It appears that AI-ready datacenters will be growing at a compound annual growth rate (CAGR) of 33 percent per year between today and 2030. That's a heck of a lot more datacenters, which, in turn, means a hell of a lot more power consumption. Why? Well, AI sucks so much power down because training and operating modern models, especially generative AI, requires extremely intensive computational resources and vast amounts of data processed in parallel across large clusters of high-performance GPUs and TPUs.
Hyperscale operators would like high bandwidth routes, with low latency (accomplished with the most direct routes possible between locations with limited switching) and very high uptime (accomplished with good physical and cyber security and route redundancy). With the birth of AI, the need for continued improvement in all these categories has accelerated.
"What I've wanted to explore since taking office is a reset of our relationship with PG&E and what it would look like for us to partner more deeply and have real guarantees from PG&E related to performance."