Hewlett Packard Enterprise, the company better known as HPE, announced today that it acquired Pachyderm, a startup developing a data science platform for "explainable, repeatable" AI. The terms of the deal weren't disclosed nor was the purchase price. But HPE said that it plans to integrate Pachyderm's capabilities into a platform that'll deliver a pipeline for automatically preparing, tracking and managing machine learning processes.
Pachyderm’s software will remain available to current and new customers -- for now, at least. HPE says that the transaction isn't subject to any regulatory approvals and will likely close this month.
Co-founded in 2014 by Joey Zwicker and Joe Doliner, a former Airbnb software engineer, Pachyderm delivers tools for versioning (i.e., creating and managing) "enterprise-scale" machine learning and AI projects. Using Pachyderm's cloud-based and on-premises products, users could automate some aspects of AI system development through data transformations, data workflows and connectors.
Pachyderm also offered versioning features for machine learning datasets and a "Git-like" structure to facilitate collaboration among data scientists, as well as the ability to generate an immutable record for all activities and assets on the platform. It also hosted Pachyderm Hub, a fully managed service with an on-demand compute cluster for AI development.
Prior to the HPE acquisition, Pachyderm managed to attract $28.1 million in venture capital from backers including Benchmark, Microsoft's M12, Y Combinator and HEP's own Hewlett Packard Pathfinder. (Pathfinder invested in February 2022.) Among its customers were Shell, LogMeIn, Battelle Ecology and AgBiome.
HPE sees Pachyderm bolstering its flagship AI development product, the HPE Machine Learning Development Environment, which provides software to build and train machine learning models for applications like computer vision, natural language processing and data analytics. In a press release, HPE lays out what it sees as the major benefits Pachyderm brings to the table, including incremental data processing, visibility on the origin of data and the ability to track different versions of data to understand when it was created or changed.
"As AI projects become larger and increasingly involve complex data sets, data scientists will need reproducible AI solutions to efficiently maximize their machine learning initiatives, optimize their infrastructure cost and ensure data is reliable and safe no matter where they are in their AI journey," HPE EVP of high-performance compute (HPC) and AI Justin Hotard said in a statement. "Pachyderm’s unique reproducible AI software augments HPE’s existing AI-at-scale offerings to automate and accelerate AI and unlock greater opportunities in image, video and text analysis, generative AI and other emerging large language model needs to realize transformative outcomes."
Pachyderm is HPE's second AI-related acquisition since Determined AI in June 2021. Determined AI, similarly, was focused on creating a platform for building and retraining machine learning models.
HPE sees AI and HPC as a potential major profit driver, but the company has struggled to maintain momentum in the increasingly competitive market. In its Q4 2022 earnings report, HPE's HPC and AI revenue dipped 14% year over year to $862 million, bringing the operating profit margin down to 3.5% compared to 14.2% in the prior-year period.