We might be able to do better than that sooner using Stirling engines.  The proponents are claiming about $2/watt at today's prices, and 30% thermal efficiency in 25 kWe units.  If you take the ~58 kW of waste heat and feed it to an absorption chiller, you get about 30 kW of cooling (which would take on the order of 10 kW of electricity to generate).  Bingo, you've just reduced electrical demand by 35 kW as long as the sun is shining.  Oh, you've also satisfied heating and DHW demand too.
How much efficiency do the stirling engine loose by rejecting heat on a higher temperature on the cold side?
Without knowing what that temperature is (ammonia absorption can run at input temperatures around 100°C) I have no way to tell.  The pictures of dishes I've seen have no obvious radiators, so T(lo) might already be rather high.
Basically Stirling engines operate on an approximation of the Carnot cycle. The thermal efficiency is then dictated by one minus the ration of the low to high temperature. Typically the thermal efficiency may be 40% ... so 60% of the thermal energy is rejected.
Stirling engines have some 'issues' but it is a 'doable' technology, especially for stationary applications. I worked on this technology in the 70s. We coupled it to a heat pump and the waste heat from the Stirling cycle improved the effective heating COP. The good thing is that it is a closed cycle. The bad thing is that its power to weight ratio is not too good.
In the district heating and cooling schemes that I know of, chilled & heated water can only be pumped a few miles (like 2 or so).

Solar power is a low value use of land, so they are likely to be located "in the middle of nowhere".  The demand for chilled water will likely be limited to the support staff.

Perhaps the bed of the Salt River in Phoenix could be covered with raised mirrors, but aesthetics and glare into offices would prevent this.  Not at the airport, glare into pilots eyes.

OTOH, storing hot oil till the sun sets and temperatures decline may well be worth doing.  Increase the natural gas supplemental heating (only 2% on NV 1, vs. 25% before) to offset any cooling but get greater thermodynamic efficiency.

You misunderstand.  Solar power may be a low-valued use, but roofs are low-value territory.  So's the air over parking lots.

Solar power is more valuable at the point of use than far away (avoids all the expense and losses of the transmission system); if the roof of the 7-11 sports a mirror array which powers the whole thing and the waste heat air-conditions it too, you've effectively gone off-grid while the sun is shining.  Have another unit or two to shade the parking lot, and you could supply excess power to charge the cars coming and going.

I see difficulties in the economics of scaling down to "7-11 size" and anything less than "Regional Shopping Mall" size.

Rooftops are NOT so cheap.  Additional structure to support solar load (weight).  Elevating them above parking lot has some additional costs as well (like elevating above river bed).

Also, great care must be taken to not break the watertightness of a roof, and any reroofing has to be done under/around solar collectors and their supports.

Still, I good see a 200 acre array mounted over some of the parking lots in Phoenix.  Concrete around each post to prevent accident knocking down pole and spilling 750 C oil everywhere !  Capital cost of lost parking spaces due to poles & guards is a high value loss and probably kills idea unless parking becomes "surplus".

Perhaps 0.5% of Phoenix load that way.

The tech to do a 7-11 size unit is already in the California desert; it's the 37-foot diameter solar Stirling dishes.  They require a post to mount them, but you could incorporate that into a corner of the building and make it do double duty.

It might be cheaper to do it with heliostats feeding a fixed collector.  If so, the result would not look all that much different from a filling station with roofed pump islands.

We shouldn't cover river beds with anything, let alone solar power plants. Weather is becoming undependable. I mean, even more undependable than it already is.