The U.S. modular construction industry has struggled for decades without a large-scale institutional buyer. The Department of Defense may be changing that.
Today’s letter was guest written by George Banks, founder of Revel, a retail development and advisory firm specializing in destination retail, placemaking, and mixed-use projects across the United States.
Not far from me, in the small town of McDonough, Georgia, there is a Chick-fil-A that does not allow customers inside. A second-floor kitchen serves food via conveyor belt to cars queued in multiple drive lanes below. The company says it can handle 700 orders an hour.
This is not a fast food story. The McDonough Chick-fil-A is a working prototype of something much larger: a supply chain optimized entirely around the vehicle rather than the human inside it. Every convenience retail format in America, including the gas station, the pharmacy, the dollar store, and the grocery-anchored strip center, was designed around a fundamental assumption: that a person would get in a car, drive to a store, and make a purchase.
Autonomous vehicles will retire that assumption. For owners of convenience retail real estate, the question is really how much time they have to prepare. And what comes next.
Today’s letter covers:
How American convenience retail got built around the last-mile human, and why that model is structurally vulnerable
Why the autonomous vehicle conversation in real estate has been focused on the wrong thing
What the math on throughput and utilization reveals about the scale of the disruption
Which convenience retail formats are most exposed, which may survive, and why
What owners and investors should be doing now
Well, How Did We Get Here?
The history of retailing is often framed as the story of experience: visionaries like Marshall Field and Stanley Marcus created department stores where the act of shopping itself was the appeal, with serendipity and status signaling baked into the design. But Charles Saunders (Piggly Wiggly), the Albrecht brothers, and Sam Walton saw retail differently, as the story of expedience, focused on getting industrialized goods to the masses as cheaply and efficiently as possible. They and their successors have been brutally efficient at removing humans from every step of the supply chain they could.
Containerization, the Motor Carrier Act, NAFTA, labor arbitrage, and data forecasting mean that the T-shirt on the rack today is cheaper and better than the one your grandfather bought in 1950. The supply chain has been optimized at every stage but one. The last mile was never engineered out of the system. It was designed in.
Autonomous cars will change all of that.
You May (Not) Find Yourself Behind the Wheel of a Large Automobile
Pundits and sci-fi films like Blade Runner have been predicting networked driverless cars for decades. In the 2010s, a Stanford futurist gave a widely cited TED talk on the convergence of four technologies required for full Level 5 autonomy: efficient batteries, self-driving software (cameras, LIDAR, AI), increased computing power, and high-bandwidth communications. Good futurists, like broken clocks, are wrong until they are right.
The timeline to mass adoption is unknowable. For real estate owners, the timing is less important than the inevitability.
Up to now, developer discussions and ULI panels have largely focused on parking. The Sharing Economy mindset saw autonomy as a synonym for shared. Shared meant fewer cars and fewer cars means less car storage. Presto chango, no more parking decks. A car drops you off, another fetches you, and in between the cars go hang out together in some creepy robot dead-storage area.
Maybe that is how it all goes down. If you've been in my car, you might agree it won't: Where will I store my softball coaching gear, Uline catalogues, cassette tapes and broker flyers if I don’t own my car? Car utilization is not the autonomy issue real estate should focus on. Autonomy as a supply chain solution is.
The Last Mile Was Designed Into the System
Consider what arrived at the house this week from online orders. Cat food, coffee beans, compression socks, a smart lock, a book. Not one required a trip to a store. An online button was clicked, and they came, frictionless.
They came at a cost, though: the truck, the gas, the human driver. Low-margin necessity goods and groceries have largely escaped home delivery because of those low costs and because the system wasn’t set up that way. The system wants you to be the last mile. Most households live within an 8-minute drive of a grocery store because proximity reduces friction, and reducing friction is what supply chains do.
Autonomous vehicles change that calculus significantly, eliminating the driver, the fuel cost, and ultimately the need for the trip at all. They remove the last source of friction the supply chain has never been able to eliminate.
Imagine a day in the not-too-distant future: a self-driving car with the softball gear and broker flyers on the floorboard heads out to take the kids to school and comes back with a gallon of milk and a carton of eggs. Maybe it went to the store down the street, or to a warehouse by the interstate. Maybe it even borrowed the eggs from a neighbor. Does it matter?
The Store is the Bottleneck
Back to that Chick-fil-A in McDonough, the one that can process up to 700 orders per hour. The average American grocery store cannot get anywhere near that throughput. The math suggests they process about 8 percent of what that one Chick-fil-A can.
Here’s how it breaks down:
The data on average annual grocery store sales is all over the board. Some say $18M. Others $37M. Call it $32M for argument’s sake.
$32M divided by $8,000 implies that each grocery store serves 4,000 households.
4,000 households multiplied by 75 annual visits equals 55 trips to an average grocery store on an average 15-hour day.
That works out to roughly one customer per minute for a 50,000-square-foot store.
This is a story about real estate utilization, and convenience retail is badly underutilized. One customer per minute for a 50,000-square-foot store is insanely low throughput. A Walmart does better, averaging about 300 visits per hour, but that's a gigantic 188,000-square-foot store, meaning only twice the visitors per square foot (oh yeah, and it’s a grocery, too.).
Now imagine a grocery executive looking at these numbers. Doubling or tripling throughput per store would be transformative, but impossible without retooling and removing the humans. So the decision gets made: build a new kind of store.
It's most likely a warehouse, probably where land is cheap and zoning is easy, perhaps by the airport or near a highway interchange. Two floors, like that Chick-fil-A: groceries sorted above, distributed via conveyor belt to waiting driverless cars below. The cars might belong to customers, to Instacart, or to the grocer itself, snappy Publix-green robot vans with freezer, refrigerator, and warming compartments, running at all hours because pricing is now dynamic, with discounts for overnight pickup and premiums at midday. No humans. The "store" runs 24 hours a day and processes several thousand orders, and the executive realizes that the existing inventory of old-fashioned people stores is almost entirely obsolete.
A warehouse processing even 80 percent of what that one Chick-fil-A can handle runs at roughly ten times the throughput of a conventional grocery store. One warehouse does the work of ten conventional stores. So you now have ten times too many stores.
The Warehouse Replaces the Store
What happens next to the acres and acres dedicated to convenience retail: the dollar stores, pharmacies, gas stations, groceries, and washaterias?
The answer is mixed. Gas stations will likely go first: they are already under pressure from EV adoption and will have no reason to exist once fueling is no longer a human errand. Pharmacies will follow: regulatory changes and low-cost autonomous delivery will make prescription drug delivery not just feasible but inevitable.
Then one of the large-format value grocers, Aldi, Lidl, Walmart, or Target, will be among the first to pull the trigger on the warehouse model. The smart money is on Walmart, which has the logistics infrastructure, the real estate footprint, and the margin pressure to move faster than its competitors. It will take time and there will be kinks, but once it works the race will be on.
Not every format will be caught up in that race at the same pace.
Dollar stores tend to locate in lower-income neighborhoods and rural areas, so perhaps they stick around longer than other formats. The communities most dependent on them are also those the AV logistics buildout will reach last, which means the disruption in those markets will play out on a different timeline than everywhere else.
Costco's resilience comes from a different place. The serendipitous discovery model, the $1.50 hot dog, the inexplicable cashmere sweater next to the motor oil, has proven remarkably resistant to online replication.
At the upscale end, Whole Foods, Trader Joe's, and Central Market will lean into experience and social signaling. The entire purpose of Whole Foods is to demonstrate to the other parents that you are a better one. That is difficult to do from a sofa.
And then there are the grocery obsessives. On a lazy weekend I’ve been known to go to five different ones, six if you count the church farmers’ market. Like a casino whale getting a complimentary car from the Wynn, maybe Wegmans sends a robot van, and the AI assistant and I chat about the coming week's menu on the ride over. No, I didn't know veal cutlets were on sale, and yes, Saltimbocca does sound delicious.
The formats that survive longer will be the ones worth the trip. Autonomous vehicles will just make that bar higher.
What This Means for Real Estate Owners
So what are real estate operators to do in this future Convenience Store Westworld, before Yul Brynner goes crazy and kills us all? A few options.
Buy gas stations. Almost always great real estate. Wait for them to close, repurpose, and redevelop. That is the definition of a Happy Deal: happy if they stay, happy if they go.
Get to know the Walmart surplus property department. A former mentor was the largest buyer of old Walmart locations in the early 2000s. There is a tidy business in re-tenanting these if you have the patience.
Play the long game. Think about what would happen if a tornado flattened an existing shopping center. Would the redevelopment be denser and mixed-use? If so, start that groundwork now. In high-rent neighborhoods, zoning is hard and barriers are high. Entitle it today while the neighbors aren't paying attention.
In low-rent neighborhoods, don't wait. The rent is low, the redevelopment case is thin, and parking requirements may disappear before a viable use emerges. If there is no higher and better use on the horizon, sell now.
Lean into lifestyle. Next to the Central Market, add a cooking school, a James Beard restaurant, a natural wine bar, some wellness concepts, and an Alo Yoga, or a Vuori for the tech bros. Social signaling never goes out of style.
Sell the box, keep the shops. Sell the grocery box today and keep the inline shop space. Get a ROFR if they go dark. Shadow space, the retail square footage that depends on an anchor tenant for its traffic, is an unglamorous hold. But local businesses are more Waymo-proof than a grocery anchor and tend to outlast them. Sell the outparcels too, cash flow the asset into the ground, and let the foreclosure lawyers clean up the byzantine HOA mess when the anchor finally goes dark.
Autonomy is coming. Today it’s happening gradually, but, like bankruptcy, it will one day happen suddenly. And to the owners of convenience retail, now is the time to start planning for that day when that last mile is fully optimized.
George Banks is the founder of Revel, a retail development and advisory firm specializing in destination retail, placemaking, and mixed-use projects across the United States.
