There are two foundational technology changes rolling through the car industry at the moment; electric and autonomy. Electric is happening right now, largely as a consequence of falling battery prices, while autonomy, or at least full autonomy, is a bit further off - perhaps 5-10 years, depending on how fast some pretty hard computer science problems get solved. Both of these will cycle into essentially the entire global stock of (today) around 1.1bn cars over a period of decades, subject to all sorts of variables, and both of them completely remake the car industry and its suppliers, as well as parts of the tech industry.
Both electric and autonomy have profound consequences beyond the car industry itself. Half of global oil production today goes to gasoline, and removing that demand will have geopolitical as well as industrial consequences. Over a million people are killed in car accidents every year around the world, mostly due to human error, and in a fully autonomous world all of those (and many more injuries) will also go away.
However, it's also useful, and perhaps more challenging, to think about second and third order consequences. Moving to electric means much more than replacing the gas tank with a battery, and moving to autonomy means much more than ending accidents. Quite what those consequences would be is much harder to predict: as the saying goes, it was easy to predict mass car ownership but hard to predict Wal-mart, and the broader consequences of the move to electric and autonomy will come in some very widely-spread industries, in complex interlocked ways. Still, we can at least point to where some of the changes might come. I can't tell you what will happen to car repairs, commercial real-estate or buses - I'm not an expert on any of those, and neither can anyone who is - but I can suggest that something will happen, and probably something big. Hence, this post is not a description of what will happen, but of where it might, and why, with some links to further reading.
Moving to electric reduces the number of moving parts in a car by something like an order of magnitude. It's less about replacing the fuel tank with a battery than ripping out the spine. That remakes the car industry and its supplier base (as well as related industries such as machine tools), but it also changes the repair environment, and the life of a vehicle. Roughly half of US spending on car maintenance goes on things that are directly attributable to the internal combustion engine, and much of that spending will just go away. In the longer term, this change might affect the lifespan of a vehicle: in an on-demand world vehicles would have higher loading, but absent that, fewer mechanical breakages (and fewer or no accidents) might mean a longer replacement cycle, once the rate of technology implementation settles down.
Next, gas itself is bought in gas stations, of which there are about 150k in the USA. Those will also go away (unless there are radical changes in how long it takes to charge an EV). Since gas is sold at very low margins, these retailers make their actual money as convenience stores, so what happens to the products that are sold there? Some of this demand will be displaced to other retailers, and some may be going online anyway (especially if an Amazon drone can get you a bag of Cheesy Puffs in 15 minutes). But snacks, sodas and tobacco sell meaningful proportions of their total volume as impulse purchases attached to gasoline. Some of that volume might just go away.
Tobacco in particular might be interesting - well over half of US tobacco sales happens at gas stations, and there are meaningful indications that removing distribution reduces consumption - that cigarettes are often an impulse purchase and if they're not in front of you then many smokers are less likely to buy them. Car crashes kill 35k people a year in the USA, but tobacco kills 500k.
Gasoline is taxed, much less in the USA than in many other developed markets: it is 4% of UK tax revenue, for example. That tax revenue will have to be replaced, with other taxes on things that may be more elastic, and there will be economic and political consequences to that. In the USA, for example, highways are funded partly from gas taxes that have not risen to match inflation since 1993 - if just keeping it flat in real terms was politically impossible, how hard will it be to take that revenue from some other part of the economy?
Conversely, in many places (especially emerging markets) fuel is subsidised by the state - coal, gasoline and kerosene (for light and heat - see for example kerosene subsidies in India). EVs on one hand and solar on the other may change this as well.
Meanwhile, of course, we will still actually need to charge our EVs. Most estimates suggest that charging a fully electric fleet would lead to 10-20% more electricity demand. However, a lot depends on when they're charged: if they're charged off-peak this might not need more total generating capacity, though it would still change output and perhaps local distribution. The carbon impact of shifting electricity generation in this way is pretty complex (for example, over 75% of French electricity generation today comes from nuclear power), but in principle at least some grid generation almost always now comes from renewables.
More speculatively (and this is part of Elon Musk's vision), it is possible that we might all have large batteries in the home, storing off-peak power both to charge our cars and power our homes. Part of the aim here would be to push up battery volume and so lower their cost for both home storage and cars. If we all have such batteries then this could affect the current model of building power generation capacity for peak demand, since you could complement power stations with meaningful amounts of stored power for the first time.
The really obvious consequence of autonomy is a near-elimination in accidents, which kill over 1m people globally every year. In the USA in 2015, there were 13m collisions of which 1.7m caused injuries; 2.4m people were injured and 35k people were killed. Something over 90% of all accidents are now caused by driver error, and a third of fatal accidents in the USA involved alcohol. Looking beyond deaths and injuries themselves, there is also a huge economic effect to these accidents: the US government estimates a cost of $240bn a year across property damage itself, medical and emergency services, legal, lost work and congestion (for comparison, US car sales in 2016 were around $600bn). A similar UK analysis found a cost of £30bn, which is roughly equivalent adjusted for the population. This then comes from government (and so taxes), insurance and individual pockets. It also means jobs, of course.
Even simple 'Level 3' systems would cut many kinds of accident, and as more vehicles with more sophisticated systems, moving up to Level 5, cycle into the installed base over time, the collision rate will drop continuously. There should be an analogue of the 'herd immunity' effect - even if your car is still hand-driven, my automatic car is still much less likely to collide with you. This also means that cycling would become much safer (though you'd still need to live close enough to where you wanted to go), and that in turn has implications for public health. You might never get to zero accidents - the deer running in front of a car might still get hit sometimes - but you might get pretty close.
That, in turn, has consequences for vehicle design - if you have no collisions then eventually you can remove many of the safety features in today's vehicles, all of which add cost and weight and constrain the overall design - no more airbags or crumple zones, perhaps. A decade ago the NHTSA estimated that the safety measures that it mandates collectively added $839 (in 2002 dollars so $1,136 now) and 125 pounds of weight, which was 4% of both average cost and average weight - this is probably a lower bound. That, of course, presumes that there are no other changes to the design as a result of removing the human controls - which is like removing the reins from a horseless carriage and thinking nothing else will change.
What else, though?
As more and more cars are driven by computer, they can drive in different ways. They don't suffer from traffic waves, they don't need to stop for traffic signals and they can platoon - they can safely drive 2 feet apart at 80 mph. There is a whole range of human behaviors that reduce road capacity, especially on freeways: it's not just that people make mistakes, but that computers can drive in totally different ways to even a perfect human driver. The video below illustrates one of these issues, familiar to anyone who's been stuck in a traffic jam on a highway and got to the front to find no apparent cause - human behaviour causes traffic waves, which cause 'phantom jams'. Computers wouldn't do this, and if they did, we could stop them.
A full autonomous road system changes traffic less from fluid dynamics than from circuit-switched to packet-switched, or, more precisely, from TDMA to CDMA. No lanes, no separation, no stopping distances, and no signals, (except of course for pedestrians to cross), means profoundly different traffic patterns.
Clearly, all of this will have some effect on congestion and road capacity. Accidents themselves cause as much as a third of congestion (estimates vary a fair bit and depend whether you're talking about highways or city centres), even if there are no changes from different driving behavior. How much changes over all, though - how much more traffic can a highway hold? How much more quickly do you get to school in the morning if you drive at the same speed but don't have to stop at every stop sign just in case there's someone there? We'll find out.
However, the impact of autonomy on traffic and congestion is more complex than just making driving itself more efficient. Though automatic driving should increase capacity, we have known for a long time that increased capacity induces more demand - more capacity means more traffic. If you reduce congestion, then more people will drive, either taking new trips or switching from public transport, and congestion might rise back to where you started. Conversely, removing capacity can actually result in less congestion (and there's more complexity here too - for example, Braess' paradox). So, autonomous driving gives us more capacity, and in a sense it does so for free, since we don't have to build roads, just wait for everyone to buy new cars, but it also gives us more use.
Parking is another way that autonomy will add both capacity and demand. If a car does not have to wait for you in walking distance, where else might it wait, and is that more efficient? Does that enable better land use, better traffic routing and more or less congestion? And, in parallel, everything that you do to make traffic, driving and now also parking more efficient tends to generate more demand.
So, the current parking model is clearly a source of congestion: some studies suggest that a double-digit percentage of traffic in dense urban areas comes from people circling around looking for a parking space, and on-street parking ipso facto reduces road capacity. An autonomous vehicle can wait somewhere else and an on-demand one just drops you off and goes off to collect other people. On the other hand, both of these models create new trips as well - both your car and an on-demand car would have to come to get you (though, since cars will be automatic, they will form an orderly queue). But with enough density of on-demand, the car you get into might be the car that's already passing, or that dropped someone else off 50 feet away - it all depends on the load factor.
Parking itself is important not just as a part of the traffic and congestion dynamic but as a cost and as a use for property. As mentioned above, some parking is on-street, and so removing it adds road capacity or allows you to add more space for pedestrians. Some of it is at work or retail, or more generally in city centres, and so that land becomes available for other uses. And some of it is at home, either on-street (again using capacity) or in drives and garages, parking lots or parking structures, which add to the cost of housing. The extreme case here is Los Angeles: it has been estimated that 14% of the incorporated land of LA county is used for parking. Adding parking to a new development pushes up construction costs: parking garages cost money, and so does leaving land vacant for parking lots. A study in Oakland, in the San Francisco Bay Area, found that government-mandated parking requirements pushed up construction costs per apartment by 18%. Back in LA, adding underground car-parking to a shopping mall might double the construction cost. If you both remove those costs on new construction, and make that space available for new uses, how does that affect cities? What does it do to house prices, or to the value of commercial real-estate?
Pretty much all of these themes feed into the potential of on-demand. If you remove the cost of the human driver from an on-demand trip, the cost goes down by perhaps three quarters. If you can also remove or reduce the cost of the insurance, once the accident rate has fallen, it goes down even further. So, autonomy is rocket-fuel for on-demand. This makes it much easier for many more people to dispense with a car, or only have one, or leave their car at home and take an on-demand ride for any given trip.
This obviously has consequences for parking - an on-demand ride to work or a restaurant removes parking in the city centre, and not owning a car and substituting on-demand entirely removes demand for residential parking. And, as mentioned above, using an on-demand ride instead of looking for parking gets rid of one kind of traffic but creates a new kind - potentially a smaller one, through.
However, truly cheap on-demand has more consequences still. For example, it displaces demand from public transport - though the cost of a bus driver is also large part of the cost of the trip, and those drivers might not be needed either, so buses might also be cheaper. Conversely, if congestion falls then buses could become more attractive than other forms of transport (both cars and also subways) because the journey time would be shorter (or at least more predictable). This of itself has all sorts of cascading effects. Do you end up with reduced bus schedules? Do marginal bus-routes close, pushing people onto on-demand who might not otherwise have used it - if they can use it? Does a city provide, or subsidise, its own-demand service to replace or to supplement buses in lower-density areas? Does your robotaxi automatically drop you off at a bus stop on the edge of high-traffic areas, unless you pay a congestion charge? This all then ripples back into congestion - buses carry people at higher density than cars, and so replacing a fully loaded bus with cars would inherently create more traffic volume, but buses do not in fact travel full all of the time, and can create their own congestion (an endemic issue in London's Oxford street, for example). And, especially on Oxford Street, they carry more people than cars because they're aggregating people onto a single route who might otherwise have taken many other separate, more direct or more efficient routes. If 50 people on a bus switch to cars, they won't all be on the same road at the same time. Meanwhile, the fixed cost of a bus creates a minimum loading level and density at which a bus is practical - breaking this apart into smaller vehicles - maybe with one passenger, maybe with 10 - might extend 'public' transport to many more people.
Perhaps the most useful way to think about this is that, just as on-demand erodes the difference between marked and mechanically metered taxis and car-services, so it also erodes the difference between both of those and buses. What exactly are the differences in traffic dynamics between a Lyft Line shuttle with 5 passengers and a municipal bus with an off-peak load of 10? Recall, too, that buses weren't always municipal, and there are parallel commercial alternatives today - see Chariot, or matutus.
The point here is not remotely to suggest that it is inherently good or desirable to replace public transport with cars, but that it now becomes possible to do so, if we want, and that it might be cheaper and more efficient in some circumstances. And, indeed, that the distinction between 'car' and 'bus' might break down.
Then, of course, there are the drivers. There are something over 230,000 taxi and private car drivers in the USA and around 1.5m long-haul truck-drivers. The question of what happens to taxi and on-demand drivers has been discussed too widely and publicly for me to add anything here, but long-haul truck drivers have some interesting nuances (I'm here excluding local delivery drivers as they're often needed for more than driving the truck itself and robotics is a whole other conversation). The average age of a long-haul driver is now 49, and around 90 thousand leave the industry every year, half though retirement. The industry thinks it has a shortage of around 50,000 drivers, and growing - people are leaving faster than they can be replaced. Truck driving can be an unhealthy, uncomfortable job with a difficult lifestyle. Hence, on these numbers, over half the current driver base will have left in ten years, around the time that most people think full, level 5 autonomy might be working. In the short term, level 4 autonomy makes truck-driving more attractive, since you can rest in the back of the truck until you're needed instead of having to stop at mandated times. But on a 20-30 year view, which is really the timeline to think about this transition, effectively all current truck drivers will have quit anyway - you won't replace them, but you won't necessarily put anyone directly out of work - until you start looking at truck stops, which takes us right back to the convenience store discussion at the beginning of this piece. And meanwhile, truck-stop operators are already starting to think about the fundamentally different trucking patterns that might come from a shift in the logistics industry away from serving traditional retail and towards serving ecommerce (i.e Amazon).
Pulling all of these threads together: if parking goes away, road capacity increases by, perhaps, several times, and an on-demand ride is the cost of a coffee, then one needs to start thinking much more generally, not just about cars, trucks and roads but cities, land use and real-estate. In fact, one needs to think about cities. Cars have remade cities over the past century, and if cars are now going to change entirely, cities will change too.
So, big-box retail is based on an arbitrage of land costs, transport cost and people's willingness to drive and park - how does autonomy change that? How do cities change if some or all of their parking space, especially in town centres, is now available for new needs, or dumped on the market, or moved to completely different places? Where are you willing to live if 'access to public transport' is 'anywhere' and there are no traffic jams on your commute? Does an hour-long commute with no traffic and no need to watch the road feel better or worse than a half-hour commute stuck in near-stationary traffic staring at the car in front? How willing are people to go from their home in a suburb to dinner in a city centre on a dark cold wet night if they don't have to park and an on-demand ride is cheap? What happens to rural pubs if you don't have to worry about drink-driving anymore? And what do you DO in the car, while it's taking you somewhere? Long Netflix and brewers, short BAT - and medevac helicopters.
Finally, remember the cameras. Pretty much every vision of automatic cars involves them using HD, 360 degree computer vision. That means that every AV will be watching everything that goes on around it - even the things that are not related to driving. An autonomous car is a moving panopticon. They might not be saving and uploading every part of that data. But they could be.
By implication, in 2030 or so, police investigating a crime won't just get copies of the CCTV from surrounding properties, but get copies of the sensor data from every car that happened to be passing, and then run facial recognition scans against known offenders. Or, perhaps, just ask if any car in the area thought it saw something suspicious.