In addition to its own data, Everpure has significantly expanded its ability to classify data for AI and make that data available to models, agents, and apps across its fleet of systems, using AI to better manage, monitor, and secure fleets and the data they hold.
He said the IT industry has shifted its focus from a traditional app-centric view to a data-centric one. While each application has tended to have its own data stack, AI models and agents require cross-app silos, universal data intelligence.
universal data intelligence
CEO Charles Giancarlo said that for 50 years, “data existed to help design applications and, indeed, computer operating systems. For the past 30 years, it has been about applications. App sprawl and AI are reversing this hierarchy.” A data-first message is emerging.
“A typical enterprise today has dozens, hundreds, or even thousands of applications, each with its own data set, and typically many copies of it. The most common enterprise activities, such as generating quotes for customers, require access to multiple applications and data sets that often do not share consistent information.… A “customer” in a CRM has a different meaning than a “customer” in a billing system. An “asset” in ERP is not the same object as an “asset” in a supply chain platform. ”
To fix this, we need to bring to the surface and make visible the data context—its semantic meaning—that was previously hidden within the silos of individual apps or app suites. This requires a semantic knowledge graph or graph database. Giancarlo says, “Data or a particular dataset must act as a system of record that multiple applications can use for enterprise workflows. … Data must be self-describing, and relationships between different systems of record must be recorded in self-describing metadata.”
Essentially, “context is created, saved, and maintained” along with the data itselfas metadata. …The core of the value is data that is controlled, managed, and understood by the enterprise and can be used for multiple purposes by multiple applications and agents. ”
This is a universal data intelligence layer, and its capabilities are based on 1touch, a contextual data intelligence startup that Everpure acquired in February. This allows Everpure to “discover, classify, and contextualize data across enterprise assets, including on-premises, cloud, SaaS, and legacy environments, and build catalogs and semantic knowledge graphs around these datasets of information.”
This data intelligence layer provides:
-
Universal Discovery: Provides visibility. Contains structured and unstructured data regardless of storage format Major databases such as SQL Server and Oracle. Show exactly where your critical application data resides.
-
Automated governance: Automatically scan your entire system to identify sensitive information (such as PII and PHI), track lineage, and provide a complete map of your data landscape to ensure secure compliance.
-
AI-enabled context: Map raw data directly to real-world business definitions to create a semantic knowledge graph. This enables modern AI agents to instantly understand, query, and securely act on information across your enterprise.
Prakash Darji, GM of Digital Experience BU at Everpure, provided the following use case:Consider, for example, a major sports medicine network. Data breaches here don’t just result in non-compliance fines. It directly impacts an athlete’s trading value and drives insider bidding. We use Everpure Data Intelligence to automatically isolate and protect exactly this type of risky information. Everpure Data Intelligence scans disparate sources and instantly flags 42 separate files containing highly sensitive athlete data that have escaped traditional controls. From there, build a true data blueprint so your team understands exactly where that information resides, how it’s used, and how it moves through your organization. ”
Unified data plane
The company put forward the concept of a unified data plane and intelligent control plane under the banner of enterprise data cloud (EDC) at its Accelerate 2025 event and is currently honing them with additional components.
Everpure announces unified data plane updates Provide a shared foundation across your enterprise, eliminate performance silos, maximize resource efficiency, and bring cloud-like scaling directly to your physical data architecture.
We have FlashArray//XL 190 systems powered by PurityTurbo, Everpure Data Stream, Cloud Azure native virtual machines, and Evergreen//One Overdrive.
The FlashArray//XL 190 is Everpure’s flagship “big iron” array, with DRAM-class latency of less than 100 microseconds and a claimed 930 percent increase in IOPS per rack unit compared to traditional systems, although details are unclear. It was first revealed in September of last year, when Everpure announced 25% more capacity, 50% less latency, and 100% more performance than the existing //XL170.
Exclusively uses the Turbo version of the Purity operating system. Typically, Purity has both controllers in Everpure’s dual-controller array accepting IO requests from the host, but only one controller writes to and reads from the system’s SSD storage. Purity Turbo changes this We partially enhance the basic design to be active/active, allowing a second controller to read from the SSD and actively process read I/O. Purity Turbo also expands the cache footprint across both controllers.
Although the XL190 is a flagship system, it is not a top-performance FlashArray, or ST, as Everpure’s positioning diagram shows.
Everpure Data Stream optimizes (prepare) your data for AI. “Vectorize your data into pipelines so your data is organized the way you want your AI agents/apps to access it.” Through automated pipelines, you can make your data AI-enabled “in days instead of weeks.”
Evergreen//One Overdrive provides a temporary cloud-like performance boost for on-premises storage, seamlessly absorbing traffic spikes of up to 25% above your baseline without the need for permanent subscription upgrades.
Darji said.Evergreen One Overdrive has traditionally supported on-demand capacity across all service levels. With 2% headroom, overdrive provides performance on demand, eliminating the need for overprovisioning and leading to greater asset efficiency. ”
Everpure Cloud Azure-native virtual machines run in the Azure public cloud and separate storage from compute, reducing costs for storage-intensive workloads. Customers can lift and shift code to Azure without refactoring and manage VMs natively in the Azure portal like a first-party service.
intelligent control plane
This embeds AI directly into daily operations, turning manual, reactive storage management into a self-optimizing, secure environment. Powered by natural language orchestration and predictive behavioral analytics, it delivers:
-
Workload rebalancing and mobility: Automatically move active workloads across your fleet without downtime, optimizing capacity distribution and ensuring consistent application performance.
-
CoPilot Workflow Execution: Enables storage administrators to use natural language to plan, validate, and trigger secure end-to-end operations across global infrastructure assets.
-
Enhanced Cyber Anomaly Detection: Monitor telemetry across your environment to identify tailored and suspicious login patterns and behavioral drifts that individual storage arrays may miss on their own.
-
Fusion Compliance and Agentic Triage: Automatically detect hardware or software configurations
-
Drift leverages agent AI to suggest root causes and provide instant technical remediation while strengthening global corporate governance.
“Because our platform is open and operational, we have monitoring capabilities, we’ve added open telemetry and the ability for people to take action using the Fusion MCP server, Fusion topology, and we’ve provided enhanced intelligence with the Pure1 MCP server. From an intelligence and compliance perspective, last year we introduced fleets, preset workflows, and workflow orchestration, and now we’re actually using LUN Tier 1 application intelligence to crack open and understand what’s inside your data and cyber.”Anomaly detection is behavioral, with compliance monitoring and remediation to check for configuration drift. ”
On the agent side specifically, he says, “Customers use automation in four ways. The first is workflow automation. For example, let’s say you have a critical vulnerability. Before you create a ticket, Copilot searches for a solution. Another approach is something like, ‘Hey, show me which volumes don’t have safe mode snapshots.’ Instead of asking questions every day, schedule them.
“And now, customers can schedule questions instead of resubmitting them every day. Customers using OpenClaw or their own automation can capture intelligence from their Pure fleet assets through Pure1 MCP and automate actions through Fusion MCP servers. This allows them to incorporate this into their existing automation workflows. And finally, agent collaboration, where humans and agents can work together to solve problems.”
Everpure offers performance triage workflows. “We deliver this as part of our platform, so when you access workflows directly within Pure1, you find intelligent automated workflows that monitor high-latency scenarios across all arrays, from dozens to hundreds. They consistently monitor latency and resolve issues in a secure and controlled manner.”
“What does a triage agent do? It’s pretty much what a human would do, but faster and smarter. So in this case, the agent goes ahead and looks for the problem. It triages the problem. Once it finds the problem, it has to verify that the problem actually originated externally, suggest a root cause, and hand it off to a human or another agent to remediate the problem.”
If the human hands off, and it’s a backup processing issue, “in Slack, the human can simply say, ‘I’d like to apply QOS or rebalance the workload.’ If you select Rebalance, the workload will be rebalanced. A monitoring agent is assigned to run the next two or three backup jobs to ensure that the problem does not reoccur and close the case. This gives you a safe and controlled way to automate agents depending on their latency. Solve problems in your fleet and simplify workload management for your customers. ”
Enterprise data cloud blueprint
Everpure also introduces a step-by-step EDC success blueprint for building and scaling an enterprise data cloud. It begins with a hands-on readiness assessment to identify pressing infrastructure and security risks, then plots a path across 10 operational pillars to move your environment from manual management to a highly automated and efficient architecture.
Customers have access to the EDC Success Blueprint Guide and workshops designed to implement their customized plan.
With this, Everpure states: “EDC provides a unified, secure data fabric that shows exactly where your data lives, how it is connected, and what it means.…Governance policies protect data based on its inherent meaning rather than application silos, AI workflows optimize around shared enterprise context, and infrastructure automatically adapts to real-world usage.”
availability
-
FlashArray//XL190 is available now.
-
Purity Turbo is available now.
-
Everpure data stream now available
-
Everpure Cloud Azure Virtual Machines will be generally available in July.
-
Evergreen/One Overdrive is scheduled to launch in the third quarter of 2026.
-
Intelligent control plane features:
-
Workload Rebalancing and Mobility – Q4 2026
-
CoPilot Workflow Execution – Q2 2026
-
Enhanced Cyber Anomaly Detection – Q2 2026
-
Fusion Compliance and Agent Triage – Q4 2026.
-
