Everpure Unveils Data-First AI Strategy, Targets $21B Intelligence Market

Everpure (NYSE:P) outlined an expanded technology strategy centered on what executives called “data primacy,” arguing that artificial intelligence is pushing enterprises to organize IT architectures around data rather than individual applications.

At its financial analyst meeting in Santa Clara, Chairman and Chief Executive Officer Charlie Giancarlo said the company believes its business is accelerating while its opportunity is expanding into additional high-growth areas. Giancarlo said AI is driving a fundamental architectural shift in which organizations increasingly organize around data, rather than applications, creating new opportunities for the company.

Data management expansion

Prakash Darji, general manager of Everpure’s Digital Experience Business Unit, said enterprises have historically purchased applications and then spent heavily integrating data among them. He estimated that 80% to 90% of enterprise spending goes toward integration work across technology assets.

Darji said the growing use of AI agents could further increase those costs because relevant context is often fragmented across applications, data warehouses and other systems. Everpure’s proposed approach is to create shared context around data wherever it resides, allowing organizations to reuse data in place rather than continually copying it between systems.

“Agents work better directly on data,” Darji said, arguing that data context must persist with the data as it moves across storage locations and performance tiers. He said the company views the storage operating system as a place to maintain persistent context and policy controls for data.

The company positioned its approach against what Darji described as “walled garden” strategies from hyperscalers, application vendors and data platforms that require customers to move data into their respective platforms. Everpure instead aims to provide a neutral data-management layer that can work across cloud, on-premises, mainframe, structured and unstructured data environments.

Ashish Gupta, general manager of Data Management, said Everpure Data Intelligence discovers and classifies data regardless of where it resides, including on Everpure systems and third-party storage. The offering is designed to identify sensitive information and contextualize data based on business processes and relationships among records.

Gupta said customers are increasingly bringing together CIOs, CISOs and chief data officers when buying data-management tools. He said this broader group of stakeholders can increase spending on data-intelligence tools, and cited Everpure’s estimate that the data-intelligence market could grow from $6 billion currently to $21 billion by 2030.

Gupta highlighted several customer examples, including a credit card company that scanned 14,000 databases in two weeks. According to Gupta, the company reduced the time required to respond to a data-subject access request from 21 person-weeks to less than three minutes. He also said a payment processor identified 91 million credit-card records in an unprotected database and reduced its estimate of personally identifiable information records from 189 million to 39 million after identifying duplicated data.

Unified data platform and autonomous operations

Chadd Kenney, Everpure’s vice president of product management, described the company’s Enterprise Data Cloud architecture as a combination of Evergreen hardware architecture, a unified data plane, an Intelligent Control Plane and Data Intelligence capabilities.

Kenney said the unified data plane is intended to consolidate block, file and object data services under a common operating model, reducing the number of management consoles and isolated systems customers must operate. FlashArray is used for low-latency transactional environments, while FlashBlade targets scale-out analytics and AI workloads, according to Kenney.

The Intelligent Control Plane, including Everpure Fusion, is intended to let customers set policies and desired outcomes for data, such as resiliency, location and access rules, while the infrastructure manages those requirements autonomously. Kenney said Everpure receives telemetry from systems every 30 seconds and uses that information to help define infrastructure intent based on operating experience.

Fusion has been adopted by 2,300 customers, Kenney said, adding that adoption has nearly doubled during the past six months. He said adoption includes both existing and new customers, with existing users increasingly considering consolidation of older infrastructure as they shift toward a unified data-plane model.

Kenney also discussed AI-oriented capabilities including Data Stream, which is designed to vectorize data for retrieval-augmented generation workflows, and key-value acceleration technology. The latter stores prior generation history and, according to Kenney, can deliver inference speeds up to 20 times faster while reducing GPU usage and token costs in certain workloads.

Neocloud AI opportunity

Chief Technology and Growth Officer Rob Lee said Everpure is also targeting AI-focused “Neocloud” providers with FlashBlade//EXA. He described Neoclouds as an increasingly important point of concentration for training, inference and agentic AI workloads that enterprises and hyperscalers may not be able to host entirely on their own.

Lee said storage can become a bottleneck for GPU utilization because AI environments require high performance across a mix of training, inference and agentic workloads, while also demanding reliability, availability, security and operational simplicity. He cited a third-party study showing AI clusters generally operate below roughly 85% GPU utilization because of storage bottlenecks, reliability issues or availability constraints.

FlashBlade//EXA extends the company’s FlashBlade platform for larger-scale deployments by separating metadata and data performance layers, Lee said. The architecture is intended to scale performance and capacity independently while relying on open standards rather than requiring software agents on GPU servers.

Lee said the platform ranked first across tested configurations in the most recent MLPerf 3.0 storage benchmarks. He also cited STN, a U.S.-based Neocloud customer, as an early FlashBlade//EXA deployment. According to Lee, STN simplified a multi-layer storage configuration, increased throughput from 30 gigabytes to 40 gigabytes per second to more than 100 gigabytes per second per node, and reported 10% to 20% faster end-to-end inference and training workloads.

During the question-and-answer session, executives said Everpure expects to compete in data security posture management and data-management markets while also partnering with platforms such as Databricks where customer use cases require interoperability. Darji said the company’s strategy is based on the expectation that enterprise data will remain distributed across multiple systems rather than consolidate into a single vendor environment.

About Everpure (NYSE:P)

Everpure is a water-treatment and filtration brand owned by Pentair plc, rather than a separately listed company trading under the NYSE symbol P. The brand provides filtration and water-conditioning solutions for residential, commercial, foodservice, and beverage applications.

Its product portfolio includes water filters, cartridges, heads, manifolds, reverse-osmosis systems, scale-reduction products, and related treatment equipment. Everpure systems are used to improve water quality and help protect equipment in applications such as drinking-water dispensers, coffee and tea equipment, ice machines, restaurants, and other foodservice operations.

Everpure was established in 1933 and serves customers through Pentair’s global water-solutions business.