AI and energy use? Not an energy problem — it's a hype problem.

Let’s be clear: blaming AI for increased electricity demand is like blaming your toaster for climate change. A kilowatt-hour is a kilowatt-hour, whether it powers ChatGPT or a hospital.

Yes, AI is driving up electricity demand. But so do electric vehicles. So do growing economies. Different forecasts put AI data center energy consumption at 945-1500 TWh by 2030. During the same period, Africa's electricity demand is expected to grow by 600-700 TWh. So, Africa alone will likely consume the same volume of energy as all data centers in the world.

And it's more likely that African demand and supply will hit those targets, while with AI:

🤖It is not clear whether there will be so many use cases for it (there are not many now),

💰Whether AI companies will make any money (they are not making any now),

⚡️And whether data centers will need that much energy (Deepseek case shows that maybe they won’t).

The real issue isn’t who uses the energy. It’s how we build the supply.

If we’re serious about scaling energy to meet global demand, we know what works:

✅ Solar

✅ Wind

✅ Batteries

✅ Maybe some natural gas for balance

And what doesn’t?

❌ SMRs (still not real)

❌ Hydrogen as fuel (looks good on slides, not in practice)

❌ Concentrated solar (20 years late and still expensive)

Nuclear? Sure — if someone, anyone, figures out how to build it on time and on budget.

So let’s stop freaking out about AI energy use. We don’t have an AI problem. We have a supply problem. And we definitely have a hype cycle problem.