It's 5% of power plants. These plants are massive. The paper does state (and shows data for the top 10) indicating that the largest emitters also happen to be

It's 5% of power plants. These plants are massive. The paper does state (and shows data for the top 10) indicating that the largest emitters also happen to be generally more inefficient (in terms of CO2 per power) than their national average (for fossil fuel plans). EDIT: I just wanted to draw out the implications. Based on paper figures, we can ballpark then that those 5% of plants produce on the order of 30-40% of fossil fuel production energy. Shutting down those 5% of plants would be relatively efficient in terms of % CO2 reduction vs energy production size needed. That being said, since the goal long-term goal is basically 0 fossil fuels before 2050, what would be important to have policies that ensure that replacement capacity is meant with major decrease in emissions. Like, we could for example get a ~33% reduction in emissions by replacing some of these plants 1 for 1 with more efficient gas plants. But ideally we want even more draw down - that means having policies in place to ensure a significant amount of capacity is replaced with renewable sources. That's really the challenge (and has been the challenge the entire time).