Casey Crownhart writes:
In the age of artificial intelligence, the biggest obstacle to progress is not money. but energy. This should be especially alarming here in the US, where huge data centers are waiting to come online and the country doesn't look like it can build a stable power supply or the infrastructure needed to service them.
It wasn't always like this. For about a decade before 2020, data centers were able to accommodate increased demand. with increased efficiency. However, the demand for electricity in the United States is now growing, with billions of requests for popular artificial intelligence models being generated every day, and efficiency gains have not kept pace. With too little new capacity coming online, the strain is starting to show: Power bills are rising for people living in places where data centers are putting increasing strain on the grid.
If we want AI to have the ability to deliver on big promises without causing sky-high energy prices for the rest of us, the US needs to learn some lessons from the rest of the world about energy abundance. Just look at China.
China installed 429 GW of new electricity generation capacity in 2024, more than six times the net capacity added by the US during this period.
China still produces most of its electricity from coal, but its share of the mix is ​​declining. Rather, the country is focused on installing solar, wind, nuclear and gas power at a record pace.
Meanwhile, the US is focused on reviving its weak coal industry. Coal-fired power plants are polluting and, most importantly, expensive to operate. Aging power plants in the US are also less reliable than before, producing electricity only 42% of the time, compared with a capacity factor of 61% in 2014.
This is not a good situation. And unless the United States changes something, we risk becoming consumers rather than innovators in both energy and artificial intelligence technologies. Already, China earns more from renewable energy exports than the United States from export of oil and gas.
