Here's a few considerations that you probably haven't...

Here's a few considerations that you probably haven't recognised. (Note that I'm not disagreeing with you, per se - merely noting a few factors that have to be considered when deciding on the size of a unit.)

There are three attributes that (to a first approximation) completely define a battery unit: energy; power; and efficiency. Efficiency is a measure of how much energy you get out, compared to how much you put in. Lithium typically gets 95% plus. ZnBr gets, from what I've seen, somewhere around 80%. That's pretty much fixed; scaling will not alter it in any meaningful way.

Energy is your kWh (or MWh, or GWh, or joules - all of those units are essentially the same, with only a scaling factor needed to convert from one to another.) The energy you can store in a battery is a function of how much electrolyte you have: the more electrolyte you have, the more energy the battery can store.

Power is your kW (or MW, or GW - again, just a scaling factor is needed to convert from one to another.) This is a function of the battery chemistry, how many electrodes you have, and the design of those electrodes. A lithium cell, for example, has a nominal voltage of between 2.3 to 3.8 volts (depending on the exact chemistry of the cell). Power is equal to voltage times current; the more current you can put out, the more power you get out. But the current is going to be limited by the electrodes - how quickly the chemical reaction that produces the current can proceed. Alternatively, you can 'stack' the electrodes - have multiple pairs (one pair has one negative and one positive electrode) - to scale the current or the voltage, depending on how you stack the electrodes.

So put in twice as many electrodes, and you can get twice the power, right? Well... sort of. The more electrodes you have, and the more densely you pack them, the more engineering problems you have to solve. This is one of the reasons why the Redflow battery has a maintenance cycle every three days: over time, as the battery charges, you get stray bits of zinc metal building up in thin tendrils, instead of forming a nice even plate on the electrode. If those tendrils - dendrites, as they're called - manage to short out electrodes, you have a problem (which will last until you discharge the battery and dissolve the dendrites - if you can). If they manage to puncture some critical membrane within the cell, you have a permanent problem.

If you simply scale up the size of the tank and the existing electrodes, you can store more energy. But you will have the same charging and discharging power that you had originally. To get more power into (and out of) the system, you need more electrodes. That adds manufacturing complexity and cost.

The question then becomes: is the added manufacturing complexity and cost greater than, or less than, the value that the larger sized battery brings? I'm not going to even try to answer that question, because there's so much I don't know about the engineering of these cells, and hence the problems that such a scale up would bring.

In short: I'm not saying that you are wrong. I am only saying that the question you pose is a much more complicated one than it appears on the surface. Given where we are, would it be worth Redflow's while to design a new system that holds more energy and is capable of more power, and to then convert (or build out) the factory to build that new system? That's a pretty significant investment, and it would render a fair chunk (25%? 50%? I don't know) of the existing investment useless. That raises the bar required for the return on investment to make such a decision worthwhile.

My suspicion is that, for better or worse, Redflow is close to locked in to the 10 kWh/3 kW battery design they have - at least until they have a significant positive cashflow. It's not impossible to change, but it would require a significant level of trust from shareholders to fund that change. So whether or not a bigger unit would be better on the market is a moot point - Redflow isn't likely to make such a shift in any way beyond what they're currently doing in bundling units together.