Forecasting Gets Complicated

Forecasting Gets Complicated

Not a valid attachment ID.

DeepSeek's  latest claim that  artificial intelligence can make  energy more efficient is just the latest reason  why the energy industry  needs to rethink all its assumptions about future demand.  Through nearly 200 charts, clean energy expert Nat Bullard lays out a  host of other surprising factors  influencing decarbonization  projections in his annual presentation  published online.
Bullard, an advisor and co-founder of energy information platform Halcyon,  says new information about  rapidly developing fields like AI and diabetes  drugs that could reduce food  consumption adds layers of uncertainty to the  outlook.
The world is  already on a complicated path to net zero, according to Bullard, who previously held roles at BloombergNEF — including  head of content —  through 2023.
For example, 2024 may have been a record year for  renewables, but  demand for coal and gas  is also  up.
"We're burning more fossil  fuels and emitting more CO2 than  ever before," he says. "At the same  time, we are  producing and installing more  wind turbines, solar panels and batteries than ever before.  We live in complex  times."
Here are some other highlights.
Data  is not everything.
Global electricity demand growth is  expected to accelerate  in the  coming years, but only a fraction of that  demand is  expected to  be devoted to AI. In a scenario  designed to reflect the current  political landscape, the International Energy Agency showed in an October report that data centers  will play a limited role in  the increase in electricity demand between 2023 and  2030.
Not a valid attachment ID.
AI has greater potential for efficiency
Data centers in the  United States and Europe  consume electricity and  raise concerns about their climate footprint.  In Virginia,  home to more than 300 centers, the industry  will account for about a quarter of electricity consumption in 2023, while in Ireland the sector  will use more  electronics than urban households,  Bullard says.
But all that could change. Chinese AI startup DeepSeek says its  basic open-source models  require only a fraction of the training hours required by equivalent  U.S. versions.
Bullard says the development likely means a wave of infrastructure  assets like power plants and data centers in the  U.S. could peak and then  recede.

[attachment deleted by admin]