Why fast AI starts with what it runs on

When Amrita Das isn't deep in product strategy or customer conversations, she's on the sidelines of a school game cheering on her two boys. She’s a working mom who moves fast, prioritizes hard, and brings that same clarity to everything she does at Everpure, and she's spent her entire career as a software PM.

She started as a firmware engineer and moved into PM roles across enterprise technology, and eventually landed at Everpure, always on the software side, always focused on the operating system, the ecosystem, the experience that sits on top of the hardware.

That’s why she finds it quietly maddening that people still think of Everpure as only a hardware company.

She left Everpure and returned after a year in Big Tech, a return that made complete sense once she lived the alternative.

And what she's building now is the most ambitious product she's worked on yet: FlashBlade//EXA, Everpure's file storage platform built for AI infrastructure at exabyte scale. Purpose-built for neoclouds and AI factories,the organizations running AI inference across tens of thousands of GPUs, FlashBlade//EXA delivers 10+ TB/s read performance in a single namespace and exabyte-scale capacity, making it one of the highest-performance file storage platforms available for large-scale AI workloads.

The problem nobody was solving

For years, the default assumption in AI was simple: if you need more performance, buy more GPUs. Everything else would keep up.

It doesn't.

“Storage shouldn’t be the bottleneck,” Amrita says. “Everything gets written to something in the backend. It should be able to scale, it should just work. The GPUs shouldn't be idle, because that's your most costly resource.”

That's the problem FlashBlade//EXA was built to solve for a new category of customer building AI infrastructure at a scale the industry hadn't fully planned for: the neocloud.

“We started with just two things,” Amrita says. “Performance and scale. And we found out that even without a full feature set, customers were interested because of what the platform can do, and because of what Everpure has built over the years. They trust us.”

Built for the AI era. Rooted in what works.

What Amrita's team is building isn't just a new product. It's a new chapter for Everpure's entire product portfolio, and that means being intentional about how FlashBlade//EXA grows without losing what makes the rest of the portfolio great.

When she talks to neocloud customers, the conversation comes back to the same things every time.

“As we were having those customer conversations,” Amrita says, “it became really clear: simplicity and reliability aren't nice-to-haves in this market. They're critical. And those are exactly the things Everpure has always been great at. Listening to customers told us that what made FlashArray and FlashBlade so trusted is what this market needs too.”

So from day one, FlashBlade//EXA product requirements document includes the full Everpure ethos: simplicity, non-disruptive upgrades, integration with Everpure Fusion, a place in the Enterprise Data Cloud. Not all of it lands on day one (Amrita is candid about that) but none of it gets forgotten either.

It's a deliberately different approach to how new platforms have been built before, and it positions FlashBlade//EXA not as a standalone product, but as the newest member of the fleet that covers everything from backup to Tier Zero applications.

The halo effect is real

FlashBlade//EXA opens doors for the entire Everpure portfolio.

Amrita describes it as the halo effect: earn trust with one product, prove the performance, and the relationship with customers grows. A neocloud comes in for FlashBlade//EXA, but their infrastructure needs don't stop there. Containerized workloads, block storage, and file and object outside of training and inference: across all of it, customers get a consistent experience backed by a platform they already trust.

“What that means for the customer is simpler operations, fewer vendors to manage, and a partner who understands their full data environment,” Amrita says. “From backup to Tier Zero apps, that's a powerful story for the customer's business.”

One confirmed example: Everpure's first FlashBlade//EXA customer has become a reference, a signal that the performance is real and the relationship is built on results, not just benchmarks.

The strategy for FlashBlade//EXA is grounded and deliberate. “You can be the second vendor, get your foot in the door, prove that you can do this, and then take on the market. We understand we're still building. But we have to be the number one or number two vendor of choice. You make yourself indispensable to the customer, and then you expand. And we do all of this without losing sight of what really matters: the business value. What’s the outcome for the customer? How are we making their lives simpler?”

Why Amrita came back

Amrita left Everpure for Big Tech because she was curious. She wanted to see what product life felt like at a different scale and also really liked the manager, with whom she’s still in touch. What she found was a different kind of organization entirely. The kind of influence she wanted to have (over product direction, customers, and what actually got built) was hard to come by. In a company of that size, the ability to drive outsized impact gets diluted. Decisions travel through many layers which impacts the ability to move quickly.

Amy Fowler, General Manager, Commercial, and Brian Gold, VP of Software Architecture, both at Everpure, kept checking in on Amrita throughout that year. “They were constantly checking in on me,” Amrita says. “It was always, ‘I think you’re making the wrong decision, but I’m going to support you.’ That’s the kind of leaders they are.”

There's something else, too. At a company of Everpure's size, with a product like FlashBlade//EXA at the stage it's at, a PM can change things. The product roadmap isn't set in stone. The customer conversations are real and direct. The decisions you make today show up in the product tomorrow.

“At a massive organisation, you can be excellent at your job and still feel like you're not making a difference,” she says. “Here, you feel it. The work you do matters, not just to the team but to the customers relying on the platform.”

We're on a journey, and that's the exciting part

Amrita doesn't dress up where FlashBlade//EXA is in its lifecycle. “We still have a long way to go to be really successful in this space. At times it does feel like, ‘what did we get ourselves into?’ But it's been so much fun because we're working on problems we haven't worked on before. The scale of the problem, the kind of customers, the kind of requirements we're getting — it's genuinely interesting and challenging.”

Every conversation starts as a point-of-contact, because FlashBlade//EXA is still building its presence in this market. That means more customer conversations, more hands-on engagement, more direct exposure to the problems neoclouds and AI factories are trying to solve, than most PM roles anywhere would offer.

“It's not just working with engineering,” Amrita says. “It's about developing the broader ecosystem of partners, alliances, and customer references. You're doing it from the ground up, for the first time. That's the kind of PM work that actually builds your skills in every direction.”  For Amrita, it boils down to building the foundation that AI runs on. That’s the company Everpure is becoming. And we have everything it takes to get there.”