Make America Green Again

“You need double the electric of what we have right now.”

President Trump's advisers told him just how important energy was to the future of AI.

They are absolutely right about the scale of the challenge. By 2030, the US economy will consume more electricity for data processing than for manufacturing all energy-intensive goods combined (aluminum, steel, cement, chemicals).

Drilling down on fossil fuels seems to be the answer from the Whitehouse, using AI as a lever to open up opportunities for fossil fuels; the old classic line of drill, baby, drill continues.

Behind every AI model, every cloud app, every CI/CD pipeline, there’s a data center consuming electricity. Millions of servers, stacked in warehouse-sized buildings, pulling power 24/7 to fuel the rise of compute.

Since 83% of data centers are located in regions heavily reliant on fossil fuels (over 100gCO₂e/kWh), rising energy demand doesn't just strain the grid, it also drives up carbon emissions from electricity generation.

Our answer? Follow physics, not politics. Electrons, not elections.

We're shifting compute to data centers in the lowest-co2 regions, within the US (where possible) but also around the world.

Following Carbon-Aware Compute, Not Political Power

Instead of being fixed and hard-coded, what if regions were flexible?
What if your workloads could flow where the grid is cleanest; live and intelligently?

Why keep running jobs in Virginia at 488gCO₂e/kWh
when you could shift them north to Québec at 30gCO₂e/kWh
or to California when the sun is shining
or Seattle and Washington when hydro is humming?

Turns out... you can. And you don’t need to wait for government or grid reform to do it. It's a one-line code change that saves up to 90% of carbon emissions.

What Compute Can Be Shifted?

Not all workloads are created equal.

Batch processing, CI/CD builds, and training jobs can run anywhere without impact. GitHub Actions are perfect candidates for selective shifting—they're non-latency-sensitive workloads that can chase clean energy across regions.

• GitHub Actions (locked to Azure): average intensity → 302gCO₂e/kWh
• CarbonRunner (live, multi-cloud, carbon-aware): average → 25gCO₂e/kWh
• That’s a 12x improvement, just by shifting to cleaner regions

And it’s not just cleaner; it’s often cheaper, and makes for a more resilient infrastructure; espeically in times for heatwaves.

Where Can It Be Shifted?

We respect the real-world constraints of infrastructure, compliance, and sovereignty. That’s why CarbonRunner supports flexible region preferences — so you can shift compute within the boundaries that matter to you.

Want to stay in the US? No problem! We’ll prioritize the cleanest US regions available; like the Oregon or Washington when hydro is flowing or California during solar peaks.

Don't mind going out of the country? Canada offers some of the best low-carbon regions in the world, with Québec and British Columbia running almost entirely on hydropower. They’re ideal for sovereignty-conscious workloads that still want strong carbon savings.

Open to global shifting? If your workloads aren’t restricted by region, we can push even further — to Norway, Sweden, or France, where grids are powered by hydro, wind, and low-emission nuclear.

By default we utilise all of our CarbonRunner Cloud, finding the greenest regions for you and running your workflows there. It runs across multiple clouds like AWS, Azure, GCP and other providers, in 46 low-carbon regions. This brings resiliance to the forefront.

Even inside constraints, there are better options.

Isn’t It Just a Load of Hot Air?

A heatwave is sweeping across the US. With millions cranking up air conditioning, electricity demand is surging, and the grid needs to respond quickly. In moments like this, it often falls back on fossil fuels to meet the spike.

Even in places like Oregon or California, which usually lean on hydro or solar, extreme weather shifts the equation. Hydro can be down due to drought. Solar fades at night. And gas fills the gap.

That’s why carbon-aware compute can’t just rely on historical assumptions. It needs to track the grid in real time — because what’s clean today might be dirty tomorrow.

Sometimes, the regions we think are green… aren’t, they are forever changing; You have to listen to the grid in real time

Let’s be clear: we’re not shutting down gas plants with carbon-aware compute… not yet. But we are helping them run less often.

By choosing cleaner regions and shifting compute away from fossil-fueled peaks, we ease demand right where the system is most fragile.

And when we pick cloud providers that invest in Power Purchase Agreements (like GCP), we’re helping to bring more renewables online — not just avoiding emissions, but funding their alternatives.

If we can lower the peaks, we use less carbon. Less burning. Less strain. Less… gas.

Some may still be full of hot air — we’re just trying to pass wind more responsibly.

So What Do We Do About It?

Clean energy isn’t static. It varies by region, by season, by hour.

To truly reduce emissions, we need to shift compute based not just on where the power usually comes from — but on where it’s clean right now.

That’s what CarbonRunner does.

It listens to the grid. It watches the carbon. It routes jobs in real-time to where electricity is greenest, moment by moment.

When hydro flows in Washington? It runs there. When solar peaks in California? It shifts. When the grid goes dirty? It reroutes.

With access to over 46 low-carbon regions worldwide, your compute becomes more resilient — even during heatwaves, water scarcity, and rising baseload demand.

Conclusion: Invisible Data, Tangible Impact

We have a real opportunity right now to take meaningful action on the invisible compute that is driving energy demand across the world.

As heatwaves intensify and data centers strain already fragile grids, the choices we make — about where and when our workloads run — carry real weight.

By supporting greener grids and choosing carbon-aware infrastructure, we can shape a future where our digital tools work in harmony with the planet. These systems power our productivity every day. It’s time we power them more thoughtfully.

This isn’t a silver bullet — but it’s a scalpel. Less peak demand = less fossil response. That's the lever.

Some want to drill. We prefer to deploy.

AI may not be "his thing," but energy is everyone’s problem.

Let’s Make America Green Again — with code, not crude.

---

Want Some More to Read?

• NY Times - Renewable Energy Is Booming in Texas
• McKinsey - The cost of compute
• Grist - Trump and the energy industry are eager to power AI with fossil fuels
• Ember - US Electricity 2025 Special Report
• The Guardian - Eastern US swelters from heatwave
• Heatwaves Strain Power Grids as Data Centre Demand Peaks
• Heat Waves Endanger Data Centers That Power AI