Photo: Tara Winstead / Pexels
Aptos moves Shelby storage into private beta with real customer workloads
Three early enterprise customers are now running AI, spatial intelligence, and 3D rendering jobs on the decentralized storage network
Shelby, the decentralized storage platform built by Aptos Labs with Jump, has moved into private beta. It is now handling real customer workloads.
Shelby entered private beta in October 2026. That came after its testnet shut down on August 3, 2026.
Three early enterprise customers are on board. Each one stresses the system in a different way.
Teepin represents the enterprise AI side. AI pipelines read huge amounts of data repeatedly, so storage speed directly shapes how fast models train and serve.
Pictor Network works on distributed 3D rendering. Rendering jobs split heavy scene files across many machines, then reassemble the results.
PathPulse focuses on spatial intelligence. That generally means working with location and mapping data, which tends to be large and constantly updated.
Shelby is pitched as a verifiable global object storage system tuned for AI workloads. It targets sub-second read performance.
Shelby is S3-compatible. S3 is the storage interface popularized by Amazon Web Services, and many developer tools already expect it.
AI, tech, and the markets they move—in one daily briefing.
Daily. Free. Join 34,000+ readers across crypto, finance, and policy.
The blockchain part sits underneath. Aptos smart contracts coordinate the network and verify data integrity.
Every write produces a sequenced cryptographic record on-chain.
Shelby also uses erasure coding. This technique splits data into fragments with added redundancy, so files can be rebuilt even if some pieces disappear.
The design also aims to cut the number of data copies made during computation.
Shelby is specifically going after “hot” storage. That means data needing frequent, fast access, as opposed to cold archives that sit untouched for months.
Aptos Labs first announced Shelby in mid-2025. The testnet phase focused heavily on AI and distributed computing use cases.
Core performance during that period held up well. Testers came away confident that the underlying system worked.
Feedback flagged a need for better developer tooling and stronger API compatibility.
So the team chose to polish the developer experience and operational compatibility rather than keep expanding network capabilities.
Aptos is trying to compete on cloud-like performance while adding something conventional clouds do not offer by default: on-chain proof of data provenance.
For the APT token, the link is indirect but real. APT is used for transaction fees and coordination within the Shelby framework.