In an earlier post, we shared what Private Beta means for Shelby: controlled customer onboarding, built around real data and real workloads.
Now the first of those workloads are running.
Teepin, Pictor Network and PathPulse are among the first teams working with Shelby through Private Beta, bringing three different data and compute patterns onto the platform.
That range is intentional. The aim now is to prove Shelby’s performance across real applications, datasets and infrastructure - and to start turning our product thesis into measurable customer evidence with paying customers from the outset.
Meet who’s building
Teepin: enterprise AI around proprietary data
Teepin is building a cloud-services layer for enterprises that want to use proprietary or regulated data with open-source models.
Its platform gives enterprise teams a familiar environment for storing data, deploying models, running batch jobs and scaling AI pipelines while abstracting the underlying infrastructure.
The rollout onto Shelby starts with data, creating a consistent layer that can support Teepin as it works across multiple GPU providers, with broader compute services to follow.
That gives Teepin more flexibility to switch between providers without having to rework the data layer each time. For Shelby, it shows how enterprise AI teams can preserve more choice over where their compute runs.
Pictor Network: data for distributed 3D rendering
Pictor Network is building infrastructure for distributed 3D rendering, where rendering workloads can draw on compute across a broader network rather than remaining tied to a single environment.
For a distributed rendering network, the relationship between data and compute is particularly important. GPU capacity can exist in different places, while the assets required to run the workload still need to be accessible to that compute.
Pictor is an early example of Single Address, Global Access applied to a workload where compute is inherently distributed. The team has already integrated Shelby into its development environment and reported a smooth initial integration experience.
PathPulse: spatial intelligence from video
PathPulse is building a spatial-intelligence platform powered by a large fleet of drivers capturing real-time video of road and traffic conditions, including potential traffic violations.
That can generate petabytes of data in a short period of time, which then needs to be uploaded, processed and delivered across different geographies.
For Shelby, PathPulse shows how the data layer can support applications where large volumes of data are generated continuously across many locations, while processing and delivery may happen elsewhere.
Three workloads, three different paths
These workloads show where Shelby fits best: when data and compute do not stay in one place, and the data layer needs to keep up with changing infrastructure.
Teepin starts with enterprise data and expands toward compute services. Pictor operates across distributed GPU capacity. PathPulse creates large volumes of video data that may need to be processed in different environments.
Across those workloads, the common thread is the same: data needs to remain usable as compute changes.
That is where Shelby can create the most value - reducing the operational work around copies, staging and synchronization, making it easier to use compute in another environment, and giving teams more freedom to choose infrastructure based on the workload rather than where the data already sits.
What comes next
This is just the first cohort of customers active in Private Beta. We’re already onboarding additional teams across use cases including distributed training, cross-cloud analytics and AI video generation.
As these customer engagements progress, we’ll share more detail on each use case, how much data is involved, what changed once Shelby was introduced, and the results we can measure.
We’ll also begin opening Shelby to more developers, alongside new capabilities shaped by what we’re learning through Private Beta.
That will give us a clearer picture of where Shelby works best and why it matters across different use cases.
If these challenges sound familiar, we’d like to hear from you.
Private Beta gives customers the opportunity to work directly with the Shelby team and help shape the capabilities built around real infrastructure needs.