Introducing Frame Compute

Distributed Data Centers

Delivering hyperscaler levels of AI compute faster and cheaper than conventional data center buildouts.

The Micro Data Center

Frame Compute is Canada's first distributed data center using existing infrastructure and underutilized power capacity to meet expanding AI compute demand.


It's designed to deliver the compute of a hyperscaler, faster and cheaper than conventional data center buildouts.

Up to 64 PFLOPS of FP4 AI Compute

Frame Compute FC-1

What's Inside

Choose from three GPU density configurations

FC-1 Overview FC-1 Detail 4x GPU Node 8x GPU Node 16x GPU Cluster

A Giga Problem

AI is transforming global energy demand, ending decades of flat growth.

By 2030, U.S. electricity demand could outpace supply by over 1,300 TWh due to infrastructure bottlenecks.

Supply forecast rises from 20 to 25 GW by 2028. Reference demand rises from 20 to 74 GW. High demand rises from 20 to 90 GW.
Historical Supply forecast High demand Demand gap

1.3K TWh

Projected demand gap by 2030

7K TWh

High demand forecast by 2030

5.7K TWh

Supply forecast by 2030

Sources

Historical: EIA Monthly Energy Review, Table 7.6
Supply forecast: EIA Annual Energy Outlook 2025 Reference Case
Reference demand: IF / Pew Charitable Trusts, America's Electricity Future
High demand: Grid Strategies, National Load Growth Report

Tapping into
Unused Capacity

Frame Compute's micro data centers connect to existing behind-the-meter power infrastructure in homes and businesses, unlocking untapped capacity for AI inference workloads.


Up to 180A ready to compute.

Home Home

140A

available

Business Business

180A

available

Underutilized Capacity
Safety Reserve
Typical Max Power Use
Sources

EARN ON INFERENCE

Distributed Datacenter Model

AI Compute Request
Frame Compute Sand Beige Node $0.00
Frame Compute Sand Node $0.00
Frame Compute Lakeside Black Node $0.00
Frame Compute Lakeside Sand Node $0.00
Frame Compute Oceanside Light Node $0.00
Frame Compute Oceanside Black Node $0.00

Frame Compute at Home

Calculate Host Income

at $1.07/GPU/hr

12¢/kWh

Conservative

Est. Cost: $152

$0/mo

56% util.

Expected

Est. Cost: $214

$0/mo

86% util.

Optimistic

Est. Cost: $242

$0/mo

100% util.

Conservative / Expected / Optimistic based on current marketplace utilization
*Electricity cost estimated at selected rate per kWh.

Sources

Estimates use current marketplace pricing and utilization snapshots. Actual earnings depend on your listing price, machine reliability, GPU demand, and availability.

Monthly demand prices - vast.ai

Frame Compute Cloud

Frame Compute Cloud turns a vast network of individual AI chips into one powerful, unified system.


Our smart coordination layer automatically manages this fleet to run AI tasks closer to you. This ensures lightning-fast performance, no matter how large the workload grows.

Frame Compute Node

Purpose-built to be easily installed outdoors. Each node is self-contained, quiet, modular, serviceable, and secure.

Compute

  • Up to 16x NVIDIA RTX PRO 6000 GPUs
  • 4 AMD EPYC CPUs
  • 3 TB Memory
  • Fiber Networking

Cooling

  • 35,000 BTU Heat Pump
  • Liquid Cooling to Compute Module
  • 60 dB Quiet Operation

Resilience

  • 15 kWh Battery
  • Reliable Power for Node
  • Home Backup During Outages

Frame Compute at Home

Host Benefits

Compute Income, High-Speed Internet, Battery Backup, and Supplementary heating.

Compute Income

Your compute unit will be sourced to the Frame Compute Cloud generating a recurring income

High-Speed Internet

Fiber connectivity that benefits both the compute node and your home network.

Battery Backup

Power bank backup to supply your home and the compute unit during outages.

Supplementary heating

Excess heat from compute nodes help supplement in-home heating.

Light Controls