Pure Storage Is Now Everpure. Here’s Why CDOs Should Care.

The storage industry has a long and undistinguished history of companies reinventing themselves right before someone else makes them irrelevant. Pure Storage is not doing that. When it announced on February 23 that it was changing its name to Everpure and acquiring AI data intelligence startup 1touch, it was pivoting toward a fight it had already quietly won the right to enter.

The distinction matters. Pure Storage, founded in 2009, built a genuinely exceptional company: USD3.7 billion in annual revenue, an 84 Net Promoter Score that is the highest in its industry by a distance, and an efficient hardware architecture that consumes one-fifth the power of comparable all-flash arrays.

The problem wasn’t the product but the label. “Storage company” had become a ceiling, blocking access to data scientists, application developers, and CDOs, who are now the most important buyers in enterprise technology and have no obvious reason to return a call from someone selling flash arrays.

So, Pure Storage, now as Everpure, demolished the ceiling.

The door that kept closing

Nathan Hall, vice president and general manager for Asia-Pacific and Japan at Everpure, is diplomatic about the brand problem but precise about its consequences. “Our innovation has gone far beyond where our brand currently is,” he said in a recent media briefing. “We need our brand to shift perceptions of our company, of our offerings — from storage to data management.”

The truth that CDOs know but often too late is that storage is where most enterprise AI projects quietly die. Not from lack of compute or bad models but from data that is siloed, unlabeled, and weeks stale by the time it reaches inference. In reality, it can be even worse when a company has its data spread across repositories that no single team has fully mapped. “Data is the fuel for AI,” Hall noted. “And that data has to be prepared to actually be useful — and that takes a lot of human effort.”

The new name splices “Pure” and “Evergreen,” with the latter being the company’s signature architectural promise of non-disruptive upgrades and hardware that never ages out. For its customers, it means never having to plan a forklift migration again. Embedding that promise directly into the brand means the conversation starts at the first meeting rather than the third year of the relationship. That’s not a small thing in enterprise sales.

The acquisition that changes the argument

A name change is a press release. The 1touch acquisition is the strategic signal.

1touch brings the ability to discover, classify, and enrich data across any environment, from SaaS to the edge, at the moment data is created, before it ever needs to be “prepared.” This matters enormously because the current alternative is a long, manual chain of ETL processes that takes weeks to run and produces outputs that are already stale by the time they reach a model. Everpure’s vision, now architecturally within reach, delivers enterprise data to the inference layer already described, contextualized, and AI-ready. The data knows what it is.

“Data is the lifeblood of the AI era, but without the proper controls and semantic context, it remains an untapped resource,” said Ashish Gupta, chief executive officer and president of 1touch, in the rebranding press release. He added that the combined company would aim to “provide a level of contextual intelligence that is unmatched in the industry — giving customers the foundation they need to move AI projects from pilot to production at record speed and trust.”

For many CDOs who have watched AI pilots stall at the data preparation stage, this is a direct answer to the most common reason enterprise AI doesn’t scale. The deal is expected to close in Q2 FY27.

Three products, one argument

The name change and acquisition didn’t arrive alone. Three product announcements, released in quick succession, show a company that has been building toward this moment.

The headline is Evergreen//One for FlashBlade//EXA, which extends Everpure’s as-a-service consumption model to its most powerful AI storage platform. FlashBlade//EXA recently set the highest score ever recorded on the SPEC Storage AI_Image benchmark, sustaining over 6,300 simultaneous AI training jobs, moving data twice as fast as its closest competitor from less than half a rack of space, and sustaining over 90% GPU utilization across large NVIDIA Hopper clusters. In an AI data center, idle GPUs are the most expensive line item on the balance sheet. EXA keeps them working.

Everpure Data Stream, entering beta later in 2026 and co-engineered with Supermicro on the NVIDIA AI Data Platform reference design, tackles the last-mile problem: even with exceptional storage, data still has to travel from where it lives to where the model can use it. Data Stream automates that orchestration entirely, cutting the manual handoffs that slow pipelines and introduce errors.

ActiveCluster for File, announced March 12 and available in Q2 2026 via a non-disruptive Purity OS upgrade, solves a more foundational issue. In most enterprises, file data is locked to the specific array where it was written, making migrations painful and mobility nearly impossible. ActiveCluster introduces fleet-wide, policy-driven mobility for file workloads: data moves based on SLAs, not human intervention. “We’ve moved from hardware-to-data thinking to an app-to-data model,” said Shawn Hansen, vice president and general manager of FlashBlade and FlashArray at Everpure, in a separate press release.

Read separately, these are product announcements. Read together, they form a coherent stack with performance at the infrastructure layer, automation at the pipeline layer, and intelligence at the data layer. It describes a platform rather than a product line and is a materially different conversation than the one Pure Storage was having two years ago.

The sleeper issue: Sovereignty

Buried beneath the AI narrative is something CDOs in Asia-Pacific will recognize immediately: data sovereignty is no longer a compliance footnote. As governments across the Asia Pacific fret over where data physically lives and who controls it, it is becoming a strategic decision, not an IT one.

Everpure embeds geofencing policies directly into datasets at the storage layer, so compliance metadata travels with the data rather than relying on a separate governance layer to enforce it. For organizations managing data estates across multiple regulatory jurisdictions, this architectural choice eliminates an entire category of risk.

Hall calls this approach an accelerant for new conversations. And in a region where data localization laws are multiplying faster than most legal teams can track, and existing ones are being made tighter with stronger enforcement, he’s right.

The honest assessment

Everpure is not a finished product. The 1touch acquisition hasn’t closed. Data Stream is in beta. The full vision — an enterprise data estate that contextualizes itself at the point of creation, eliminating ETL pipelines as a category — is still being assembled. Anyone presenting this as a solved problem is getting ahead of the engineering.

But the direction is coherent, the underlying platform is genuinely best-in-class, and the R&D commitment is real. Everpure spends roughly 20% of revenue on research and development — Hall claims that that is a significant gap over their nearest competitor. Nevertheless, that gap compounds over time.

“It’s no longer just about the storage,” Hall said. “It’s really about the data itself and the context of that data — that’s what becomes essential to having success in AI.”

Infrastructure companies that successfully make this kind of transition tend to become indispensable in ways that are very hard to displace. Whether Everpure fully delivers on that ambition in the next 18 months is the only thing worth watching.

Image credit: iStockphoto/Toms93