Intermittent operation blows up CAPEX costs.
There's no heat in the system until you start dumping electricity into the resistive heating of your cells - you won't ever be able to run at scale at room temperature. A gallon of gas per second unit - something nicely sized for intermittent operation at a gas station - would need to vent approximately 120 megawatts of heat.
Current densities are 1-2 OOM lower than water electrolyzer.
Cell CAPEX isn't nearly as catalyst dependent as the daily 'We've invented a replacement for platinum in fuel cells' papers would make you believe. Nafion, balance of hardware, and assembly man-hours are $$$.
I've been where you are (more or less): beware the superficially justifiable assumptions.
We will have to shed heat from the system to the environment, but that can be done passively.
Current densities are likely going to be substantially lower, but our capital costs are also going to be much lower. We don't need to operate at 30 bar for example, nor further compress our product.
Our capital costs aren't magic, they're just good enough in balance with everything else.