The problem is the super low thermal runaway onset temperature of around 160degC or so in LCO/NCA chemistries. Stopping on-going thermal runaway might be even more difficult, but at very least it means cooling every internal spot of every cell cell below 160degC. Given high heat output, and limited thermal conductivity of cell materials, this is impossible by regulating cell envelope temperature by boiling - especially since boiling causes voids (gaseous water conducts heat more poorly than liquid water).
This is why it's almost impossible to stop thermal runaway once it's started. If the cells are small, with cooling channels between each cell, and you pump 20degC water at massive flow rate evenly through the whole thing, it is remotely possible. If you expect the evaporating water to do the trick - no dice. So basically with the pool boiling solution, you let it burn, and it will produce smoke, and then you just hope the pool prevents it from propagating within pack.
Sure, given enough material thickness and enough water, then you can just the thermal runaway go to completion, but the fact they don't do this suggests it requires too much materials and weight.
The total energy released during thermal runaway of a fully charged cell is approximately twice the stored electrical energy (e.g., see
https://www.fire.tc.faa.gov/pdf/TC-TN16-22.pdf ). Thus a 50-gram 15860 cell, storing ~30Wh = 108 000J of failure energy, would require 108 000J/ (2257 J/g) = 47 grams of water.
This is an interesting finding, so if you have 1 ton of batteries, another ton of water would be required to fully evaporate during thermal runaway. In reality, you would need significantly more, say double that, so that only excess is evaporated, and all cells keep submerged. While at it, you would want thick enough metal walls so that blowtorch effect in localized hotspots (remember, the cells have internal oxidizer, so this can and will happen underwater) can't make a hole causing water to drain.
The net effect would be destruction of the gravimetric and volumetric energy density. Battery storage is already expensive and large; engineering it completely fire safe would likely kill the whole business model. So instead we seem to accept the risk that these things sometimes burn, producing a lot of toxic smoke, and fire brigade trying to control it by pumping a lot of water through the failed unit and protecting the neighbor units.