Inside Pure Storage’s Everpure Pivot And How 1touch Adds the Data Context It Was Missing For AI

The acquisition of 1touch by Everpure, formerly known as Pure Storage, is aimed at giving the data storage and management company the data discovery, classification, enrichment, and contextualization capabilities needed for building the AI use cases of the future.

The planned acquisition of 1touch by Everpure, which until Monday was known as Pure Storage, brings the company new capabilities around the classification, enrichment, and contextualization of data as a way to bring that data to AI.

Prakash Darji, Everpure’s manager of digital experience, told CRN that 1touch is a 60-person company focused on data.

“They’ve got connectors that connect to every data source under the sun,” Darji said. “That could be NFS, SMB, S3 databases, or every database under the like sun I could think of like Oracle, SQL Server, Postgres, Mongo. They get into the content of the data.”

Everpure declined to discuss financials around the planned acquisition, which is scheduled to close some time during the second quarter.

1touch found its product market fit many years ago in regulated industries and financial industries finding things like PII (personally identifiable information), regulatory, information, data governance, lineage, and compliance, Darji said. 1touch works as an OEM provider for IBM’s Guardium AI-driven security platform, he said.

The 1touch technology, unlike that of most DSPM, or data security posture management, companies, does pattern matching all the way through to embedding AI classifiers into data characterization, Darji said. For example, he said, it could allow a user to scan an X-ray and find a bone break because it could compare to an SLM (small language model) or LLM (large language model).

“Everpure has been talking about the Enterprise Data Cloud and getting into managing the context of data sets, workloads, and presets and things we’ve been talking about in Fusion,” he said. “That’s great when you get to a certain level. But if you really want to solve problems in the world of the future world of AI, you need context about what’s in the data, such as, this PDF has a P.O. number which is associated with this database, so when you’re training an AI use case for purchasing history, both the data in the PDF invoice and in the database table are associated and matter and can be fed into my training.”

In an example around AI training clusters, Darji said there are problems with bringing data to GPUs and training GPUs.

“Imagine you had an ITOps and storage context, plus a data context, and you use that larger context across storage and data to improve the accuracy, training, and efficiency of future AI use cases,” he said. “This is where we see the world going in a few years. I’m not saying that exists today as a market, but if you followed Pure for any period of time, you know we look at the next decade in flash and the next decade in Evergreen//One and subscription services. Now we’re taking a forward-looking view of where we believe data infrastructure needs to be. That’s why this technology is interesting to us.”

Other companies offer DSPM technologies mainly for identification of private and sensitive data and applying governance and regulated industries around data, Darji said.

“What’s different about this company is, yes, they can do that,” he said. “But this isn’t a bunch of security privacy guys looking to build a business. There are bunch of data guys that happen to also do security and privacy. That’s why we were interested in this company versus others in the space.”

That philosophy fits in with why Pure Storage changed its name to Everpure, Darji said.

“We found that the word ‘storage’ in the Pure Storage name was great in terms of the category we’re in, but it was also category-limiting from doing more than storage,” he said. “We already started getting some pressure last year that people saw unidimensionality when storage was in the name of the company. Largely, the design principle was, let’s drop ‘storage’ from the name of the company so it’s less category limiting. And we thought we should put another classifier in the name, like data. But who knows if in 10 years will that be category-limiting as well.”

Everpure was always focused on bringing evergreen value to customers, Darji said.

“‘Evergreen’ had a lot of brand value to us,” he said. “It is part of who we are. We’re continuously delivering value in an evergreen way, and we will apply that to any category we ever get into. So largely the best attributes of evergreen and pure led us to Everpure.”

Darji said that Everpure did not previously partner with or invest in 1touch prior to this acquisition.

Prior to acquiring 1Touch, Everpure did not have data discovery, classification, enrichment, and contextualization capabilities, Darji said.

“We were in storage management and data set management with Fusion, but we didn’t get into looking into the data,” he said.

Everpure found 1touch via a research team in Darji’s business unit that works with a corporate development team to find potential acquisitions, Darji said.

“We have a lab, and we always evaluate a lot of tech,” he said. “Our approach is never believing in the PowerPoint ‘marketware.’ We usually take tech for a tryout, and that’s where we discovered that the extensibility of the 1touch technology platform could be used in more AI and data-centric use cases versus just DSPM. We’ve tested other technology in the space as well, and we found this to be something that had unique attributes for extensibility and classifiers.”