This is the second article in our series of three on the Clean Transportation Revolution. In the first one, we took a close look at electric vehicles. In this second one, we will focus on autonomous vehicles (AVs). Driverless cars are coming. The first models could be available for purchase within two years and some predict that by 2030, 95% of passenger miles travelled in the US could be in on-demand autonomous vehicles. It is very likely that children born this year will probably never drive a car.
Billions are being invested by companies across the value chain, from traditional automakers to innovative software start-ups, not to mention the tech giants. This revolution may be slower than the most enthusiastic commentators are saying but it will happen, driven by two clear benefits: greater safety and mobility for all.
AVs will be electric, so they will participate in decarbonising transport and reducing air pollution. However, will they make cities more liveable, or on the contrary will they lead to ever greater congestion? This will depend on their usage model, in shared fleets or individually owned. We will come back to this key question in our third and last article.
In any case, this is an enormous technological revolution, with fantastic investment opportunities for daring angels – in software for driverless cars or fleet management but also in more immediate connected cars technologies which can become AV plays in the longer term, and in the myriad sectors which will be disrupted by AVs, from parking or public transport to retail or entertainment.
The five shades of autonomy
Autonomy is not black or white: levels of autonomy have been classified from zero (no automation at all) to five (full automation), and many vehicles on the roads today qualify at least for level two – such as Tesla’s autopilot function.
Some in the industry see autonomy progressing in a continuum, with increasingly autonomous generations of vehicles ultimately reaching level five after passing the previous levels. Some others focus on the development of fully automated vehicles in one step. In this article, we are discussing the ultimate outcome: the prospects for fully automated cars.
Description and comparison of automated vehicle functionality levels
Autonomous vehicles have arrived
For most of us, driverless cars remain, at best, a science fiction concept. However, in reality, they already exist.
- Autonomous vehicles developed by a number of different firms have already safely driven millions of miles over the past few years. In particular, vehicles tested by Google’s Waymo have driven three million miles and now clock up 25,000 miles every week.
- Autonomous shuttles by companies such as Navya or Local Motors are operating in a number of sites, including public areas in cities in Switzerland, Germany, France and the US.
- Finally, driverless taxis are been tested on the streets of Pittsburgh by Uber and Volvo, and of Singapore by Nutonomy.
Just two more years to wait
If we are to believe manufacturers, driverless cars will be available for purchase within two years.
- All the main manufacturers have an AV programme and have announced dates: the first fully autonomous cars are due in 2017-2018 (Tesla, Volvo), and many will more come in the following years: Audi in 2020, GM in the early 2020s, Ford, Daimler and BMW in 2021, Hyundai in 2022, Honda by 2025, Nissan-Renault before 2027, etc.
- Apple’s CEO Tim Cook has recently confirmed that the company is working on driverless car technology.
- In the UK, beyond today’s relatively modest experiments with small cars, delivery vans and minibuses, a group named ‘Driven’ (led by Oxbotica, a startup making software for driverless cars) is due to test AVs on roads and motorways in 2019.
It is very likely that children born this year will probably never drive a car.
Mass-market acceleration expected between 2025 and 2030
We have compiled forecasts from different sources: according to experts, we should expect a rapid expansion of the driverless car market from the second half of the 2020s:
- The first 10m self-driving cars could be on the road by 2020. This is a large number, which seems too ambitious to a number of experts – and in any case, AVs would still be a tiny portion of the 1.4bn vehicles on the roads globally in 2020;
- However, looking further to 2025, there could be 30m AVs sold in that year. This would represent approximately 30% of vehicles sold globally, hence a massive take off;
- Then, by 2027-2028, virtually all cars sold could be autonomous. This bold vision is notably backed by Elon Musk: “almost all cars produced will be autonomous in 10 years”;
- Longer term, the range of estimates widens: some expect that a quarter of all cars will be self-driving by 2030, or that 25% of miles driven in the US could be in shared self-driving electric cars. The most bullish analysts model that by 2030, 95% of passenger miles travelled in the US could be in on-demand autonomous vehicles.
What are the sceptics saying?
A number of serious sceptics do not believe that fully autonomous vehicles can be available in the next ten years.
Unsurprisingly, an important stumbling block is software development. The challenge is for vehicles to be able to operate without restrictions in any environment, facing a virtually unlimited number of different ‘use cases’. For example, in the absence of lane markings or on unpaved roads, the system must be able to guess which areas are appropriate for driving – which can be difficult if the road surface is not different from its surroundings e.g. when covered with snow.
John Leonard, an expert MIT scientist, lists other types of technical ‘corner cases’ such as left hand turns into heavy traffic, or the interpretation of eye and hand gestures – the latter issue pointing to a broader question about how well robotic cars will mix with human-driven cars.
Accidents have happened
To be able to negotiate the plethora of scenarios, the cars’ decision-making systems must undergo comprehensive ‘training’, with artificial-intelligence engines been taught to make smart inferences from a necessarily finite set of examples. All this still requires development, testing and validation.
Broader issues with self-driving cars include their cost, as well as the likely regulatory and legal hurdles, with questions such as: who is liable for what and when, what components are required in cars, etc. At the very least, sceptics say, sorting out these questions will take decades rather than years – and they have a point.
Another caveat for AVs is a broader one that can be associated with any automation technology, namely the risk of job destruction. Given the very large number of jobs related to driving, driverless vehicles could put millions out of work, the first in line being taxi, bus and truck drivers.
Still it is likely that, at some point, widespread adoption of AVs will happen. Indeed, AVs are expected to bring massive benefits to society as a whole, the two greatest being: a step change in road safety, and the potential to deliver affordable mobility to everyone.
Benefit #1: Revolution in Road Safety
Even though it may be a very remote idea for the public today, a key driver for the adoption of AVs should be a massive improvement in road safety – for all users: drivers but also cyclers and pedestrians.
What are the facts today?
- Globally, nearly 1.3 million people die in road crashes each year, and 20-50 million more are injured or disabled. Road crashes are the leading cause of death among young people ages 15-29, and the second leading cause of death among young people ages 5-14.
- We know that more than 90% of car accidents are caused by human errors.
- Autonomous vehicles eliminate human error and are already safer than human drivers.
- A recent analysis by Rocky Mountain Institute showed that Google’s Waymo had fault in only one accident and, when comparing this to the more than two million miles driven by their cars, this made them by far the lowest at-fault rate of any driver class on US roads— about 10 times lower than the 60-69 year-olds, the safest human drivers.
In fact, this impressive result is no wonder, since AVs have superhuman senses. They can see hundreds of yards around the vehicle in all directions; they use radar to see a slowdown three cars ahead of them, infrared to detect a pedestrian on a dark night, and lasers to map an incredibly detailed view around; and they learn from one another in real time, which means that when one car sees a new pothole or construction detour, the whole fleet knows — immediately.
Benefit #2: Affordable transportation to everyone
Available anywhere, anytime with a tap on a smartphone, self-driving cars can offer convenient door-to-door transportation that is as fast as driving yourself for the price of taking the bus. Such rides can be available to all people, in all neighbourhoods, any day, anytime. That means affordable transportation options in places where public transport is not available and freedom of mobility for people who do not drive including the youngsters, elderly and disabled.
Good for the environment? That remains to be seen!
Contrary to what we sometimes read, the autonomous cars revolution will not, on its own, reduce CO2 emissions or pollution in cities. If AVs do help the environment, it will be not thanks to their autonomy, but simply because they will be electric vehicles – see our previous article on EVs.
The key new benefit of AVs in cities is their very real potential to reduce congestion – even assuming the same traffic. Indeed, self-driving cars will communicate with one another, optimise routes not one by one but ‘as a team’ and hence reduce traffic jams.
Vehicle-to-vehicle and vehicle-to-infrastructure communication to reduce congestion
However, it is also easy to imagine a scenario whereby the move to AVs could make things worse rather than better, compared to human-driven EVs, in terms both of CO2 emissions and congestion.
For example, if we all replaced our cars with AVs, the number of miles driven could increase exponentially – due to AVs’ convenience: more people could start using cars instead of buses or trains; some could send their AVs on every day errands rather than walk; many could accept longer commutes, since instead of having to drive themselves they could use the time in their car for productive things… or entertainment.
In such a scenario, AVs would lead to a skyrocketing in energy use, and potentially to less liveable cities, with more rather than less traffic.
The key decision factor here will be how AVs will be adopted in real life: as individual cars, or as fleets of robotic taxis – as described notably by Robyn Case, co-founder of ZipCar. As she puts it, we should focus on FAVES, or Fleets of Autonomous Vehicles that are Electric and Shared. This will be the subject of our third article, on Mobility-as-a-Service.
How to invest in the AV revolution?
- First obviously, a number of companies are involved in building the self-driving vehicles. This includes large car manufacturers – some of whom will win whilst others will be disrupted – but these are not the type of investment that Green Angel Syndicate contemplates. Then, there are the start-ups building their own autonomous vehicles. However, as detailed above, a number of those are already well advanced and funded.
- A key component is software, an area where a lot of value still has to be added. We expect many more investment opportunities in this field, although it is likely to be a crowded domain for companies and a hotly contested space for investors.
- Slightly off the track, we would expect interesting transition plays in the connected cars area, in other words opportunities to back companies developing tools that can be useful in today’s human-driven cars, but will then have a path towards adding value in the autonomous world. Examples can include devices or software helping us to become better drivers by collecting and analysing driving data, or fleet management software.
However, the biggest opportunities linked to self-driving cars may be in the myriad ways in which they will transform the rest of the world – from parking, public transport and charging to real estate, retail, entertainment and insurance. In many of these areas, new companies with new technologies and business model will emerge and at Green Angel Syndicate we will back the exciting ones which will make a positive impact on the environment.