Google DeepMind chief Demis Hassabis says concerns over artificial intelligence’s energy use must be weighed against the long-term benefits the technology could deliver in climate and grid efficiency.
Speaking at the All-In Summit, Hassabis says his team is focused on making models both powerful and efficient because Google must serve billions of users every day at low cost and speed.
He notes that techniques like distillation, where large models train smaller ones to mimic their performance, have already delivered big leaps.
“Over time, if you look at the progress of the last two years, the model efficiencies are like 10x, you know, even 100x better for the same performance.”
But he highlights that efficiency gains have not reduced overall demand for energy.
“Now, the reason that that isn’t reducing demand is because we’re still not got to AGI yet. So also the frontier models, you keep wanting to train and experiment with new ideas at larger and larger scale, whilst at the same time, at the serving side, things are getting more and more efficient.”
Hassabis argues that AI will eventually contribute far more to energy solutions than it consumes.
“From the energy perspective, I think AI systems will give back a lot more to energy and climate change and these kind of things than they take in terms of efficiency of grid systems and electrical systems, material design, new types of properties, new energy sources. I think AI will help with all of that over the next 10 years that will far outweigh the energy that it uses today.”
Looking ahead, Hassabis ties these breakthroughs to a broader vision of scientific progress.
“If we will have AGI in the next 10 years, full AGI, and I think that will usher in a new golden era of science. So a kind of new renaissance. And I think we’ll see the benefits of that right across from energy to human health.”