Sunny Side Up
AI and Climate: The Math Actually Works
AI’s energy use makes headlines. Its climate benefits don’t. Here’s the full picture.
You’ve seen the headlines: AI is an energy hog. Data centers are boiling the planet. But the full story is more interesting — and more hopeful — than the doom narrative suggests.
Yes, AI uses energy. Here’s the context.
Data centers do consume significant electricity. The International Energy Agency projects global data center demand will rise to around 945 terawatt-hours by 2030. That’s a big number.
But here’s what the University of Waterloo found when they actually ran the numbers: AI’s overall influence on worldwide emissions is small. The effects on climate “are not that significant” at a global scale.
Local impacts? Yes, some regions will feel strain from data centers. But nationally and globally? It barely registers.
The benefits nobody talks about
Here’s where it gets interesting. The World Economic Forum projects that AI’s net impact on emissions is “overwhelmingly positive” — if it’s intentionally applied to low-carbon technologies.
What does that look like in practice?
Wind energy optimization
Google DeepMind’s wind energy AI has boosted renewables’ economic value by 20%. That’s not theoretical — it’s already deployed, already working.
Extreme weather prediction
As Yale Climate Connections reported, AI is dramatically improving predictions of extreme weather events. Better forecasting directly protects people from climate impacts — you can’t evacuate if you don’t know the storm is coming.
Green hydrogen production
Protium Green Solutions raised £31 million to use AI for optimizing green hydrogen production. AI-driven optimization of electrolysis processes could make renewable hydrogen economically viable at scale.
Grid optimization
Power grids are incredibly complex. AI helps balance supply and demand, integrate intermittent renewables, and reduce waste. These aren’t glamorous applications, but they compound.
The real question
The debate shouldn’t be “is AI bad for climate?” It should be “how do we make sure AI gets applied to climate solutions?”
The IEA is clear: AI’s net impact depends on how it’s deployed. Right now, there’s no guarantee the most beneficial applications will scale. But the potential is real.
As one researcher put it: “For people who believe that the use of AI will be a major problem for the climate and think we should avoid it, we’re offering a different perspective.”
The bottom line
AI consumes energy. AI can also reduce energy consumption elsewhere — in grids, in buildings, in transportation, in manufacturing.
The math isn’t “AI = bad for climate.” The math is “AI could be net positive if we’re intentional about it.”
That’s a story worth telling.