From a stock perspective, I'm long TSLA. The car business is nice, but its just a vehicle for changing the world (and already priced into the stock, really). The real things we have to price is how the technology in a car can apply to all transportation problem/solution sets, how legitimate electric is as a solution for fuel/energy problems (and, long, for space), and also - can TSLA avoid existing regulation barriers and use the US market/economy in time.
The plan is to offer grid stabilization services -- voltage boosting, phase correction, and demand shifting/response -- by aggregating together home batteries, utility-scale batteries, and scheduling EV charging. These are all valuable services for the utility. The revenue from these would be passed on to the end customer, "closing the loop" in our current perverse residential electricity pricing.
What do I mean by perverse pricing? Ideally each customer would be billed based on and tailor demand to the real-time generation+distribution cost at their location on the grid (accounting for their individual line loss and utilization of their distribution and transmission connections), but this is far too complicated and prone to market manipulation. Even TOU pricing has an information lag from actual demand because of changes in demand and the rising percentage of renewable and distributed generation. And net metering is subject to the "solar death spiral" problem, which is quite real (though overhyped by utilities for their own purposes).
One of the big advantages to making their own inverters is that each one acts as a grid monitoring station, relaying voltage drop and phase lag information back to the mother ship. Their EVs can already do this as evidenced by their OTA rollout of charging software that slows charging when overheating wires (voltage sag) or arcing (voltage spikes) are detected.
In the not-too-distant future Tesla will have more raw information on the grid than the operators themselves, though obviously it's even more powerful when combined with their existing data. This combination can be done directly through data sharing or emergently simply by interfacing with the existing pricing system.
None of this is speculative btw -- SolarCity's CEO Lyndon Rive[1] and Tesla's CTO and SolarCity board member J.B. Straubel[2] have talked directly about these plans before.
[1] http://www.ustream.tv/recorded/87401050 (mentions planned "smart inverter" that can monitor grid, returning revenue to customers, and that the goal with storage isn't off-grid systems but grid cooperation)
[2] https://www.youtube.com/watch?v=zWSox7mLbyE (mentions aggregating storage and EV charging over the internet)
Tesla's showrooms are also a great asset in promoting both solar and battery tech for the home, and the car is important in rationalizing the expense of these showrooms (at least for now) as well. Moreover, a very consumer facing brand and product in the car helps tremendously.
Anyhow, not to nitpick. You're right on about the overall strategic rationale and why this makes so much sense as a combination. But it's too far to say the car doesn't really matter. In fact, the car is a very important part of the strategy and especially of the nearer-term tactics.
You will end up paying a larger grid connection fee. If you touch any part of the grid you are going to help pay for it's upkeep. The traditional model is changing but the grid itself still has extensive maintenance. Want to ensure that some power generation plant 2 states over doesn't accidentally dump a metric ton of electricity and fry every substation between your house and Kalamazoo? Or what happens when your town has an earthquake, tornado, hurricane, blizzard that disables your solar generation, you'll be forced to rely on the grid getting back online to help assist the people that eventually will assist your solar. People will stop paying for generation but I don't believe people realize what goes into keeping this kind of infrastructure up and running safely.
That said, the conflicts of interest and pretty clear move to bail out a failing company still bothers me
http://www.pv-tech.org/editors-blog/solarcitys-1gw-buffalo-f...
Over the last year states have started increasing fees to hook solar into the grid and at the same time are reducing net metering rates (payouts for power generation). This has hurt Solar City over the last year and likely would've killed it long-term.
I see Tesla using the Power Wall to make an in-home grid that uses solar + battery as the primary power source, and the grid is the backup power (for charging the battery with off-peak rates) . That way people can avoid paying fees for net generation, and the net metering rates don't matter.