Vgrinshpun, I think you missed my point. I'm not talking about how much profit Tesla may need to price into this. That is the supply side. What I am talking about is the demand side. How big is the market at $800 per kWh? Maybe 1 GWh. How about $400? 10 GWh. $200? 100 GWh. $100? 1000 GWh. I'm just making this up to illustrate a demand curve. So if Tesla wants to address a $4B market $400 is low enough, but if they want to reach a $20B market they've got to press the price down to $200. In the short run, they can go after a small market priced at say $500, but longer term with the gigafactory Tesla will want to sell 15 GWh per year. So the demand question is, how low will they need to take the price? The supply question is whether they can be sufficiently profitable at that price. Tesla has the opportunity to follow an experience curve strategy. As they double their cumulative production they will be able to drive down the cost and price by 10 to 15 percent. As they drive down the price the size of the addressable market grows which sustains the ability to double cumulative production.
This is very much the sort of path that SolarCity is on. They know that as they drive down both their cost per W and the price they offer to customers, they are able to enter new markets. So they are relentless in driving down their installed cost per W, and this is what enables them to keep doubling their customer base every year. It's not just a matter of beating the competition or hitting some predetermined profit metric, it's about constantly expanding the addressable market.
I did get your point. I think that you are seriously underestimating demand - and that was the point I was trying to demonstrate. Stationary storage has many advantages over a gas fired peaking power plant - from the greatly reduced maintenance to a far superior load response time - this could improve stability of the grid, which is s big consideration. So my point is that your assumption that one need to drive cost much lower than break even point ($842/kWh per the EPRI study) might not be correct.
This is similar to a comparison between an ICE and a Tesla EV: once cost parity is achieved, EV inherently wins hands down: "not a fair fight". So explosion of demand does not require cost to be much lower than an ICE. As soon as cost parity is achieved, an EV competes with ICE not by offering lower price, but by offering much superior overal experience.
There are many examples of this. Condider flat screen TVs. As soon as they reached cost parity with tube TVs, the demand went way up: nobody in their mind would buy a tube TV over a flat screen TV given the same cost.
In summary I believe that as soon as stationary storage is available at a cost that is slightly lower than the break even cost, there will be explosion of the demand.
That I believe what JB had in mind. The demand is not characterized by the demand curve - it is more like a step function.