Neoclouds and data centers

Extend your cloud.
Sell the storage.

We put S3 object storage in your data center. You sell it to your customers inside your own product, right next to your compute.

Talk to an engineer How it works

Your building

YOURSComputeAI, HPC, analytics and app workloads

YOURSWEKA / VAST scratchHot working set, checkpoints in flight

cross-connect

OURSFil One Capacity · HDDArchives, backups, completed jobs, logs<1 ms · no internet path · no egress charge

internet

Beyond your buildingOur resilient cloud

OURSFil One SovereignErasure coded across several data centers. Survives a site loss, with fast transfer and sharing between sites. Stays in-jurisdiction. Deployed in days, not weeks.

Other clouds

>50 ms over the public internet. Egress metered (AWS, GCP) or capped (Backblaze 3× stored, Wasabi fair use).

<1 msTo your compute, on a cross-connect

Your priceResell under your brand, keep the spread

$0 capexWe fund, rack and run the hardware

The problem

Scratch is too expensive
to park data between runs.

WEKA, VAST and local NVMe are priced for the run in flight. Without a cheaper tier in your building, data waiting on the next job ends up leaving.

Scratch is priced for the run

A growing share of what sits on your most expensive storage hasn't been read in weeks.

Nowhere else to go

So customers move it to a hyperscaler or another cloud, and the storage revenue goes with them.

Disks are scarce

High-capacity drives are on allocation, with long lead times. Build your own tier and you wait on hardware while demand walks out the door.

See how a zone fixes this →

The product

Object storage built to be resold.

Local

In your building, on a cross-connect. Under 1 ms to your compute, so data parked between runs stays close and bills as yours.

Regional

Erasure coded across several data centers, so data survives a site loss. Keeps each tenant in-jurisdiction and still bills as yours.

We fund the hardware

Our drives, racks and operations. No capex, no drive lead times, no new hires.

S3 as they know it

Standard S3 API and tools. Per-tenant keys and usage metering feed your billing. Nothing to rewrite.

Talk to an engineerWe fund and run the hardware. You resell it under your brand.

Today · read back from another region

The egress meter goes to zero.

Datasets and checkpoints read back100 TB / mo

Hyperscaler egress, public list$0.09 / GB

Egress alone, per month≈ $9,000

Per year, before a byte of storage≈ $108,000

On a cross-connect in your own building$0

in egress. Scratch holds the working set it was bought for, and the same 100 TB becomes a line on your invoice.

Price a zone

The commercials

You sell it.
You price it.

We bill you one wholesale rate per terabyte, with no egress or retrieval component. You set the retail price your customers see and keep the spread.

White label, end to end

Your brand on the console, the docs and the S3 endpoint

Per-tenant keys and RBAC: no customer reaches another's data

Per-bucket metering exported into your billing system

Partner API to provision tenants from your own tools

SSO through your identity provider

Request wholesale pricing

The pilot

One workload. One site. Start earning.

Pick the workload that annoys you most: checkpoint sprawl, dataset reloads, a backup target you overpay for.

Pick the workload

One cluster, one bucket pattern, agreed numbers.

Regional endpoint live

A tenant on our nearest site while we survey yours. Measure throughput and egress saved.

Go live locally

If the numbers hold, we rack the Capacity tier in your building.

Start earning

Sell it under your brand. Storage revenue that used to leave now shows up on your invoice.

Be first in your region.

We are expanding the network now. The first partner in a region helps shape how its zone is built. All we need to start: a name on the workload, and an hour with the infra team that owns it.