Loading organizations...
§ Private Profile · Seattle, WA, USA
XetHub is a technology company.
XetHub provides a data management platform scaling Git for machine learning datasets and models, enabling AI teams to adopt software engineering best practices. It operates as a version-controlled blob store, facilitating efficient collaboration on massive, evolving datasets. Content-defined chunking and deduplication ensure optimal storage and streamlined access by integrating changed data portions within Git.
Founded in 2021 by Yucheng Low, Ajit Banerjee, and Rajat Arya, XetHub originated from their experience scaling Apple's internal machine learning infrastructure. Their background at Turi and Amazon highlighted the critical need for a robust, Git-compatible solution for versioning and collaborating on large AI datasets, inspiring XetHub's development.
XetHub primarily serves machine learning and remote development teams, optimizing workflows for large data and models. Beyond AI, the company aims to empower technical teams across hardware, biotechnology, and gaming. Its mission is to facilitate rapid experimentation and innovation with full historical traceability and reproducibility, driving efficient development in data-intensive fields.
XetHub has raised $8.0M across 1 funding round.
XetHub has raised $8.0M in total across 1 funding round.
XetHub has raised $8.0M in total across 1 funding round.
XetHub's investors include Madrona Ventures, Addition, Hyde Park Venture Partners, Liquid 2 Ventures, Pioneer Square Labs, Unlock Venture Partners, Kevin Nazemi, Sujal Patel.
XetHub has raised $8.0M across 1 funding round. Most recently, it raised $8.0M Seed in January 2023.
| Date | Round | Lead Investors | Other Investors | Status |
|---|---|---|---|---|
| Jan 1, 2023 | $8M Seed | Madrona Venture Group | Addition, Hyde Park Venture Partners, Liquid 2 Ventures, Pioneer Square Labs, Unlock Venture Partners, Kevin Nazemi, Sujal Patel | Announced |
XetHub is a technology company that builds a version-controlled blob store enabling machine learning (ML) teams to collaborate on massive datasets and models at terabyte (TB) scale, treating data like code in Git repositories.[1][2][3][4] It serves hybrid and remote ML practitioners and data science teams in technology and data management sectors, solving the problem of slow, error-prone workflows caused by fragmented tools for handling large files, versioning, and reproducibility.[1][2][3] Founded in 2021 in Seattle, XetHub raised $7.5M in seed funding led by Madrona and was acquired by Hugging Face in August 2024, after which its technology integrates into Hugging Face Hub to scale Git for AI development.[1][2][4][6]
The platform offers content deduplication for efficient storage of data versions, time travel on petabyte-scale repos, model comparisons across branches, and tools like data summarization and visualization, now accessible to Hugging Face's global ML community.[3][4][6]
XetHub was founded in 2021 (with operations starting in February 2022) by Yucheng Low (CEO), Ajit Banerjee, and Rajat Arya, all Apple alumni who built and scaled Apple's internal ML infrastructure handling over 100PB of data for dozens of teams.[1][2][3][4] Low previously co-founded Turi (acquired by Apple in 2016), where he advanced ML algorithms and data scaling; Arya contributed expertise in distributed systems from Microsoft, AWS, and Turi; Banerjee brought entrepreneurial experience from Amazon.[2][4]
The idea emerged from their Apple tenure, where they addressed collaboration challenges on massive datasets, leading them to create a Git-scalable solution for external ML teams after leaving in 2021.[2][4] Early traction included a public beta with 20GB free storage, Windows support expansion, and $7.5M seed funding from Madrona, fueling hires and features like data drift detection.[2]
XetHub rides the explosion of AI/ML requiring massive datasets and models, where traditional tools like Git LFS fail at TB scales amid rising remote/hybrid teams and open-source AI collaboration.[4][6] Timing aligns with Hugging Face's dominance in ML model hosting and the shift to "software-style" AI development, amplified by post-2023 AI hype driving demand for reproducible, scalable workflows.[1][4]
Market forces favoring it include exploding data volumes (e.g., >100PB internal scales at firms like Apple) and needs for efficiency in versioning amid talent shortages; its Hugging Face integration influences the ecosystem by democratizing large-scale storage for millions of users, benchmarking superior performance over competitors like DVC or S3.[3][4][6]
Post-acquisition, XetHub's tech will embed deeply into Hugging Face, accelerating features like advanced data drift tools and optimized versioning for all users, potentially redefining ML infrastructure standards.[4][6] Trends like agentic AI, multimodal models, and enterprise AI adoption will amplify demand for its scale, with influence growing via open-source releases and community benchmarks.[6]
As the bridge from Apple's elite infra to global ML, XetHub positions Hugging Face to own AI collaboration at petabyte levels, transforming clunky data workflows into seamless dev practices.