Everpure, formerly Pure Storage, announced a strategic shift towards a "data primacy" model at its Accelerate event, asserting that data, not applications, should be the core asset in enterprise IT. CEO Charlie Giancarlo highlighted that the biggest obstacle for enterprise AI is not the models themselves, but the complex data pipelines required to feed them, with a survey indicating 94% of IT leaders see data quality as crucial for AI value.
To facilitate this new architecture, Everpure unveiled several key offerings. Everpure Data Intelligence, derived from the 1Touch acquisition, discovers, classifies, and contextualizes both structured and unstructured data across hybrid environments, creating an enterprise-wide semantic knowledge graph. This is designed to map data dependencies and provide real-time context for AI agents, thereby minimizing AI token costs and enhancing compliance through automated governance of sensitive data like personally identifiable information.
Additionally, Everpure introduced Data Stream, a full-stack solution built on the NVIDIA AI Data Platform, aimed at automating the Retrieval-Augmented Generation (RAG) pipeline. This product ingests, curates, classifies, embeds, indexes, retrieves, and generates data, reducing what previously took weeks or months into minutes. The company also announced performance upgrades to its Enterprise Data Cloud platform, emphasizing a move to ensure data is "AI-ready" by making it self-describing and embedding its semantics and relationships as metadata.
Giancarlo explicitly addressed claims of profiteering amidst a memory crunch, stating that the company's transition to a data-centric model will help customers achieve token and storage savings. He also made a bold claim about the company's continued strength in traditional storage, stating Everpure plans to "ship more flash than really anyone else" as it continues to outspend rivals in storage research and development. The new strategy is presented as an inversion of the traditional application-centric hierarchy, arguing that AI necessitates data becoming the primary asset, with applications serving as consumers of this data.
The company's vision extends to providing universal access to data, not just on Everpure's own arrays, but across public clouds like AWS and third-party storage arrays from vendors such as Dell or NetApp, utilizing over 175 different connectors to reach various data repositories.
