Demand for artificial intelligence and machine learning is booming—and so is AI’s energy demand. AI requires massive amounts of computing power, which means huge data centers need to be built to enable AI’s widespread adoption.

AI’s energy demand is expected to grow faster than new generation comes on line. Solving that will be hard enough, but differences between the tech industry and the electric power industry make it even more difficult in some ways. The tech industry evolved with a move-fast-and-break-things ethos, while the electric power industry is one of the most heavily regulated and conservative industries in the country. Developing a new data center takes 12, 18, maybe 24 months, whereas adding new generation and transmission takes years.

Tech and power are odd bedfellows, and while they are getting to know each other better, they still have plenty more to learn.

Ben Lee, a computer scientist at the University of Pennsylvania, is one of the handful of people in the country who works in the space between the electric power industry and big tech. He says there are opportunities for the two to work more productively together. There are also some things that likely will not make much of a difference to accommodating AI demand in the near future, such as time-of-use rates.