Self-driving cars are likely to be a revenue bonanza.
How high is up?
That’s the rhetorical question often posed when someone asks how much money self-driving cars will potentially be able to make.
Part of the willingness and enthusiasm of VC firms and major automakers to invest in driverless car tech is due to the belief that there is a huge pot of gold at the end of the self-driving car rainbow.
Estimates provided by automotive industry analysts are at times wildly all over the map in terms of the revenue potential. Some have pegged the number into the $500B range for annual revenues, while others say it is more akin to the $1T-$5T in big bucks making per year.
The initial scramble for putting down bets on driverless hardware and software startups has cooled somewhat, partially as a result of the realization that the hyped timetable for achieving true self-driving cars is going to be longer than what many talking heads had gushed about.
Those eager and breathless investors that wanted to grab up land right away were oftentimes doing so with one eye closed, or maybe both eyes were shut and blissfully happy to blindly plunk down cash into anyone or anything that promised to soon-to-deliver driverless cars.
Driverless Kits Afoul
One of the most egregious such offerings involved those that said they could make a low-cost kit for transforming a conventional car into a self-driving car.
Imagine that!
In the United States alone, there are an estimated 250 million conventional cars.
If you could merely buy a kit at your local auto store, or maybe even at Best Buy, wouldn’t it be tremendous to take home the kit, slap it onto your beat-up old jalopy, and voila, you’d now have yourself a genuine driverless car.
The shelves would be stripped clean of those kits and people across this country would be overnight letting the AI drive for them since they would have miraculously turned their cars into AI self-driving passenger-toting vehicles.
It’s a wonderful world.
Of course, the reality is quite afield of that kit notion.
As I’ve repeatedly exhorted, you are not going to be able to shove a bunch of spiffy tech onto a conventional car and have it become a true self-driving car.
Most of the kits that have been floated into the marketplace involve allowing you to trick and jury-rig your car into using automation to steer the wheels and push the pedals, but this is a highly dangerous gambit and fortunately, not many people have fallen for the ruse.
In one sense, crafting a true self-driving car involves brain surgery, and letting the everyday consumer operate in such a capacity is raft with concerns and calamity.
By the way, conspiracy theorists argue that anyone that disses on the kit’s possibilities is doing so to protect the man, the large corporations, and wants to prevent the little guy from being able to afford and have driverless cars. This was frequently used in the pitches for such kits.
All I can say is that I’d welcome finding a means to make driverless cars as inexpensive as possible and have a mobility-for-all advent, though this has to be done safely and with the realization that cars are multi-ton vehicles that can readily harm and kill, so whatever tech is used had better be darned good.
I realize the conspiracy believing camp will retort that I am obviously saying this as part of the conspiracy card-holding schemers.
Oh well, let’s move on.
Back-Of-The-Envelope Approach
When you are sometimes faced with a situation involving an estimation problem, you can resort to using a back-of-the-envelope approach.
A now-classic question asked during job interviews for techies includes pressing them to craft an estimate of the number of manhole covers that there are in New York City (NYC).
The initial reaction to such a question is that it seems utterly irrelevant and an impossibility if one doesn’t have to access to Google search or an equivalent to look-up the answer.
What you are supposed to do is (cleverly) make use of the fact that NYC is laid out on a grid pattern and by using the number of rows and columns, i.e., streets and avenues, you can readily use simple multiplication and arrive at a figure that approximates the manhole cover count (assuming that there is an average of one manhole cover per such intersecting point).
One caveat is that this question presumes the interviewee already knows that NYC is based on a grid shape and happens to know how many rows and columns there are, which is therefore biased toward those that perchance live or are especially familiar with New York.
Some interviewers are okay with a candidate not knowing those facets, as long as the candidate asks probing questions of the interviewer to ultimately ascertain the answer (but, how many of us are willing to ask questions of the interviewer in such a tense setting?).
Some say it’s a rather awkward or untoward question to ask and merely makes an already uncomfortable interview an even more dismal experience.
Returning to the topic of back-of-the-envelope estimations, the notion is that you try to use overall indicators to arrive at a reasonably sound estimate.
The estimate might be off by a bit, perhaps a lot, yet at least you have some general number around which you can then carry on your noodling and use for various added considerations.
People that live their lives in a precise way are usually uncomfortable with back-of-the-envelope methods. For them, the idea of making sweeping assumptions and forgoing getting into the meaty details is antithetical to their world view.
Yes, back-of-the-envelope estimates can be rather raw and crude.
Yes, you ought to interpret such estimates with a grain of salt, or maybe a boxcar full of grain.
Yes, these kinds of estimates can be misused, and inadvertently take on a life of their own as a form of gospel.
Nonetheless, using a back-of-the-envelope technique can be instructive, along with getting the creative juices going and lead to or inspire others to dig deeper and come up with something that either has the underlying details to support the guess or that showcases that the guess was woefully inadequate.
With that important preamble, let’s see if we can estimate via back-of-the-envelope the revenue potential for true self-driving cars.
Remember, please, it’s just back-of-the-envelope.
The Levels Of Self-Driving Cars
It is important to clarify what I mean when referring to true self-driving cars.
True self-driving cars are ones that the AI drives the car entirely on its own and there isn’t any human assistance during the driving task.
These driverless cars are considered a Level 4 and Level 5, while a car that requires a human driver to co-share the driving effort is usually considered at a Level 2 or Level 3. The cars that co-share the driving task are described as being semi-autonomous, and typically contain a variety of automated add-on’s that are referred to as ADAS (Advanced Driver-Assistance Systems).
There is not yet a true self-driving car at Level 5, which we don’t yet even know if this will be possible to achieve, and nor how long it will take to get there.
Meanwhile, the Level 4 efforts are gradually trying to get some traction by undergoing very narrow and selective public roadway trials, though there is controversy over whether this testing should be allowed per se (we are all life-or-death guinea pigs in an experiment taking place on our highways and byways, some point out).
Since semi-autonomous cars require a human driver, the adoption of those types of cars won’t be markedly different than driving conventional cars, so I’m not going to include them in the back-of-the-envelope estimation for true self-driving cars.
For semi-autonomous cars, it is equally important that I mention a disturbing aspect that’s been arising, namely that in spite of those human drivers that keep posting videos of themselves falling asleep at the wheel of a Level 2 or Level 3 car, we all need to avoid being misled into believing that the driver can take away their attention from the driving task while driving a semi-autonomous car.
You are the responsible party for the driving actions of the car, regardless of how much automation might be tossed into a Level 2 or Level 3.
Self-Driving Cars And Potential Revenue
For Level 4 and Level 5 true self-driving cars, there won’t be a human driver involved in the driving task.
All occupants will be passengers.
Many have predicted that we’ll have a mobility transformation and essentially become a mobility-as-a-service economy, shaped around ridesharing on steroids.
Some believe that only large companies such as the automakers and maybe the prominent ridesharing firms will own and operate driverless cars. This will be done based on large fleets of self-driving cars that they opt to establish and deploy.
Furthermore, other large companies that today have nothing to do with cars at all are predicted to jump into the self-driving car bonanza.
It would seem easy enough to accomplish. Buy a bunch of pricey driverless cars, put them onto an online sharing network, and count the money as those AI-driven cars do your bidding and give people rides.
I’m known as a somewhat contrarian since I claim that we’ll also still have individual ownership of such cars.
My logic is that there is potentially a lot of money to be made by owning and leveraging your own self-driving car, thus we all will have a monetary incentive to want to do so.
Why let only the big companies get the dough when the mom-and-pop can do so too (see my detailed explanation here)?
In any case, there doesn’t seem to be any disagreement per se that mobility via driverless cars will become more friction-free and the advent of self-driving cars will be a mighty disruption to our society (hopefully, a positive disruption).
The roughly 70 billion annual hours that Americans spend driving their cars today will shift those drivers into becoming passengers. That’s a lot of time opened-up for other activities, doing so while commuting in a self-driving car that’s outfitted with Internet access and other goodies.
It seems fair to assume that riding in a driverless car won’t be for free and you’ll need to pay some price to do so.
What will be the price and how will it be calculated?
Nobody knows for sure.
How much will people ride in driverless cars?
Nobody knows for sure.
Well, maybe we can take a poke at this and come up with something.
Currently, in the United States alone, we rack-up an estimated 3.22T miles of driving annually. You might not have seen that statistic before, and upon reflection, it really is somewhat staggering. That’s a lot of miles.
For a back-of-the-envelope estimation, let’s go ahead and pretend that all those miles of travel would be shifted entirely to driverless cars.
Thus, we are going to make a series of assumptions that include the aspects of having only driverless cars and no conventional cars on our roadways, and that people will be going the same kind of distances that they did while they were drivers of cars, etc.
We’ll come back to those assumptions momentarily.
What kind of price might we come up with as a per-mile fee for using a driverless car?
Today’s estimates tend to suggest that ridesharing via the use of human drivers is around $1.00 or more on a per-mile price (this can be higher and lower in some locales and various times of the year), and some pundits have predicted that the per-mile price for driverless cars is going to be about one-third of that amount, so let’s use $0.35 as a round number.
By the simplest of math, we can calculate this:
· Estimated annual revenue = miles driven x per mile fee
· $1.12T annual revenue = 3.2T annual miles x $0.35 per mile fee
In the broadest of manners, we now have an estimate of the annual revenue that could be derived by self-driving cars.
It’s a tad over a trillion dollars.
If you compare that estimated number to published charts of the annual revenue of various industries in the United States, the revenue of self-driving cars would be near the top of the chart and be on par with the real estate industry and other gigantic industries.
That’s nice.
It also showcases why there are big companies eyeing the driverless car evolution. There does seem to be a pot of gold at the rainbow, really, for real, and not just a mirage.
The question today is how long will it take to reach the pot of gold and whether you can stomach making the investment now that will gradually and eventually turn to gold, but meanwhile, it is pretty much all R&D and not something that’s turning a buck right away.
With our business culture being seemingly run by quarterly reports, it is hard to stay the course on an investment that seems attractive and yet it is bleeding money for now, and perhaps for quite a while ahead.
You’ve got to have some pretty strong and visionary guts to keep in the game.
This is not for the faint of heart.
There’s More To The Estimation
I’ve mentioned that the 70 billion hours of driving time today will become 70 billion hours of sitting inside a driverless car.
This is worthy of consideration as it offers some additional and quite promising revenue opportunities.
Think of the advertising possibilities!
You’ll have people trapped inside their self-driving cars, which I mean in a positive way, suggesting that they will be riding around in self-driving cars and presumably have nothing to do since they no longer need to be attentive to the driving task.
Plus, driverless cars will be outfitted internally with various LED displays so that you can watch video streaming via the 5G (and greater) online access to the Internet.
Advertisers will salivate at this prospective market potential.
Beam your ads into those driverless cars, capturing the attention of the riders, and you can use both short-form ads and longer-form ads (if the average commute time says 30 minutes to an hour, just imagine how much ad time that allows for).
There are lots of ways to further monetize driverless cars, including in-car online entertainment and infotainment, Internet of Things (IoT) access, and so on.
For the moment, it seems reasonable to push upward the $1.12T by suggesting that there will be at least new ad dollars arising via self-driving cars too. We don’t know how much money that will bring in, but let’s just use 20% on top of the existing estimate, and we’ll go with an expanded revenue of say $1.34T.
Another facet to consider is the likelihood of induced demand to appear.
Induced demand refers to the notion that once you make something available, there might be additional demand that comes out of the woodwork that otherwise was suppressed.
Some believe that driverless cars will make finally a mobility-for-all world possible.
Perhaps the number of miles driven today is too low a base to use since it doesn’t account for the induced demand that will be unleashed via driverless cars.
Also, the number of miles driven in the United States has been rising lately on the amount of about 2.8% per year, therefore by the time that driverless cars are prevalent we might be consuming a lot more miles than we are today.
The expanded formula becomes:
· Estimated annual revenue = ((Normal annual miles + Growth annual miles + Induced annual miles) x $0.35 per mile fee) + Ads annual revenue + Infotainment annual revenue + IoT annual revenue + Other annual revenue
As a side note, let’s not deal with future dollars and for ease of discussion calculate everything in today’s dollars, which notably ignores inflation and other economic elements that will impact future dollars.
Let’s uplift the number of miles from 3.22T by perhaps 20% for an overall increase in ridership over the next several years and then another 20% for the anticipated induced demand, so overall boosting the 3.22T by about 40%, turning it into 4.50T, and then use our $0.35, producing $1.58T, and then toss-in the ads too, landing us at a revenue of around $1.89T.
We could also play with the $0.35 and raise it up since there is the possibility of charging more for driverless cars.
Try to not bicker and grumble about these fast-and-furious numbers, since it’s once again our snapshot back-of-the-envelope approach coming to play.
All-told, we seem to be arriving at a number of at least between $1T to $2T.
Is this reasonable?
You might have seen some pundits that have estimated in the $5T to $7T range, which is generally aligned with my estimates herein since those larger predictions are based on a global or worldwide estimate.
Notably, keep in mind that so far I’ve only discussed the United States market potential, yet one would be wise to see the larger picture of the massive size of the international markets too.
Conclusion
One of the frequent jokes made about the act of making assumptions is that when you “assume” it makes a “mule” out of you and me (there’s a different word that goes in place of the word mule, I’m trying to keep things clean).
A rather imposing assumption earlier stated was that we would have all driverless cars and no conventional cars on our roadways. I’ve said over and over that the roll-out of driverless cars is going to take place over many years, likely decades.
Thus, earning that back-of-the-envelope $1T to $2T of annual revenue is quite some distance off in the future.
As part of this noodling exercise, I decided to go beyond just a hastily scrawled envelope and opted to craft an overarching spreadsheet that postulates a gradual build-up of driverless cars, starting at 0.01% of the total volume of cars (some consider this a key milestone), and then reaching crucial thresholds such as at 1% of all cars, 5% of all cars, 10%, 20%, 50%, 75%, 80%, 95%, and 99% (perhaps never reaching 100% per se in our lifetimes).
Interestingly, the chart suggests that once we reach the 50% mixture we’ll cross the $500B mark, and then the vaunted trillion dollars arrives once we get above 75% (but that’s more of the back-of-the-envelope kind of guesswork being employed).
I also included time as a factor, trying to anticipate when this will all emerge.
Once again, it’s all back-of-the-envelope based, and the inclusion of a spreadsheet doesn’t in itself make the numbers more prophetic.
There’s a fascinating allied question of how many driverless cars we will need on our roadways to sustain the incurring of those 3.22T (or more) miles, which is a topic that I’ll be covering in a later column.
Generally, most believe that we won’t need 250 million cars of a driverless nature, and instead can provide sufficient coverage by utilizing a substantially lower number of such self-driving cars, which I also tend to agree with.
Another serious and obviously necessary topic entails what will be the costs associated with deploying a driverless car. The revenues might be mighty, but perhaps the costs are mighty too.
It is useful to realize, for example, the assertion that most driverless cars will only last about 4 years in-service due to the heightened usage demands (see my details here for further explanation).
Anyway, hope you’ve found this exercise useful.
I’ll grab up a few more envelopes to noodle some more on the myriad of allied topics and see what added scrawling can produce.
