Infrastructure / Storage & Retrieval

AIStore

AIStore is a distributed storage cluster built for AI workloads, presenting an S3-compatible object interface over local drives while also operating directly on remote cloud buckets. It scales out linearly, runs with or without Kubernetes, and offloads ETL and batch transformation into the storage layer so training jobs read transformed data in place. NVIDIA develops it.

Verified 2026-08-18 via the GitHub API and the LICENSE body.

Openness

5 high confidence

5.0

license: MIT(OSI)
source: public
core-gated: ungated

MIT license body confirmed. The repository is public and unarchived and builds the whole product, and the README describes no paid tier, enterprise edition or license-gated build beside it, so source is public and the core ungated.

https://github.com/NVIDIA/aistore/blob/main/LICENSE recorded 2026-08-18
LICENSE file is the verbatim MIT text
https://api.github.com/repos/NVIDIA/aistore recorded 2026-08-18
Repo metadata - license spdx_id MIT, private false, archived false, default branch main - for NVIDIA/aistore.
https://github.com/NVIDIA/aistore/blob/main/README.md recorded 2026-08-18
README describes a lightweight distributed storage stack for AI applications that scales out linearly, deploys with or without Kubernetes from one Linux machine to a bare-metal cluster, and operates natively on in-cluster and remote data, with no paid tier, enterprise edition or license-key-gated build beside it.

Adoption

2 low confidence

2.0

1,914 GitHub stars, which lands in the 1K-10K stars band of the stars scale, level 2. The `aistore` package on PyPI is the client-side SDK for talking to a cluster, not the cluster itself, so its downloads measure client installs. With no deployment figure published the stars scale applies, and at 1,914 stars it bands at 2.

https://api.github.com/repos/NVIDIA/aistore recorded 2026-08-18
Repo metadata - stargazers_count = 1,914 - for NVIDIA/aistore.

Capability

4 medium confidence

4.0

Banded on the category feature matrix as complete retrieval system, distributed or cluster-capable. Placed one band below the vespa anchor, on a matrix that bands on how much of the persist-index-retrieve path a product provides itself, at what scale and over how many retrieval modes.

https://github.com/NVIDIA/aistore/blob/main/README.md recorded 2026-08-18
README still documents a lightweight distributed storage stack for AI applications that scales out linearly, deploys with or without Kubernetes from one Linux machine to a bare-metal cluster, and operates natively on in-cluster and remote data.

Verified 2026-08-18