There is a reason we don't often use electrolysis for hydrogen production -- it scales poorly and is expensive. Scaling that up several orders of magnitude is not a realistic option.
This is true of most electrolytic industrial chemistry, for example the hydroxides used in many carbon capture schemes. What works for the small-scale needs of today for industrial chemistry are not intrinsically scalable. In software terms, industrial chemistry is not "embarrassingly parallel" for reasons that are not immediately obvious to the layman, so scaling is hard.
That’s simply not true. Electrolysis scales JUST fine (the process for making aluminum is much the same although occurs at often higher temperature, and we used to use electrolysis for ammonia’s hydrogen). We just don’t do it because steam reforming is usually cheaper. Simple as that.
Hydrogen use 4x the amount of electricity per ton as aluminum, and that assumes you are not recycling aluminum. It is orders of magnitude less if it is recycled, which most aluminum production is. You'll need to come up with a less specious example than aluminum if you want to be credible.
We also produce significantly more hydrogen than aluminum in any case.
I’m not sure why you consider that “specious.” Aluminum requires less electricity per ton because hydrogen is the least dense element. Per mole, hydrogen uses less electricity than aluminum. :)
And we don’t produce massively more hydrogen than aluminum. Global primary aluminum production is ~64 million tonnes per year. Hydrogen for ammonia production (which is the majority of hydrogen I think) is about 40 million tonnes per year out of about 75 million tonnes total pure hydrogen. Certainly the same order of magnitude.
But again, this is missing the point. Before around 1950 (when steam reforming became widely available), hydrogen for ammonia fertilizer was produced commonly by electrolysis, often using cheap hydroelectric power. It was done at massive scale and only stopped because steam reforming happened to be cheaper.
And what is your point, here, by saying we make more aluminum than hydrogen? “We can’t credibly do things that are any different, even if just different by degree, than what we happen to do right now.”?
This is true of most electrolytic industrial chemistry, for example the hydroxides used in many carbon capture schemes. What works for the small-scale needs of today for industrial chemistry are not intrinsically scalable. In software terms, industrial chemistry is not "embarrassingly parallel" for reasons that are not immediately obvious to the layman, so scaling is hard.