From storage to intelligence: WWT's takeaways from Everpure //Accelerate '26

Last month, World Wide Technology sent one of its largest delegations yet to Everpure's annual tech conference, //Accelerate '26. The team covered the full map. From enterprise, global financial services, commercial, and public sector, as well as joined by storage SMEs from our GS&A Storage practice and partner management. And yes, we walked away with hardware: WWT was named Everpure's Global Partner of the Year at the Accelerate Partner Forum.

The trophy is nice. But the announcements are the real story.

This is the first //Accelerate since Pure Storage rebranded as "Everpure" earlier this year, a name change that turns out to be more than marketing. It signals a genuine strategic pivot: from storage company to data company. Storage still matters (roughly 20% of annual R&D budget goes to FlashArray and FlashBlade enhancements), but the headline theme at //Accelerate '26 was data: how to find it, understand it, secure it, and make it work for you in an AI-first world.

Here's a few highlights from the keynotes, breakout sessions, and show floor conversations that stood out to our team. 

The keynote: numbers, a name change, and a new vision

Chairman and CEO Charles Giancarlo opened with the metrics that matter. Everpure's Net Promoter Score remains the highest in the industry at 84. In a space full of vendors promising everything, customers are actually staying and recommending. That's rare.

Charlie addressed market headwinds such as volatility, NAND supply constraints, and competitive flash shipment data. Based on 2025 numbers, Everpure was neck-and-neck with the other two major storage OEMs. With Q1 2026 results in, they're now projecting that by year-end, they'll have shipped as much flash storage as all competitors combined. Bold claim, but the trajectory supports it.

Then came the "why Everpure" explanation, which reframed everything that followed. Historically, data has always been pulled to the application through fragmented copies and siloed storage. This is expensive to manage, hard to secure, and increasingly impossible to govern. In the agentic AI era, that model simply doesn't work anymore.

Everpure's answer is a Data-Primacy model: bring the applications to the data through Shared Contexts and Universal Data Intelligence. Their recent acquisition of 1touch.io is the first major move in that direction, adding an intelligence layer on top of the Enterprise Data Cloud (EDC) stack introduced at last year's //Accelerate.

Everpure Data Intelligence: your data has secrets and it's time to find them

The 1touch.io platform is now called Everpure Data Intelligence, and it does three things: Discover, Classify, and Contextualize your data.

In one of the breakout sessions, the demo started simple: connect your data sources — file shares, databases, cloud storage — and the scanner gets to work, surfacing PII such as social security numbers, credit card numbers, and national IDs. It shows you exactly where sensitive data lives, with links and visualizations.

That alone is useful. The real power kicks in when you provide a structured source of truth. Feed it an employee table (names, addresses, SSNs) and a customer table (account numbers, credit card data), and now the tool classifies and contextualizes your entire environment against those definitions. When that sensitive data appears anywhere else, you're alerted immediately, with full audit trails and compliance reporting that your security team will actually appreciate.

From those findings, you can apply access controls, lifecycle policies, and automated alerts to move sensitive data toward being AI-Ready. Which is where Data Stream comes in, and if you haven't heard of it yet, you will.

Data Stream: the easy button for AI-ready data

The biggest challenge most organizations face in their AI journey isn't finding a model. It's getting their data ready to use with one. AI-ready data has to travel a pipeline: Ingest → Curate → Transform → Vectorize → Index → Serve. Most teams are managing that pipeline by hand, which is slow, inconsistent, and hard to audit.

Data Stream automates the handoff between every stage. Built on the NVIDIA AI Data Platform reference design, it's containerized software running on commodity servers with NVIDIA accelerated computing, using Portworx and FlashBlade under the hood.

Connect Data Stream to your unstructured datasets and use a no-code GUI to move raw data to AI-ready through guided wizards. Choose your persona view, select your LLM models, verify each stage with real-time previews, and get full code output and run-to-run audit trails. The semantic layer is particularly worth noting: define the objects and relationships specific to your business, constrain what the agent can access, and you get consistent, auditable output every time.

For regulated industries and organizations managing AI governance concerns, that consistency isn't a nice-to-have — it's the whole ballgame. The ROI case is direct: fewer manual steps, faster time-to-insight, and a repeatable path to unlocking value from data you already own.

Fusion: autonomous storage is getting closer

Fusion, the fleet-level management layer built into recent Purity OS releases for FlashArray and FlashBlade, received two significant announcements at //Accelerate '26.

The first is Rebalance and Workload Mobility, coming in a Purity release later this year. When an array can no longer meet a workload's performance or capacity requirements, Fusion won't just flag the problem; it will fix it. Fusion presents recommendations, runs pre-checks on connectivity, and offers 1-click remediation to migrate workloads across the fleet. Under the hood, it uses proven ActiveCluster technology, so the move is invisible to clients and applications. No maintenance windows. No runbooks. No 2 a.m. calls.

The second update is Compliance Reporting and Remediation. Fusion's Workload Presets standardize deployments with prescriptive settings for snapshots, replication, QoS, and more. Over time, configurations drift. Now, when your security team updates a requirement,  for example, bumping Tier 0 snapshot retention from 7 to 14 days, Fusion identifies every non-compliant workload and offers 1-click remediation with a complete audit trail to hand to your security team. No manual sweeps, no error-prone reconfiguration across every array in the fleet.

Together, these features move storage from a passive appliance to an active participant in your security and operations posture. WWT is planning to get these features into our Advanced Technology Center so customers can see them in action before making decisions. If you want to understand where your organization stands today, Everpure's EDC Blueprints offer self-assessments, guided workshops, and road mapping tools to help you build a clear path forward.

XL190 Turbo Mode: because sometimes "fast" needs a boost

Earlier this year, Everpure introduced the FlashArray//XL190 to compete directly with high-end 8-engine systems from legacy vendors. The //XL190 is the largest dual-controller system in the FlashArray family — maximum capacity, maximum performance — in a 5RU chassis. The space and power savings compared to traditional high-end platforms are substantial.

New this conference: Purity Turbo, now generally available via a Purity//FA software update. When databases, analytics, and VMs compete for the same array resources during peak periods or backup jobs, performance SLAs can slip. Purity Turbo allows read requests coming through the secondary controller to be served directly from DFMs or the 3TB of local read cache — bypassing the need to forward every request to the primary controller as the current architecture requires.

One important clarification: Purity Turbo addresses contention scenarios. It's not a mechanism to extract more performance than the hardware was designed for — oversubscribing the system would create its own problems, particularly during controller maintenance. Use it with intent and it's a genuine SLA safety net.

WWT's performance benchmarking lab is going to put this through its paces. We'll share what we find.

ActiveCluster for File: block doesn't get to have all the fun anymore

For years, block workloads have benefited from ActiveCluster — synchronous replication between arrays with transparent failover. Network outages, power disruptions, planned maintenance: the application doesn't notice. File workloads have been watching from the sidelines, hoping for their turn. Their turn is here.

ActiveCluster for File is now available via a Purity//FA update at no additional cost, extending synchronous replication to NFS and SMB file workloads as part of the unified data plane in the Enterprise Data Cloud.

A file system runs active on one FlashArray with synchronous data replication to a second array. When an outage occurs — or you need to move workloads around within your fleet — the full file system personality fails over to the second array within 15 seconds. When the primary comes back online, replication automatically reverses. NFS and SMB clients experience a brief pause, then continue as normal.

No runbooks. No manual backend reconfiguration. Setup is wizard-driven, covering realm creation and file system extension across arrays. For organizations running file-heavy workloads — and most do — this is a meaningful step forward in what a resilient storage architecture looks like without adding cost or complexity.

What this means for your organization

The through-line from //Accelerate '26 is coherent and deliberate. Get smarter about your data with Data Intelligence. Make it AI-ready with Data Stream. Manage it autonomously with Fusion. Protect it with ActiveCluster for File. Do all of it on the most efficient flash platform available.

WWT was present for every announcement, and our team is already putting these capabilities through their paces in our labs. Whether you want to see a demo, pressure-test how these features fit your environment, or simply understand where to start — we're ready to have that conversation.