AI-Ready Data Foundation
Everpure Debuts Data Intelligence and Stream Platforms for AI-Ready Enterprise Data Governance
Key Points
Everpure’s GM Prakash Darji said the company is moving “away from role-based access controls in apps to attribute-based access controls,” arguing that data needs middleware akin to where API management is headed—an “API gateway” equivalent for the data era.
SiliconANGLE reported Everpure’s GPU-accelerated ingestion architecture goes beyond preparation and “extends all the way to the inference stage,” and Giancarlo said it enables compute and storage to “scale independently.”
CEO Charles Giancarlo framed data primacy as a reversal of the last three decades, saying: “Data existed to serve the application design… Over the last 30 years, applications have been the center of gravity… App sprawl and AI are reversing this hierarchy.” He also highlighted real-world semantic conflicts, noting that “customer” can mean different things across CRM vs. billing systems.
ITPro said Everpure’s Data Intelligence platform includes “universal” discovery across both structured and unstructured data “regardless of storage format,” plus “automated governance features aimed at bolstering compliance capabilities” and data mapping to drive “AI-ready context” for agents.
Blocks&Files reported Everpure’s approach uses AI not only for classification/discovery, but also for operational control: the company said AI helps “manage, monitor and secure the fleet and the data it holds.”
Everpure, the company formerly known as Pure Storage, launched two major products this week at its Pure Accelerate conference in Las Vegas: Everpure Data Stream and the Everpure Data Intelligence platform. The company is betting that the biggest obstacle to enterprise AI is no longer building models — it's getting clean, well-organized data to those models fast enough. CRN reported the launches as part of a broader software push aimed at helping companies discover, govern, and scale AI workloads.
CEO Charles Giancarlo framed the shift as a historic reversal. "Data existed to serve the application design," he said. "Over the last 30 years, applications have been the center of gravity. App sprawl and AI are reversing this hierarchy." The strategy, which Everpure calls "data primacy," puts governed data at the center — not the apps built around it.
GPU-Accelerated Ingestion Cuts Prep Time from Months to Minutes
Everpure Data Stream uses NVIDIA GPUs to speed up the process of preparing raw data for AI use. The company says ingestion timelines that once took months now take minutes. The product is built on NVIDIA's AI Data Platform reference design and automates the messy, manual work of cleaning and formatting data before it reaches a model. VMBlog reported that the platform enforces access controls directly at the data-stream level, not inside individual apps.
SiliconANGLE noted that the architecture goes further than most storage-focused tools. It "extends all the way to the inference stage," meaning it supports AI workloads from raw ingestion through to the moment a model generates an answer. Giancarlo said the design lets compute and storage "scale independently" — a key requirement when AI workloads spike unpredictably.
Data Intelligence Platform Crawls Mainframes, SaaS Apps, and Everything Between
The second launch, Everpure Data Intelligence, tackles a different problem: enterprises often don't know what data they have or where it lives. The platform crawls and indexes data across four major environments — SaaS apps, databases, object and file stores, and legacy mainframes. ITPro described it as "universal" discovery that works "regardless of storage format," with automated governance features built in to help companies meet compliance requirements.
Once data is found, the platform classifies it and applies policy controls. The goal is to produce what Everpure calls an "AI-ready" view — a clean, governed snapshot that AI agents and models can actually use. The system relies on a semantic graph and self-describing metadata, which add context to raw data automatically without manual tagging.
Security Moves from User Roles to Data Attributes — a Major Shift for AI Agents
GM Prakash Darji explained one of the platform's most significant technical changes. Everpure is moving "away from role-based access controls in apps to attribute-based access controls." In plain terms: instead of granting a user access based on their job title, the system grants access based on what the data itself contains. Darji compared the new approach to an "API gateway" — a control layer that sits between data and the apps or AI agents that consume it.
This matters especially for AI agents, which can carry high-level permissions but should not read sensitive personal data. Giancarlo also highlighted a concrete example of why consistent definitions matter: the word "customer" means something different in a CRM system versus a billing system. The semantic graph is designed to resolve those conflicts so every AI query returns a consistent answer.
AI Also Runs the Storage Fleet Itself, Blocks and Files Reports
Beyond organizing data for AI models, Everpure is using AI to run its own infrastructure. Blocks and Files reported that the company said AI helps "manage, monitor and secure the fleet and the data it holds." That means the storage environment can flag problems, enforce policies, and respond to threats without waiting for a human administrator.
Everpure cited IDC research showing that data quality is now ranked as the top factor in determining AI's return on investment among enterprise technology leaders. The company's broader argument is that the market has reached a turning point: buying better AI models delivers diminishing returns, but fixing the data pipeline underneath them delivers real, measurable value.
Sentiment
Mostly PositivePublishers
15
Articles
16
Reach
31
