One Year of “Ask Us Everything”: John “Coz” Colgrove on Everpure’s Past, Present, and Future

Over the past year, our uniquely interactive Ask Us Everything series has given IT practitioners direct access to Everpure experts. Audience questions—not a slide deck—drive the conversation. For our first anniversary edition, we welcomed a particularly special guest: Everpure cofounder and technical visionary John “Coz” Colgrove.

The result was a wide-ranging conversation covering Everpure’s origins, some of its biggest technical bets, AI, infrastructure strategy, fleet management, costs, and where data management goes next.

“What was the thinking behind starting the company in the first place?”

Coz summed up the origins of Everpure (formerly Pure Storage) in characteristically straightforward terms: it was his opportunity to “fix everything that ever annoyed me about the storage industry.”

After years of working around complex storage systems, he wanted to build something powerful, but simpler—storage that worked without forcing customers to become experts in unnecessary complexity. Downtime was annoying. Upgrades were annoying. Why make customers deal with them?

That philosophy still influences how Everpure approaches new capabilities. Every feature adds complexity, Coz explained, so the goal is to provide good defaults, minimize configuration and, ideally, remove complexity elsewhere at the same time. And because this is all about your organization’s data, quality comes first.

For storage admins accustomed to legacy platforms, that distinction matters. Those systems accumulate layers of tuning, configuration, dependencies, and all at a cost of operational overhead.

“What technical bet looked risky—but paid off the most?”

When an attendee asked Coz to identify the riskiest technical decision that ultimately delivered the biggest impact, his answer was DirectFlash.

Designing hardware specifically around all-flash managed by Everpure software was already a significant decision for a young company. DirectFlash took that philosophy much further by allowing Everpure to work directly with flash, instead of accepting the behaviors and limitations built into conventional SSDs.

The results included greater efficiency, reliability and much more consistent I/O. Coz shared one striking example: the longest I/O latencies dropped from roughly 600 milliseconds to about 40 milliseconds after the transition to DirectFlash—and those events happened about one-tenth as frequently.

It was a useful reminder that Everpure’s differentiation has never simply been “flash instead of disk.” It comes from rethinking the storage architecture around flash, rather than forcing a new medium into an old model – which sums up the legacy approach to flash storage.

“Will AI agents automate more storage operations?”

Another attendee asked whether AI agents could eventually automate tasks such as updates.

Coz’s answer revealed a lot about Everpure’s approach to automation. The objective isn’t automation for automation’s sake—it’s all about making the experience genuinely easier.

Everpure is already moving further toward self-service upgrades, with Coz estimating that roughly half or more of Everpure Purity OS updates are now self-service. AI agents will increasingly help, but he was cautious about handing support completely over to autonomous agents before the technology can deliver the quality that customers expect.

His philosophy: use AI to augment people today, and expand its role as the technology improves.

“Can I manage all my data from one place?”

Coz was asked a question familiar to almost every storage administrator: Will there eventually be a single pane of glass for block, file, object, and other data?

For Everpure environments, Coz’s answer was yes—and that journey is already underway through Pure1, Everpure Fusion, and the Enterprise Data Cloud. His advice was particularly direct: all customers should be looking at Everpure Fusion for managing their fleets, to enable the benefits of the Enterprise Data Cloud.

However, managing third-party storage through the same methods is harder. Coz noted that inconsistent APIs, changing implementations, and a lack of industry-wide cooperation make it difficult to deliver the level of reliability Everpure would expect from such an experience.

“What should infrastructure teams focus on in an AI-driven world?”

Coz’s answer was one word: adaptability.

AI is moving too quickly for infrastructure teams to optimize around one architecture, model, or use case. What matters is being able to change direction as requirements evolve.

And adaptability, he argued, depends heavily on simplicity.

If infrastructure is standardized, flexible, and managed consistently, teams can respond much faster when workloads suddenly require five times more processing, or when an entirely new AI use case emerges. The Enterprise Data Cloud supports that model by allowing teams to move away from treating dozens of arrays as individually managed systems and instead define consistent classes of storage services across the data estate.

“What can customers do about rising infrastructure costs?”

With component costs increasing, Coz was asked what customers could do to plan for expansion while controlling spending.

Coz’s first recommendation was refreshingly practical: use what you already have more efficiently.

Data reduction can mean that customers have more usable capacity than expected, making fleet-wide visibility important for identifying where capacity or performance is constrained—and where resources remain available. That creates opportunities to rebalance before simply purchasing more infrastructure.

Also, Everpure Evergreen//One as-a-service subscriptions deliver cost efficiency. It keeps costs down by reducing the risk of overbuying, and by spreading infrastructure costs over a longer subscription period rather than tying them directly to current hardware manufacturing costs.

“So what’s next for Everpure?”

Perhaps the biggest takeaway was that Everpure’s next chapter is increasingly about data, not just storage.

Coz said Everpure will continue advancing Everpure FlashArray and FlashBlade, but the larger opportunity is helping customers understand what data they have, where it lives, how copies and snapshots relate to each other, and how that data can be prepared and used effectively for AI.

That direction also helps explain the evolution from humble beginnings to Everpure: expanding beyond “storage boxes”, while preserving the simplicity, quality, Evergreen philosophy, and customer experience that has defined the company from the beginning.

And that made this first-anniversary episode especially fitting. Ask Us Everything was designed as an unscripted platform, where Everpure Community members drive the discussion in every session. One year later, attendees got to direct those questions to one of the people who helped start the company—and then look ahead with him to where storage, data, and AI go from here.

Check out this and all our other Ask Us Everything sessions. And, keep the conversation going by jumping into the Everpure Community.