I'm not disputing anything, but would it have killed arstechnica to include some primary source links to back up their assertions?
This is the report: https://www.eia.gov/electricity/monthly/current_month/februa...
It has full-year data for 2025 in table ES1.B.