What might a future grocery store look like?

A while back some were speculating that Amazon was going to be introducing same-day delivery for most of its products. Amazon later denied this, with the CFO suggesting that this couldn't be done economically at a broad scale. Yet at the same time if think about it there are a lot of warehouse-like buildings for food kicking around - they're called grocery stores. What could you do if you tried to apply lots of technology to them? Would consumers like the results? (So far the attempts have been fairly unsuccessful, but I wonder if this might be due to inadequate levels of automation). Anyways, here follows some speculation as to what the future might look like... (and also outlines perhaps a bit of why I'm a bit confused as to what the workforce and the economy will look like in the more distant future).

The existing way of grocery shopping in Canada and the US should leaving environmentalists sobbing. Here for example is Energy Star's description of grocery stores:

On average, supermarkets in the United States use around 50 kilowatt-hours (kWh) of electricity and 50 cubic feet of natural gas per square foot per year — an average annual energy cost of more than $4 per square foot. For an average-size (50,000 square foot) store, this equates to more than $200,000 annually in energy costs and results in 1,900 tons of CO2 being emitted into the atmosphere — equivalent to the emissions from 360 vehicles in one year! Refrigeration and lighting account for over 50 percent of total energy use in the average supermarket, making these systems the best places to start looking for energy efficiency opportunities. Especially since the profit margins of supermarkets are so thin, on the order of 1 percent, ENERGY STAR estimates that one dollar in energy savings is equivalent to increasing sales by $59!

A big problem: lots of open fridges and freezers, trying to simultaneously keep adjacent areas at different temperatures without barriers in the way that would obstruct humans from gaining access. Compare to FreshDirect, an existing online grocery store:

Unlike a neighborhood grocer, FreshDirect’s warehouse has seven different climates to keep various products at optimal temperatures, helping the company purvey top-quality food.

One of the big obstacles to implementing electronic grocery shopping is that consumers seem to want to touch items in an attempt to search for the best ones, yet I wonder if I the potential freshness advantages of electronic grocery shopping might be enough to overturn this consumer preference. From that same article comparing traditional grocery stores to online ones:

inventory turns are faster for online grocers — it takes the average grocery store 28 days to turn over everything in stock, while SPUD’s turns in nine days, with items like produce and milk turning in 24 hours.

Right now online grocers seem to be more expensive than their non-electric counterparts, but I wonder if that might change in the not too distant future. There's always the possibility of substantial energy savings (and better quality products) from being able to store each in optimal environment. But that's not the only thing. What happens if you introduce robot order pickers - basically a souped-up version of Amazon's Kiva warehouse robots? Here are a few more things that might result:

• there'd be no need to individually label items for sale or the hassle of updating labels as prices change from week to week
• there's be no need to light the facility to the same degree
• density can be increased
• warehouse robots (already) can operate faster than humans
• 24/7 operation wouldn't be a problem
• no need to wait in a checkout line
• can pack the warehouse in the most efficient fashion to retrieve orders - taking into account frequency of purchase for example rather than stocking all the different kinds of, e.g., laundry detergent in the same area.
• few employee related issues (wages, wage negotiations, vacations, absenteeism, ...)

To expand on this a bit more, what happens if you were to add robotic delivery vehicles similar to Google's self-driving car? If there are no humans in the vehicle there wouldn't be the same need to safety devices - throwing out a load of groceries ever now and then probably makes more sense. Including just the mandatory safety devices on cars added about 86 pounds per vehicle between 1968 and 2001 and average car weight is up 800 pounds since the 70s oil shock even though the now-common SUVs and pickups are excluded from that average. That's a substantially lighter vehicle - don't forgot also the weight of a human driver. You'd also no longer need to design the vehicle to meet human needs - should it be bigger? smaller? (I'm guessing a limited-speed, light-weight vehicle would be easier to get government approval of than running full-size vehicles). You could eliminate most of the parking lot for most grocery stores and, as the vehicles would be operating within a short range of home base electric vehicles might be more feasible. Of course there are a lot of issues still be resolved to get a system like this working.

A lot of the challenges to do this seem to be related to robots and robots seem to be getting close to the stage were a lot of this is feasible. I've mentioned warehouse robots before. They already seem to be going in at the distribution level:

distribution, where robots that zoom at the speed of the world’s fastest sprinters can store, retrieve and pack goods for shipment far more efficiently than people. Robots could soon replace workers at companies like C & S Wholesale Grocers, the nation’s largest grocery distributor, which has already deployed robot technology. ... And at Earthbound Farms in California, four newly installed robot arms with customized suction cups swiftly place clamshell containers of organic lettuce into shipping boxes. The robots move far faster than the people they replaced. Each robot replaces two to five workers at Earthbound
(Source: NYT)

Lately robots are also being deployed in places like vineyards - handling pruning as well as picking grapes - showing you some of the capabilities of current technology.

Also interesting is a machine that I'll dub the burger bot. The manufacturers claim a payback period of less than one year for a machine that in a manner more consistent, more sanitary, and more fresh (tomatoes, pickles, onions, etc. sliced at the time of order rather than during prep) can whip up about 360 burgers per hour. (HT: econfuture)

What impact would this likely have on the economy if jobs disappear in quite a few sectors? How do you replace those, particularly given how automation seems to be taking over some of the knowledge sector of the economy as well? How do you avoid increased economic inequality? (A lot of this stuff is likely to prove very capital intensive). The warehouse robots are already out there or fairly feasible to extend. Automated vehicles aren't quite there yet - how long might it for something like a robotic delivery vehicle to be economical (outside certain industries)?

(Some ways in which different approaches are currently being tried: scan the code of an item on a billboard and get the item delivered and oversized vending machines)