# Sustainable compute by design

## The datacenter that doesn't exist.

Traditional datacenters consume massive energy, water, and resources. Fabric creates the same compute power using devices that already exist.

### Traditional 1GW Datacenter

What OpenAI & Meta are building

- **$30-50B** Build cost  
- **10-15yr** To construct  
- **3.5M** Tons CO2/year  
- **Billions** Gallons water

### Fabric Distributed Network

Same compute, zero footprint

- **~200x** Smaller build cost  
- **Months** To deploy  
- **~0** Additional CO2  
- **0** Water needed

## Why Fabric is different.

The first datacenter that grows like a social network, exponentially, not linearly. No construction. No cooling towers. No environmental destruction.

### No Water Consumption

Traditional datacenters use millions of gallons of water daily for cooling. Fabric devices use ambient room temperature. Zero water needed.

### No New Construction

No land acquisition, no permitting, no building massive facilities. Fabric taps into compute that already exists in homes and offices.

### Idle Power Utilization

Your computer uses nearly the same power whether idle or working. Fabric puts that wasted energy to productive use.

### Drastically Lower Carbon

By avoiding dedicated facilities with industrial cooling, Fabric reduces CO2 emissions by orders of magnitude compared to datacenters.

## Fabric vs. Traditional

For equivalent 10MW compute capacity (100,000 nodes)

### Traditional Datacenter

- ✕ 87,600 MWh/year Energy consumption  
- ✕ 35,000 tons CO2/year Carbon emissions  
- ✕ Tens of millions of gallons Water for cooling annually  
- ✕ $500M - $1B Build cost, 4-6 years

### Fabric Network

- Uses existing idle power No additional energy infrastructure  
- Near-zero additional emissions Devices already running  
- Zero water for cooling Ambient room temperature  
- $15-25M Projected Investment, 12 months to deploy

## Scaling without destruction.

At a million nodes, Fabric enters the exaflop era with 2 exaflops of distributed compute, matching the footprint of a regional cloud datacenter. At 10 million nodes, Fabric passes 1 gigawatt of compute and exceeds 20 exaflops.

- **10,000 nodes = 1 megawatt cluster ~20 petaflops**  
- **1M nodes = exaflop era ~2 exaflops**  
- **10M nodes = hyperscale 20+ exaflops, 1GW**

## Ready to get started?

Create your free account and start earning passive income in minutes.

[Sign Up Free](https://fabric.carmel.so/)