Meta revealed Prometheus, a gigawatt-scale computing cluster in Ohio expanding toward one million GPUs. The facility trains frontier AI models using custom liquid-cooled racks, dense Ethernet networks, and spare processors that reduced interruptions by 99 percent.