI environments, with extensive benchmarking and functional testing confirming enterprise-grade availability and reliability.This validation confirms that the platform can efficiently feed data to accelerated computing resources, helping organizations accelerate model training, support lower-latency inference, and improve overall GPU utilization.
Object storage designed for AI data pipelines
HPE Alletra Storage MP X10000 is designed specifically to support large-scale unstructured data environments that power modern AI and analytics workloads.
The platform combines scale-out object storage with integrated data intelligence capabilities that help organizations:
- Store and manage massive volumes of AI data
- Accelerate data pipelines feeding GPU-accelerated workloads
- Scale performance and capacity independently as environments grow
- Prepare and enrich data for AI applications directly within the storage platform
This architecture enables organizations to deploy a data platform capable of keeping pace with the scale and performance requirements of modern AI infrastructure.
Advancing the AI data ecosystem
The certification reflects the continued collaboration between HPE and Nvidia to accelerate every stage of the AI data lifecycle. By combining HPE Alletra Storage MP X10000 with Nvidia accelerated computing platforms, organizations can deploy AI infrastructure with a validated storage foundation designed to support the scale and performance demands of production AI workloads.
These capabilities build on the broader Nvidia AI Computing by HPE portfolio, which integrates HPE infrastructure with Nvidia accelerated computing, networking, and software to help enterprises deploy, operationalize, and scale AI environments more quickly and reliably.
Comments
As we celebrate the 20th anniversary of S3, many of us recall Jeff Barr's post from March 14, 2006. Since that post, object storage - and its obvious companion, S3 - has profoundly transformed not only the storage industry but far beyond.
Object storage emerged as a radical new model, embraced by a few purist players, while others simply layered an S3 interface on top of existing storage solutions. Today, we commonly refer to S3 storage as object storage, though the reverse isn't necessarily true - even if S3 has become ubiquitous.
Over time, this de facto standard has achieved universal adoption, steadily expanding its footprint through new data models, interface extensions, and backend policies - while also attracting major players from the data analytics ecosystem.
With the rise of AI, this trend is further reinforced - but it also pushes vendors to clarify their positioning, as not all object storage solutions are created equal.
For HPE, the object storage strategy has been somewhat unclear in recent years, as the company prioritized promoting its hardware regardless of the software layered on top. Many readers will recall that HP(E) supported a wide range of solutions, including SwiftStack, Cleversafe, Ceph, OpenStack Swift, Cloudian, and Scality - among others. The rationale was understandable: "Don't lose any deal over software."
More recently, however, the strategy has begun to evolve. HPE selected VAST Data for file storage, which led many to anticipate a similar approach for object storage. Instead, the product team opted to build its own solution on the new Alletra storage platform. This shift has been further reinforced by the decision to stop reselling certain partner-based file and object storage offerings.
Since then, HPE has made a strong push in this area, and its identity around object storage has evolved - positively impacting both partners and end users.
That said, object storage for secondary use cases differs significantly from primary storage, which itself comes in multiple forms. Supporting an image-based application via S3 is very different from handling demanding AI workloads. As is well understood, feeding data efficiently into processing pipelines is critical, making parallelism and scalability essential requirements.
Certification processes reflect this reality. At the Foundation level, systems are validated for their ability to support up to 128 GPUs and to perform reliably at that scale. Two additional tiers extend these benchmarks further: the Enterprise level, supporting up to 1,000 GPUs, and the NCP (Nvidia Cloud Partner) level, scaling up to 10,000 GPUs.
As we've said many times, users increasingly prefer to tackle modern challenges with modern solutions. This is precisely why companies like DDN have developed Infinia, and why HPE introduced the HPE Alletra Storage MP X10000.
Click to enlarge
Regarding Nvidia-certified solutions listed on this page, only HPE appears at the moment. However, Cloudian has also announced its certification at the Foundation level through an official press release, and DDN is expected to be added soon, along with several others. Being first is not what matters most - being included on the list is.
Subscribe to our free daily newsletter
