- Prometheus spans thousands of acres with dozens of buildings and hundreds of thousands of GPUs, and is being used in production while still being built.
- Meta created a new backend aggregation network (an “Ethernet super spine”) to support the cluster.
- Operational software must route around failures and manage complex network and distance effects.
- Hyperion is planned as a five‑gigawatt, Manhattan‑sized data center in Richland Parish, Louisiana.
- Lessons from Prometheus inform Hyperion’s design and operation, including scaled networking and software management.
- Meta is positioning this infrastructure to support training and serving of next‑generation large models.
Useful?