Everpure aims to bridge AI data gap with Universal Data Intelligence

Making enterprise data AI ready

The Everpure Data Intelligence announcement, made at the supplier’s Accelerate 2026 conference in Las Vegas this week, centres on the integration of the OneTouch acquisition that allows organisations to discover, classify and contextualise structured and unstructured data across hybrid environments.

According to Patrick Smith, field chief technology officer (CTO) for EMEA at Everpure, the pivot addresses a critical bottleneck in enterprise AI adoption where fragmented infrastructure and siloed applications can lead to poor results and spiralling costs.

“For decades, environments have been application-centric,” said Smith. “The data for applications is locked behind an application, and that means with thousands of applications in an enterprise, you have thousands of silos. In a new world, we need to pivot to being data-centric, where applications access a unified dataset.”

The UDI layer – of which Everpure Data Intelligence is a part – aims to make enterprise data “AI ready” by creating a semantic knowledge graph of relationships across datasets that allows data science teams to identify relevant information for specific use cases without manual extraction.

Everpure also expanded its Unified Data Plane with the introduction of Purity Turbo, a software enhancement for the FlashArray//XL190 that delivers mission-critical performance boosts. This system is now positioned as the fastest in the FlashArray line, targeted at the extreme performance requirements of high-frequency AI training and real-time inference.

Automating the flow of data

In the cloud, the supplier announced Everpure Cloud Azure Native for VMs, which decouples storage from compute to allow organisations to scale Microsoft Azure workloads more cost-effectively. The service, scheduled for general availability in July, allows Azure native VMs to utilise Everpure’s volume-as-a-service capabilities directly through the Azure portal.

To assist with the ingestion of data into AI projects, the supplier confirmed that Everpure Data Stream is now available. This automation layer manages AI data pipelines and integrates vector databases for RAG scenarios. By automating the connectivity between the enterprise dataset and the AI pipeline, Everpure aims to reduce the time spent on data preparation.

“AI doesn’t come first; the business challenge comes first,” noted Smith. “If you don’t have the data that’s relevant to that, then you have a problem. If you don’t understand where that relevant data sits, you have a problem.”

The management of these sprawling environments is addressed via the Intelligent Control Plane, which now includes MCP servers for Fusion and Pure1. These servers enable “agentic” workflows that allow AI agents to query infrastructure topology and performance metrics. This allows for automated actions, such as AI-driven performance triage and fleet-wide data rebalancing, which is now available in preview.

Subscription-based users under the Evergreen//One model also gain access to Overdrive, a new performance-on-demand feature. While Evergreen//One previously offered capacity bursting, Overdrive allows customers to burst performance 25% above their service-level agreement (SLA) to meet temporary business demands.

Rising to the data sprawl challenge

With these latest announcements, Everpure is consolidating its transition towards a market traditionally occupied by specialised metadata players and data governance platforms. The challenge for any storage supplier moving “up the stack” is to prove it can handle data logic and semantic contextualisation without losing focus on its core high-performance hardware. 

The supplier also introduced a “Success Blueprint” for the Enterprise Data Cloud that comprises assessments and workshops to help organisations transition from legacy application-centric architectures to the new model. Smith argued that this shift is necessary as organisations face the “garbage in, garbage out” problem with AI.

“Data sprawl has been a challenge for decades,” said Smith. “The less that is understood about that growing dataset, the worse the returns you’re going to get with your AI initiatives. Infrastructure platforms are diverse and fragmented, leading to management challenges that increase exponentially at scale.”

The new UDI layer is designed to quantify and manage this growth, providing a platform where data is known and governed before it is consumed by AI agents or large language models.