Maybe fusion will be an alternative someday but for now it's just a fantasy. We need to act based on what's proven to work today.
Fission as a solution is something that is popular on social media, for reasons that are utterly mystifying to me. The arguments are invariably a few words that reach sweeping conclusions with no actual data backing it up, and lots of data contradicting it that the individual appears oblivious to.
Current supply of storage matches current demand. Supply is low only because demand is low. However, as demand increases, supply will continue to match demand, and moreover the price will actually decrease because of the fact that the learning curve is a function of production volume.
This has been a steady empirical phenomenon for 30+ years, and it's predicted by basic economics principles. It's not going to change now!
This is true for all battery types, but especially for sodium ion and iron air, which are constituted of abundant materials. Sodium ion in particular has very similar behavior and cost to lithium ion.
This confusion you're having is you seem to be conflating manufactured goods (like batteries) with scarce goods like land or services, whereby there's a fixed supply that can't be increased and where Wright's Law doesn't apply. This is not correct.
Storage is more like televisions or light bulbs, where you can basically make as much of it as you want, and the price will keep declining as more is made. And supply will always be there for demand, whatever the level of demand happens to be (in this case, a lot).
Converting every passenger car and light truck in the US to a BEV would involve enough batteries to store something like two days of the average grid output, which is more than would be needed for a cost optimal wind/solar/battery/hydrogen system for a 100% renewable grid.
Assuming the power stored in these vehicles can be reclaimed by the grid anytime they want?
It's an argument I like to use. When someone claims "we can't use X because of reason Y, we have to do Z instead" I look to see if Z also is hit by objection Y.
Another example of this is "renewables require too much material that we can't recycle", at which point I observe that the quantity of materials produced by society as a whole greatly exceeds what renewables would involve, even if the society is powered by nuclear. The US produces 600 megatons of construction and demolition waste a year, for example. Renewable waste would just be a minor blip on this existing waste stream. So, either recycling this waste isn't actually needed, or a putative sustainable nuclear-powered society has discovered how to recycle it, so just toss the renewable waste (which is almost entirely things like steel, aluminum, and glass) into that same recycling infrastructure.
Steel mills run when power is cheap. They historically have run at night (and only minimum power during the day) because cheap power is available at night. Of course there are lots of different steel mills, older ones can't shut down - but modern ones don't run 24x7, they run when power is cheap. Even the old 24x7 ones did their yearly maintenance in December - when power demand is highest (Christmas lights).
Wind and solar are easially predicted a few days in advance with high accuracy, and thus the mills change their shifts/output to follow the cheap power. If it is cloudy/no wind they will send their employees home (with pay) or do maintenance for that week while waiting on more cheaper energy. It takes a tremendous amount of energy to melt iron and so they manage this carefully because it makes them money. They can't deal with months of no production, but they can manage a week here and there.
If this isn't about ceasing carbon emissions then none of this is necessary. Fire up the coal plants!
(1) the emissions of a 98% renewable + 2% natural gas grid that comes online in 6 years, assuming fossil fuels for t between [t, t+6 years].
(2) the emissions of a 100% fission grid that comes online in 16 years, assuming fossil fuels for t between [t, t+16 years].
If you insist on ignoring the temporal nature of cumulative emissions, then sure, you can arrive at a convenient but false conclusion. But any honest analysis will consider the emissions in that [t+6 year, t+16 year] interval.
(... it would also consider things like social licensing risks leading to early plant closures like what's happening in Germany, or the fact that nuclear will likely be paired with natural gas too because demand itself is variable, and overbuilding nuclear is expensive.)
Start both with the same (current) % for renewables and (1) have some realistic ramp-up of renewables to reach 98%, and (2) keep the renewables more modestly rising in the fission version, while fading-out fossils in favor of fission
You should also account the carbon foodprint of grid-level energy storage (yes, it will be needed, even with the natural gas plans), vs the foodprint for fission plants (undoubtedly quite bad).
Renewables are by their nature much more distributed in space, which makes them much harder to enclose and control in the way required to reproduce the current structure, especially as they are mainly being built by challengers who aren't really interested of forming monopolies with the fossil industry.
Most states currently only care about installing solar and wind -- not storage -- because they are still majority fossil fuels, and at the current moment it makes no sense to install storage if you still have fossil fuel to dislodge. The only exception is really California, who are installing storage, but their bottleneck is not the market's ability to deliver enough supply.
There are also many storage options beyond lithium ion if you only spent a moment to look.