Pure Storage has announced a series of updates designed to enable organizations to roll out AI workloads and traditional enterprise applications better and faster. The company’s aim is to make data more mobile and give organizations more control over their data, regardless of where it is located.
Earlier this year, we wrote about the Enterprise Data Cloud. The company announced this new architecture a few months ago during the annual Pure Accelerate event. You can read more about it via this link. Today, the company is releasing several updates to its portfolio, almost all of which can be seen as updates to the Enterprise Data Cloud story.
Today’s biggest update is partly cosmetic. Pure is introducing the Pure Storage Cloud. This is partly a formalization of what it already offered, such as Cloud Block Store (on Azure and AWS) and around migrating VMware workloads. By positioning this offering more clearly (and adding a few extras), it should make large migrations to the cloud easier, without organizations having to modify their applications.
AI acceleration through hardware and software
Like many suppliers, Pure Storage is also focusing explicitly on AI workloads. The Key Value Accelerator will be integrated with NVIDIA Dynamo for AI Inference. According to Pure, this combination should deliver up to twenty times faster speeds in the field of AI inferencing. It should also reduce energy costs, which is not insignificant these days.
However, today’s announcements go beyond mere performance improvements. With Purity Deep Reduce, Pure is introducing a new data reduction engine that recognizes patterns and can identify similar data. This should increase efficiency without compromising performance.
Pure continues to focus heavily on data control. The idea is that as long as organizations do not have complete control, AI initiatives, among other things, will not do what they are supposed to do. It does not matter where this data is located.
Expansion of cloud portfolio
In addition to Azure, Pure Storage is also expanding its offering in other environments. FlashArray will get new models in the XL and R5 series, which will be available later this year. These systems should be able to support both traditional enterprise applications and modern AI workloads.
When it comes to new models, the FlashArray//XL 190 stands out in particular. This new model will be generally available in Q4 of FY26. In addition, the FlashArray//X R5 and FlashArray//C R5 are now generally available.
AI Copilot is being expanded
Pure Storage is continuing to build on its AI Copilot functionality, which is designed to make the management environment more accessible. The expansion to Portworx means that users can query their Kubernetes clusters in the same way as FlashArray systems.
An interesting development is the integration with Model Context Protocol (MCP) servers. This functionality will be available at the end of 2025 and should enable IT teams to solve problems without in-depth technical knowledge. The system can automatically search for error patterns and suggest solutions using everyday language.
The introduction of Portworx integration with Pure Fusion shows that Pure also takes modern container environments seriously. This allows organizations to manage Kubernetes and virtual machines through a single platform, which should reduce complexity.
More than just marketing
Today’s announcements by Pure Storage are a logical next step for the company. For years, it has been positioning itself increasingly as a data platform, rather than ‘just’ a hardware provider. In our opinion, there is no question that organizations need an increasingly comprehensive data platform such as the one provided by Pure. Whether all companies are ready for it is another story. Many companies are still struggling with basic issues surrounding data quality and governance. An integrated platform can help, but only if the underlying processes are in order.
The collaboration with parties such as Microsoft and NVIDIA shows that Pure Storage looks beyond its own technology. That is a good thing in any case. In a world where interoperability is becoming increasingly important, it is a wise choice.