Well, the mismatch depends on where you are. Where I live, with a SW facing rooftop, peak demand is mid-afternoon mid-summer when insolation maximally heats my house and the AC runs all the time. Roof panels could conceivably support the AC load while at the same time form a third layer of matter with a ventilated air gap right above my actual roof. This would suck ambient air at around (say) 30 to 40 C in and pull it up in between the solar cells themselves (which would heat up as well as generate from direct sun) and radiate LWIR both ways. I estimate that this extra layer would reduce the difference between my physical roof shingles and ambient air by 2/3. Since I then have R-40 or thereabouts in the attice ceiling, a finished attic we don't use much in the summer, and an R-20 insulated floor between the attic and the house, I expect that it would reduce the load on the attic AC pretty substantially while generating enough electricity to easily keep the house cool to even cold all afternoon (plus running any appliances etc).
Storage would be nice, but in NC I'd be able to just pump surplus back into the grid and make the meter run backwards, "storing" the energy by reselling it to my neighbors and getting it back for free later in the evening. Which brings us back to TFA -- I personally think that communities should force power utilities to permit this resale, subject to a) validation of the equipment, mandatory inspection, etc. Private equipment should not be permitted to mess up the public grid; b) a clear statement of liability and requirement of insurance (the power company isn't responsible if your equipment fails and messes up either the local grid or the local grid connection via (say) lightning messes up your local equipment); and c) personally I think a reasonable surcharge/tax for the use of the public uitilities lines and distribution network to resell your power to your neighbors is pretty reasonable.
With that said, there is a lot of room for negotiation as to just how such a surcharge should be billed. Arizona's $0.70/month/KW is perhaps the dumbest and least equitable as it gets billed even if you don't push back any power, or if you use most of the power yourself and push back comparatively little (on average). A much fairer scheme might charge you rate A for energy delivered through your meter, and pay you rate B when you deliver energy back through your meter. In the specific case of Arizona, where they charge a comparatively little $0.11/kwh, they might consider paying you back $0.105/kwh, or $0.10/kwh. That way in a dark and stormy month when you generate little power and use all you generate yourself, you don't get dinged $7.00 just for being connected, but in a bright and sunny month when you sell back 20 kwh/day but also use 20 kwh/night, you basically pay $3.00 to $6.00 for the electricity you the otherwise "break even" month. I'd advocate the 5% rate, of course, and the power company would prefer the 10% rate, but the point is that this formula is a lot fairer than a flat rate.
It also makes sense in terms of amortization. If you buy yourself a good sized battery pack to run your house at night, it is going to cost you thousands of dollars. Call it $1000. Amortize it over 10 years, you have to pay back $100/year. Borrow the money, pay it back like a mortgage, you'll end up paying somewhere in the ballpark of $150 to $200/year. Hell, call it $120 -- $10/month. So even if it costs you only $1000 for a battery pack capable of running your house off of the grid all of the time, even if the lifetime of that battery is at least ten years (neither likely to be true, so far), even if the battery pack requires zero maintenance etc, the power company fee of $7/month to "store" your energy is cheaper to you than the cost of the money required to store it yourself.
This could change if somebody ever invents a cost-effective no memory high energy density battery with a lifetime of 10,000 full cycles (call it 30 years of daily deep cycle). And there are a lot of people working on this. But if/when it becomes cost effective at the consumer level, it will already have been or become more cost-effective at the utility level by around a factor of two, and utilities will simply start adding large battery packs to their burgeoning panel farms and start significantly reducing fuel based generation of the so-called "base load".
I see a real struggle between public utilities that cannot remain in business and maintain their lines and so on without a substantial and ultimately irreducible demand and rooftop solar with batteries -- going almost completely off of the grid -- that threaten to reduce demand below that irreducible level -- in around 10 years, with the next factor of two reduction in the cost per watt of consumer installed solar. Or, of course, Lockheed-Martin may have commercial fusion by then and everything will change and PV solar will pretty much evaporate, cut off at the legs by energy provided by what amounts to "free fuel", but without all of the hassle of intermittent supply or storage.
What I don't see is any future where coal continues to be a dominant source of electrical power for over 30 more years no matter what we do now. We could completely ignore CO_2 and build all the coal plants the world needs to provide the energy required to lift the poorest 1/3 of the world's population out of 17th or 18th century poverty, just as China and India are doing and will continue to do and be damned to AGW, and it will still become cost-ineffective to build more of those plants, without the slightest tax incentives or penalties or encouragement, within the next 15 years even without fusion, within the next 10 years with fusion. Indeed if fusion is realized, it won't take 30 years -- more like 15 to 20 and coal burning plants will be replacing their coal furnaces with fusion cores in place. RCP 6.5 is probably already overly pessimistic, and with a total climate sensitivity of around 1.8 C (my own best fit to the data) we will warm roughly one more whole degree C, maybe, before CO_2 forcing stabilizes.
What the climate will do then is still anybody's guess, because we cannot predict climate and do not understand climate and the climate is perfectly capable of starting an ice age with CO_2 several (as many as 10 to 20) times as high as it currently is (it has done so in the past, in the Ordovician-Silurian transition). The system almost certainly has chaotic nonlinear negative feedbacks as extreme as a warming induced by semi-permanent shifting of the Gulf Stream 500 miles to the south, putting the entire North Atlantic and Arctic into the icebox (while heating the tropics and maybe even overall warming the globe!) and triggering the next glacial episode. Or not. Unpredictable is unpredictable.