So.... it's a bit of an apples/oranges comparison to try to compare the economics of nuclear vs storage since a power plant is rated in kW and a battery is usually more kWh than kW.
That's because a capacity rating is kind of irrelevant for a technology that just produces power, and continues to do so, day in and day out, for 60 years or so (ok, ok, there will be planned outages and the occasional hiccup, but these things happen one at a time).
If you want to replace the grid with something different which includes storage, then you need to make absolutely sure that you don't run out of capacity. A nation-wide blackout once per decade would be much too frequent.
But... if you assume you want a 1GW facility ~50% of the day you can pay $7B for a 1GW nuclear power plant or $3B for 12GWh of Tesla Power Packs. And... that's using the 2015 price of $250/kWh. When we reach the $100kWh target that is widely viewed as an inflection point vs ICE you can effectively replace a nuclear plant as daily base load for ~$1.2B. Even storing 48 hours of energy would sill be less than the cost of a nuclear plant at ~$5B.
There is no such thing as "daily base load". You can't afford to let the grid fail.
So you're willing to pay the same amount for 24 hours of storage as you would pay for a 1 GW nuclear power plant. But so far you have bought zero actual generation capacity - only a pathetically small amount of storage! A week is not enough, you have only a single day, and you still have not deployed a single PV panel or wind turbine. I think this is an awfully bad deal.
This isn't going to happen overnight... perhaps in 10 years there will be a day that renewables generate enough to carry the country for a day. Then a few years later for a week. A few years after that for a month.... then for most of the year.
Repeating it will not make it true.
Storage will come on line as it's needed... because there will be an economic incentive to store energy when it's cheap and sell when it's expensive.
No, it will not, because the capacity factor of the storage system will be extremely low, so prices would have to be correspondingly extremely high when it's operating for it to break even. A system whose purpose is to provide reliability must be fully charged as much of the time as possible. A system whose purpose is to earn money by buying low and selling high must be producing as often as possible. Those are mutually exclusive. The only way to make it work is to recognize that controllability and predictability are valuable assets and require intermittent generators to buy the missing predictability from the storage system.
What's going to drive nuclear growth? Nuclear is more expensive than most other alternatives...
No, nuclear is very much less expensive, very much less complicated and infinitely more proven than any other way to get the
required reliability in sufficient quantity - except for fossil power. You can't get away with only one day of storage, and nobody in their right mind would buy a whole over-budget nuke's worth of battery backup to get a single day of reliability when they can buy said nuke instead and get 60 years of reliability, plus a gigawatt of 0.9 capacity factor electricity to boot.
As the situation is in the US right now, they would probably buy open cycle gas turbines for $0.5/W and run them as required instead. In Germany, where natural gas is expensive, they are building brand new coal-fired power plants. This has to stop, and batteries will not help anywhere near enough.
The lack of predictability and controllability is what's causing the wild electricity price swings and decreasing grid reliability currently seen in Europe. This can be fixed by making reliability a tradable commodity, and requiring undispatchable generators to buy the predictability they lack from someone else. The lack of a carbon tax skews the market towards carbon emitting technologies. So first make costs reflect what's actually needed to run society reliably and without CO2 emissions, then let the market sort it out. Choosing the right values for the parameters would allow us to integrate as much intermittent power as possible, in the most cost-effective way, while ensuring that the grid stays reliable.
By the way, the term "reliability" is confusing. The key concepts are predictability and controllability.