Our New Data Center — Next to a Dam. Built for Compute.
Just 800m from Beauharnois, our new data center runs on clean hydro — powering millions of low-carbon workflows.

Isn't it a beauty!?
Say hello to the Beauharnois Generating Station, a run-of-the-river hydro plant on the banks of the Saint Lawrence River in Montreal, Canada.
Our newest data center region sits just 800 metres from this powerhouse, making it a perfect addition to our carbon-aware cloud, which always selects the lowest CO₂ regions.
With 36 turbines and over 1,900 MW of capacity, Beauharnois can power nearly 400,000 homes, or run ~24,736,362 jobs in parallel, reducing emissions by up to 90% compared to traditional compute. Each 300-tonne rotor weighs as much as 50 African elephants.
If we maxed out the entire output for an hour, we could run ~24,736,362 jobs in parallel, reducing emissions by up to 90% compared to traditional compute.
Thanks to Hydro Québec for keeping the electrons flowing clean and the great facts on Elephants.
Guestimating Dam Workloads
A rough calculation to figure out how many jobs we could run in parallel using the energy from a single dam. We work backwards from how much energy one job uses to see how far that power could stretch.
1
2const energyPerJobkWh = 0.0037978278;
3const durationSeconds = 178;
4
5// 1. Find energy per kWh
6const energyPerSecond = energyPerJobkWh / durationSeconds;
7const energyPerHour = energyPerSecond * 3600;
8
9console.log("Energy used per job per hour:", energyPerHour.toFixed(6), "kWh");
10
11const totalKWhAvailable = 1900000; // 1,900 MW for 1 hour
12const numberOfJobs = totalKWhAvailable / energyPerHour;
13
14console.log("Jobs that can run in parallel:", Math.floor(numberOfJobs).toLocaleString())
15
Want to build in the Cleanest US Regions?
CarbonRunner automatically shifts workloads to the lowest-carbon regions across AWS, Azure, and GCP — live, based on real-time grid data.