Conventional wisdom is that grid storage product was offered because battery production was outstripping what they could ship in cars at the time.

Remember that heavy industry isn't like software, you don't get to Just Decide to match the product to the market and switch when it makes sense. You make a bet years in advance on resource contracts and factory capacity, and if the needles don't match you need to find some way to generate revenue from what you actually have, not what you want to produce.

But yes: LiFePO4 has a high internal resistance (which is bad for applications with high current loads) and somewhat poor energy density (bad for mobile applications), but is significantly cheaper at scale due to the elimination of Cobalt and has an absolutely massive cell durability advantage (GREAT for grid storage).

But you won't see phosphate batteries in anything but budget/commuter EVs. They aren't something Tesla is going to bet on for their main production choices, but absolutely something you'd expect to see them optimize as they scale.