The State of AI: Is the US Falling Behind in the Energy Race?

The explosive growth of AI is placing unprecedented demands on energy infrastructure, and the US may be lagging behind in meeting those needs. In a new series, Casey Crownhart of MIT Technology Review and Pilita Clark of the Financial Times examine the energy implications of the generative AI boom. Crownhart argues that energy, not money, is becoming the primary barrier to AI progress. The US faces challenges in building the necessary power supply and infrastructure to support massive data centers coming online. While data centers previously offset demand with efficiency improvements, this is no longer the case, leading to rising electricity bills. Crownhart suggests the US could learn from China's rapid buildout of renewable energy sources. Clark agrees that data centers should be more flexible in their energy consumption and supports incentives for using backup generators. She notes the uncertainty in forecasting AI's future energy needs and highlights the lack of public data on AI systems' energy consumption. Clark hopes AI itself can accelerate the global energy transition, but expresses concern over the axing of renewable projects in the US. She concludes that China's green energy leadership could give it a significant advantage in the AI race. Crownhart is skeptical about claims that AI will solve climate change, emphasizing the need for measurable progress in transitioning to green energy. She stresses that AI is changing our grid and our world, and we need to understand the consequences.