Deploy. Serve.
Earn.
Every cluster is 1,000 RevoFi devices plus one full AI server. The devices bring AI, connectivity and site services to the places customers work; the server adds private GPU capacity close to them. Earnings are shared by one published waterfall.
What a Cluster Gives Your Sites
Raven works without any hardware. A cluster is the on-site extension: it puts devices where customers work and pairs them with dedicated GPU capacity, all on the same account and bill.
Site connectivity
WiFi 6 mesh, WiFi offload and Ethernet on every device; 5G cellular on Link Pro.
Site services
Managed firewall, SD-WAN and venue analytics — the services that need a device on-site.
AI where it fits
Light, latency-sensitive work runs on the device; larger open models run on the cluster’s AI server.
One account
Raven tiers, AI usage, site services and connectivity on one account, one memory and one bill.
Cluster Specification
One cluster: 1,000 Link or Link Pro devices on-site, plus one full AI server in the cluster configuration below.
Edge Devices
1,000 × RevoFi Link / Link Pro
Per device
Role: site WiFi and offload, cellular, managed firewall, SD-WAN, venue analytics, local inference. Full device specs →
AI Server
1 per cluster · 2U chassis · 4× NVIDIA H200 NVL
4× NVIDIA H200 NVL · 141 GB HBM3e per GPU
Role: open-model inference for AI usage, private GPU capacity and isolated customer workloads.
Workloads run where they fit
On the device
Network services, site analytics and lightweight models close to users and sensors.
On the cluster server
Larger open models and dedicated GPU capacity, delivered through the same Raven account and AI usage meter.
Who Funds, Owns and Earns
Devices and the AI server are funded separately. RevoFi funds and owns every cluster server; the devices are funded by the people who deploy them.
1,000 devices
Cluster AI server
Site services and Raven tiers
1. Gross billings
What the cluster’s sites pay
Site services, connectivity, Raven tiers, AI usage and server compute
2. Less delivery costs
What it costs to deliver them
All RevoFi costs of providing those services
3. Eligible net earnings
Shared by the waterfall
Collected revenue after delivery costs. Device purchases are never part of it.
The Published Waterfall
Device capital is recovered first. After that, long-term earnings shift toward the partners who serve the sites and the company that runs the platform.
Phase 1
Until LP device capital is recovered · per $100 of eligible net
Phase 2
After recovery · per $100 of eligible net
When does Phase 1 end?
When the LP’s cumulative 80% share equals the device capital. For a $1M cluster, that point is reached after $1.25M of cumulative eligible net earnings ($1.25M × 80% = $1M). How quickly that happens depends on what the cluster’s sites use.
Direct Device Owners
Owners who buy their own devices hold both the LP and channel roles, so the two shares combine. No minimum is required — one device or a thousand follow the same waterfall. It is always a share of eligible network earnings, never a share of the device price.
Phase 1
Until the owner’s device cost is recovered
Phase 2
After recovery
Ways to Participate
Four roles share one cluster. One organization can hold more than one of them.
Fund the devices
Provides the device capital for a cluster and owns the devices.
Share: 80% until capital is recovered, then 20%
Open and serve sites
Brings the cluster to its customers’ sites and keeps them served.
Share: 4%, then 20% after recovery
Own your devices
A site owner or individual who buys devices at list and holds both roles.
Share: 84% until device cost is recovered, then 40%
Run the platform
Funds and owns the AI server and operates Raven, compute, provisioning and billing.
Share: 16%, then 60% after recovery
Where Clusters Fit
A cluster spans many sites. A partner portfolio of venues, stores or offices can share one cluster’s devices and server.
Hotels, venues and campuses
Guest and staff WiFi with offload, managed firewall and venue analytics, with Raven for the team running the property.
Typical services: site WiFi and offload, venue analytics, Raven Business
Stores and multi-location chains
Separate secure networks for customers and point-of-sale, SD-WAN across locations and on-site analytics of foot traffic and queues.
Typical services: managed firewall, SD-WAN, venue analytics
Teams that keep data close
Open models on the cluster’s AI server for organizations that want private or dedicated GPU capacity alongside frontier models.
Typical services: Raven Enterprise, AI usage, server compute
Plan a Cluster
Tell us about your sites, customers and the role you want to play. We’ll help you size a deployment.