Everpure is discovering and classifying data for AI across an organization’s data estate, enhancing FlashBlade data reduction and KV cache performance, and making open weight AI models available, saying it wants to make production AI easier to adopt.
The company is moving into data management in a serious way, scanning and classifying multiple data sources, not just Everpure-stored data, as well as incrementally upgrading FlashBlade capabilities to make it a better AI hardware and software storage engine. The announcements being made today expand on one made in June.
Prakash Darji, General Manager, Data & Digital Experience at Everpure, said: "Enterprise AI is hitting a wall not because the models are lacking, but because data is not ready for real-time, autonomous agents. We are eliminating that friction. By making enterprise data continuously governed, automated, and instantly accessible, we’re giving organizations the foundation to move AI out of the lab and into production with the necessary confidence."
Everpure FlashBlade now pre-stages AI context directly into GPU memory to deliver up to 20x faster Time to First Token (TTFT) with a Pure KVA (Key-Value Accelerator) feature. It supports enterprise multi-tenancy with zero dataset relocation, eliminating GPU idle time, increasing token throughput, and decreasing response lag for real-time apps.
The FlashBlade systems have always-On DeepReduce Data Compression. This scans storage blocks continuously across FlashBlade systems to find sub-block data similarities that traditional deduplication misses, even on pre-compressed content. Everpure says you should achieve a median 2:1 global dat reduction, for FlashBlade//E and FlashBlade//S200R2, on top of existing compression. Usable storage capacity expands automatically without impacting write performance or requiring manual scheduling, significantly reducing, Everpure says, hardware footprint and cross-cloud expenses.
Everpure also now delivers a reference architecture using open weight models to allow more control over data and predictable AI costs, cutting overall API token usage from external providers.
Data Intelligence
Everpure Data Intelligence (formerly acquired 1touch) discovers, classifies, and contextualizes enterprise information stored in Everpure and other on-premises arrays, public clouds, SaaS applications, and third-party storage, including mainframes, using CIFS/SMB, NFS, OneDrive/SharePoint and Google Workspace. New capabilities give autonomous agents and administrators direct, secure access to live enterprise context without custom API work.
Data sources include CIFS/SMB, NFS, OneDrive and SharePoint, and Google Workspace. The software It reports freshness, file age, and concentration signals at the share and folder level, and infrastructure owners get tiering and reclaim signals from the same pass that shows security teams where exposure sits. The tiering signals are passive, there being no optimized, policy-drive data placement. It might well be coming.
The classification identifies private and sensitive information like PII and PHI, tracks lineage, and uses AI and ML to go beyond compliance tags into broad classification, giving customers a complete, governed map of their data landscape.
The data intelligence feature maps raw data to its real-world business meaning, building a multi-dimensional semantic knowledge graph, “so an AI agent doesn’t just locate a customer record; it understands that record’s relationship to purchase orders, geography, and sales channels. Feeding agents only relevant, contextualized data sharpens accuracy while shrinking context windows and token costs.”
Once data has been classified it can be prepared for AI, by being vectorized using Data Stream technlogy
Unified Data Intelligence (UDI) features:
- Native MCP Integration: Implements the open Model Context Protocol (MCP) so AI agents and security tools can query live data catalogs using natural language. Agents can find relevant data and understand its sensitivity class, which is meaningful to AI, agent workflows, and analytics.
- Privacy-first approach: Shows who can access each file share and how stale it is, without reading file content, so admins can fix exposure and reclaim capacity before opening shares to AI agents.
The UDI software has two major components; a Unified Data Plane and an Intelligent Control Plane. The KVA and DeepReduce compression form part of the data plane. The control plane includes cross-fleet configuration setting capability, a Configuration Advisor (Portworx Observability) to find configurations that could be optimized or corrected, and autonomous workflows.
Regarding cross-fleet configuration setting, an Everpure blog says “Most estates carry hundreds of inconsistent per-volume configurations, set years apart by different people. Pure1 Configuration Advisor (beta) reduces them to a small governed set of policies, which removes the steady manual load of maintaining protection, performance, and QoS by hand. It recommends and changes nothing on its own, so you see the impact of a change before you approve it.”
Customer admins can codify the result through Everpure Fusion or their own AI, or export it as an executive summary, JSON, or Markdown.
The autonomous workflows feature, Pure1 workflows for Data Intelligence, is a pre-built recipe in Pure1 Workflow Orchestrator that runs on a schedule. It discovers data sources, extracts sensitivity classifications through MCP, iterates hosts to find their attached volumes, assigns data classes to those volumes, and hands the policy to Everpure Fusion to execute.
Unified Data Intelligence has turn-key deployment through the existing Pure1 console, with no need for professional services engagements or setting up separate management servers.
Separately Everpure commissioned research found that 88 percent of large enterprise C-suite leaders believe a data sovereignty failure could cost them their jobs. It says data sovereignty has evolved into a defining business risk, with challenges ranging from foreign-mandated data exfiltration to potential Denial of Service (DoS) incidents.
It says Everpure data intelligence facilitates sovereignty by providing total visibility of an organization’s data estate, enabling them to make better decisions about how data should be governed and protected to support data sovereignty regulations.
The new capabilities announced today will be available this October. Access the Global DataSovereignty Report here and read a blog about it here. A second blog examines the data intelligence area.
Comment
Everpure is moving into the general data management area in a concerted way and this will be noted by Datadobi, DataDynamics, Diskover, Komprise, and others. They will find it harder to penetrate Everpure customer accounts from now on.
CEO Charles Giancarlo said we could expect more technology tuck-in type acquisitions like 1Touch, which made us wonder about Diskover as a potential target. for example, for its ROT (redundant, obsolete and trivial) data identification technology.
Bootnote
Everpure’s Global Data Sovereignty Report 2026 surveyed 2,100 enterprise leaders across eight countries.