Want to know the future of the car? Ask a horse!
Transitions driven by technology happen fast. What can we learn from horses to predict the future of the car?
Car enthusiasts are going to find much of this article horrifying, as well as asking you not to shoot the messenger I would say read on to the end, as I think there is some light at the end of the tunnel.
Horses were the dominant form of ground transport for thousands of years, yet within a few years at the start of the 20th century, they were displaced by the car for personal transport and most other uses. Now, most horses in the developed work are owned purely for pleasure.
Spot the horse
The speed of that transition can be seen in these two photos from New York taken just 13 years apart. One car amongst a mass of horses in 1900 and by 1913 only a single horse can be seen amongst a mass of cars!
Overall the transition followed an S curve of disruption:
Horse to car transition S Curve
The car was invented by Carl Benz in 1886, but remained at the bottom of the S Curve until Ford introduced the Model T in 1908, this was the tipping point that quickly drove adoption. Price was a key driver, the Model T started at $825, but within 6 years the price halved to less than $400. A key technology enabler of the transition to cars was petroleum - a few decades earlier, in 1850, Scottish scientist James Young patented the distillation process that created petrol from oil and shortly afterwards founded the first commercial oil refinery, this created a great and cheap way of storing and delivering energy.
Why did people dump the horse so quickly after thousands of years? The answer is a combination of superior economics and capabilities. Within a short period of time cars went from a bit of a joke to complete disruption because the total cost of ownership of a Model T was soon better than a horse and it could do more and go faster. All the more impressive when you think all the infrastructure of petrol refining, filling stations and car suitable roads had to be built, while World War I and the Flu Pandemic were fought.
Internal Combustion Engine (ICE) road transport is about to go through the same S curve transition during the 2020's, by the end of the decade most will have gone electric and there will be far fewer cars due to the rapid emergence of Robotaxis, literally driven by AI / Autonomous driving, at a fraction of the cost of current car ownership.
Key Technology Enabler #1 - Lithium-Ion Batteries
Much as petrol refining was the key technology enabler of the original car, so the lithium-ion battery enabled EVs. The basic science was discovered in the 1970s and the first commercial lithium-ion batteries were launched, by Sony, in 1991. The cost of batteries has dropped dramatically and is projected to continue to do so:
The main reason EVs are currently more expensive to purchase than a conventional ICE is the cost of batteries. $100 per KWh is widely considered to be the tipping point for price parity. NEF Bloomberg project that new price parity will be reached by around 2026 and will then drop below ICE. Indeed both Tesla and VAG have detailed their plans to reduce battery costs by over 50%. In addition to reducing the cost of the battery itself, car manufacturers are driving down costs by creating dedicated EV platforms, which are lower cost to manufacture compared to retrofitting batteries and motors to existing ICE chassis. Innovations such as structural batteries will drive costs down further.
Note: battery cost reduction is largely down to improvements in battery chemistry and manufacturing techniques, not the price of the raw materials.
This varies a little by car type, but even for small cars is reach by 2027:
Read the full report: BNEF Long Form Template (Grid) (transportenvironment.org)
EV running costs are lower than ICE, due to a combination of petrol vs electricity cost, less maintenance and temporary tax incentives. Meaning the total cost of ownership can already be lower than ICE, so once new purchase price parity is reached EVs will have a much lower total cost of ownership making them economically very compelling to most buyers, even without tax incentives, which are only a temporary solution to drive early adoption and address short term cost issues.
The battery tech improvements that are reducing costs are also increasing range and decreasing charging times. While EV charging infrastructure is getting increasingly better, for example in the UK there are over 40,000 chargers and between 500 and 700 are added every month. With the combined economic advantages, removal of range and charging concerns, increasing public concern with climate change and government incentives and looming bans it seems highly likely EVs will follow the same disruptive S curve disruptive adoption curve as the rapid transition from horses in cars. If you think it will take a long time to build out the infrastructure remember how a much less advanced world 120 years ago built out the infrastructure for ICE in a very few years.
Key Technology Enabler #2 - Autonomous Driving
Road transportation is simultaneously undergoing two technology-enabled transitions. In addition to EVs, we are on the cusp of a much bigger disruption - Autonomous driving.
This isn't about level 2 driver assists like Tesla's current AutoPilot or FSD. The transformation is level 4 and 5 full autonomy. This is ripe for another very rapid S-curve transition - currently, cars have a utilisation rate of around 4% - they spend 96% of their time parked, doing nothing. Entrepreneurs know a 4% utilization industry is just waiting for a massive disruption.
Think Uber, but with no driver. Combine the running costs savings of an EV with no driver costs and the utilization rate of a taxi and you end up with profitable "Transport as a Service" model that works out between four and ten times less expensive than buying and running a new car and even two to four times less than keep running an existing car. The savings projections for a US family is calculated at nearly $6,000 per year. This scale of saving is one of the key factors that killed the horse.
Once this arrives the S curve will be very steep, meaning quick adoption, like we saw the horse transition to the car going from 10-80% within ten years. The question then is - how far off are we from large scale fully autonomous driving?
I believe the answer is not long, certainly well before 2025. There is both good evidence and reasons behind this.
Evidence Autonomous Driving is getting close
Waymo has been running fully autonomous Robotaxis in Pheonix Arizona since 2020, with full regulatory approval and is now in the setup phase to do the same in San Francisco. Take a look at someone riding in one:
Cruise, owned by GM, was also granted a license to start operating autonomous Robotaxi's in San Francisco in September 2021. Auto-X is doing similar in China and is already running driverless Robotaxi's in five cities in China. A recent test crowded and the chaotic street is an impressive demonstration:
Tesla is heading for the same goal and has said they intend to both operate their own fleet of robotaxis and allow Tesla owners to earn money by allowing them to be used as robotaxis when they aren't needed. Their approach is a little different to Waymo, Cruise, and Auto-X - rather than training dedicated robotaxis in specific cities, they are taking the harder path of training their cars all the driving techniques needed to drive anyway (or at least the US initially). This initially takes longer, but if it pays off they could be first to the mass market.
They have been developing and testing the re-written FSD software with around 1,000 customers for the last few months and as of late September 2021 have started to open up testing to more users. What has been most impressive is how the software has been learning and getting better. While it is still only as good as a human learner driver I've seen it's been progressing rapidly and I can see that if it continues to learn at a similar rate it will reach the level of a qualified driver, probably during 2022, particularly as Tesla add their DoJo supercomputer to the mix. An example I like is Lombard Street in San Francisco. I've driven it myself and it's hard. Up until very recently, FSD couldn't do it, yet here it is, in the dark, doing the road.
Search YouTube for "FSD Beta" and you will find loads of videos in all sorts of city and rural locations, even on dirt tracks, in the snow and heavy rain with lots of traffic. It's far from perfect yet, but learning fast, which is the key point.
Reasons Autonomous Driving is getting close
Fundamentally it's due to Moore's Law - compute power doubles every ~2 years, which results in an exponential increase. It's the reason your Smart Phone can take amazing photos and the latest iPhone 13 can even create convincing blurred backgrounds while shooting and editing videos, or you can play almost photo-realistic games on the latest Playstations, Xbox's and PCs. We can now carry more computing power in our pockets than a multi-$M supercomputer that filled a room only a few years ago and that keeps on doubling every two years!
In the last few years, scientists and businesses have been increasingly focussing on Artificial Intelligence and Machine Learning. Both in hardware, such as the neural processing units found in modern smartphones, graphics cards, supercomputers and autonomous driving computers and in incredible software, such as Google's DeepMind and the autonomous driving software from Waymo, Cruise, Auto-X and Tesla. These accelerating innovations in hardware and software are what underpins making autonomous driving a reality in the not too distant future.
EVs and Autonomous Driving - The perfect storm to disrupt road transport with Transport as a Service
These two massive changes coming at nearly the same time are highly complementary and will drive change - the total cost of ownership economics advantages of EVs combined with utilisation and operating cost advantages of autonomous driving will enable businesses to offer Transport as a Service (think Uber with no driver) at prices and convenience levels that will be extremely compelling to a great many families. Simply open your smartphone and be on your way within a few minutes at fraction of the price of your own car.
The Implications
Once Transport as a Service (TaaS) becomes technically possible and gets regulatory approval (remembering it already has in a few cities around the world) it will start off a rapid transition, similar to the horse to car transition 120 years ago. Here is a projection from a study by RethinkX:
Download the study from: https://www.rethinkx.com/transportation#transportation-download
Meaning that most miles will be driven by Robotaxi's by 2030. Given that personally owned cars have utilization of around 4%, but Robotaxi's will be MUCH higher, this means far fewer cars are required. RethinkX projects a rapid decline in the number of privately owned cars and the total number of cars goes down too because the much higher utilization rate of Robotaxis requires far fewer cars:
This will start in cities where car ownership is expensive and population density quickly allows great coverage of robotaxis. It will then ripple out into the suburbs, smaller towns and eventually many rural areas.
Likely we will see both per mile payment models, like Uber, but also fixed monthly subscriptions, like NetFlix. Probably these models will include some flexible options - for example standard Transport as a Service most of the time, but access to cars for an extended period of time, including cars you drive yourself, for example when going on a family holiday, or wanting to go out for a fun drive!
For this article, I've been focussing on cars, but the same dynamics almost certainly apply to the rest of road transport. All forms of trucking / HGVs / deliveries will move to EV powered autonomous vehicles. Most of this is likely battery-based as it has the lowest operating costs, but hydrogen may have a role in the longest range HGVs where greater range and refuelling time is more important. Although this is a bit uncertain as battery tech keeps improving it may even shut the door on hydrogen in these applications?
Transport as a Service - Not for everyone
There will be groups where this either doesn't make sense or people prefer to own.
Very rural locations will probably struggle with low population density where the cost and convenience don't stack up.
Car enthusiasts, including me! will often prefer to own. Although here you might see a multi-car family dropping to one car owned for the enthusiast and the rest of the family's transport needs served by transport as a service. While richer families may choose car ownership as a status symbol or simply because they like the feel of owning a car.
People with professions that make a lot of use of their transport - for example, truck owners who transport and store work equipment in their truck.
Positive Implications
Economically families and nations benefit - almost $6,000 per year per family adding over $1T to the US economy. Plus a potential $1T benefit from increased productivity - people can work during travel. Those on low incomes will have much better access to high quality, fast and affordable transport.
Climate change benefits because this drives a much quicker transition to EVs. The environment also benefits as there will be far fewer cars needing to be manufactured, so less mining and processing, resulting in fewer production emissions and less environmental impact from mining.
A lot of precious space will be freed up in towns and cities. For example, LA has around 100 square miles of car parks, enough for four Manhattans. Most cities and towns have a lot of space locked up in car parks and street parking. That is a lot of space for parks, businesses and homes. Residential streets would look a lot more attractive and be safer places with fewer cars parked on them.
A lot of lives will be saved - around 1.3 million die from road accidents and 10 million from air pollution globally per year. While EV robotaxis won't eliminate this they will almost certainly reduce it a lot. It's very unlikely regulators will grant licensees unless a high level of safety can be demonstrated.
Travel times will probably be shorter too. In congested urban environments, robotaxis can be connected to real-time traffic management that can better manage the flow of traffic. On long freeway/motorway journey's the combination of higher safety, faster reactions and real-time connection have the potential for governments to hugely increase the speed limits for robotaxis without risking safety. We might see cordoned off robotaxi lanes with either very high or no speed limits, a little like Germany has on some of its autobahns today.
Some companies will do very well out of this transition. Those already in Transport as a Service are well-positioned, probably explaining why Uber is valued at $88B and Lyft at $18B and why Google has invested so much in Waymo and GM bought Cruise. It also explains why Tesla is pushing so hard to complete its FSD software - it's not doing it just to be able to generate income from selling FSD software options to car owners, rather it's about being one of the biggest winners of Transport as a Service revolution. Those car companies that are successful in offering Transport as a Service will benefit from higher profit margins, even if total revenue may be less in some cases. For example, GM's Cruise is projecting $50B in robotaxi revenue.
As a massive market disruption, it will almost certainly create new opportunities that people haven't even thought of yet. Entrupriers are good at finding opportunities from transitions.
Negative Implications
Some types of jobs will almost completely disappear - not great if you are a taxi or truck driver. Others will shrink a lot - less need for mechanics, car salespeople, automotive production workers, etc.
The oil industry will be massively disrupted with demand for oil plummeting:
Oil demand plummets
As demonstrated by COVID - when demand for oil shrinks then the price drops very quickly. Projections for oil price with this much-reduced demand are ~$25 per barrel. That will make some forms of oil extraction uneconomic, such as much deep-sea drilling and oil sands. Great for the environment, but not so great for companies, workers and investors in the oil industry.
Take another look at this adoption curve. The dark part of the graph is stranded individual car ownership. With Transport as a Service cheaper than even running an existing car lots of people will sell their cars to save money. But that will create a very big and expanding supply of second-hand cars but a limited market of people wanting to buy them. Standard economics means the price of those second-hand cars will drop rapidly. This will be particularly severe for standard "transport" cars. Probably enthusiasts cars will hold value better as there will remain a strong demand for those sorts of cars.
Stranded assets
Some automotive companies will do well out of the transition - particularly those that lead on both EV and autonomous driving technology. However, those that lag and rely too much on the volume of cars sales could at best lose a lot of revenue and market share and at worse cease to be. Look at what happened to Nokia, Blackberry and Kodiak when the phone and photography markets went through a similar transition.
It's also likely that autonomous driving software will go the same way as other operating systems. There will only be a very few, as only a few very large companies can maintain the required investment and skill level. For example, in the computer market, you have Windows, MacOS and Linux, in the mobile market iOS and Android. One or two-car manufacturers might have the capability and scale to own their own autonomous driving software and cars - Tesla becomes the Apple of the road transport market. While others use the same core operating system, much as virtually all non-Apple mobiles use Google's Android OS. I suspect this is what Google has in mind with WayMo.
Learn more
Much of the data and conclusions from this article comes from an excellent and very in-depth analysis by RethinkX
Is there a silver lining for car enthusiasts?
Car enthusiasts reading this article will probably be quite horrified at this point. Electric robotaxis are about as far away as you can get from a rumbling V8 you drive yourself! However, I do think there is a silver lining to the EV robotaxi cloud.
The fate of the post-ICE horse tells us about the potential future for the enthusiasts' car. The horse is still around, they aren't used much for everyday transport, rather they are either used for racing or owned by enthusiasts who enjoy riding them.
Car manufacturers that want to sell to private owners will need to make their cars very appealing - look lovely and be great to drive. They will continue to race cars too.
You might have more money to spend on the car you'd love to own - from savings from Transport as a Service for the rest of the family. Alternatively, if it's driving a fun car, not owning that appeals to you then almost certainly companies will offer Transport as a Service plans that allow you to take out a fun drive yourself car when you want.
But what about that V8? Won't that succumb to new ICE sale bans that are happening around the world? Well maybe the EV Robotaxi will help prevent this - The new ICE bans are there to drive rapid adoption in order to reach aggressive CO2 reduction targets and get to net-zero by 2050 to limit the worst effects of climate change. But if EV Robotaxis take off at the rate the analysis suggests governments could find they are way ahead of the transport CO2 targets and synthetic fuels will have become more widespread to meet the needs of aviation and shipping CO2 reduction. Combine these together and maybe governments find room to allow synthetic fuel using ICE for the enthusiasts.
Join In
Comments (8)
Reading this as a petrolhead hurts, but you raise a good point and back it up well. I think the switch is inevitable and at this point I don't really see a massive benefit that'll come from questioning that. My bugbear is given EVs were primarily adopted for environmental reasons, is the switch what truly should be happening?
nah mate this guy is a tesla fanboy .....and they are always wrong with their predictions and try to live in their deity (musk)'s future dystopian world......i mean robotaxis........
sadly this article just makes no sense ....here in india electric cars arent even considered to be an option........evs are just an expensive solution for the rich.......the internal combustion engine has given freedom of movement to the people at a low cost.....electric will never replace an ice........there has been a major vote of confidence among all manufacteres that the future is not all electric but ECLECTIC....hydrogen/synthetic fuels/electric all will become an option........and not to mention synthetic fuels make an ice cleaner than an ev on emmisions:-
What’s more, a full life cycle analysis conducted by the Institute of Mechanical Engineers last year showed that a BEV powered by renewable energy would emit 58g/km across its whole lifespan (including mining raw materials for the batteries, eventual disposal of the battery and so on). That’s versus the 45g/km for a fully sustainable petrol-powered internal combustion engine car.............................................................................source:www.formula1.com/en/latest/article.pat-symonds-on-how-formula-1-are-creating-the-next-generation-of-100.6XCGNQ3ExMhbhYy338Qgi2.html
The article is based on the projection that EVs become less expensive to buy than an equivalent ICE ~2026/7. They are already cheaper to run. I would imagine India to take longer to transition, but I do see it's making a start:...
Read moresorry to say but the energy density of synthetic fuels will trump over the efficiency of evs....people will fight for the case of synthetic fuels as it has the existing infrastructure of fuel stations and also doesnt necessarily need any...
Read moreGood article,but one thing that I don't get is, we compare the transition from ICE to EV to that of horses to cars,but ICE and EVs all fall under one category:cars. So I think to compare EV transition with that of when long ago we were transitioning from Horses to cars,is abit absurd.
I agree the comparison has its limitations. Certainly, if it was a straight ICE -> EV transition that would be less of a change, but add in the probably more disruptive autonomous driving enabled Transport as a Service and the change is much...
Read moreOh so basically autonomous driving leads to the comparison...........ok I'll take it.