That the near future will be one in which driverless cars are a normality is no longer under serious question. Exactly how imminent that future is and what exactly it will look like is still unclear. What is clear is that it is happening and will bring about major changes in our lifestyle and to the economy.
With personal transport and public transport services such a key part of everyday life, the chances are that the advent of driverless cars will not simply be a superficial convenience. It will not be in the same league as robot vacuum cleaners shaving a couple of hours a week off the list of the list of domestic chores. It won’t even be comparable to the impact of the advent of p2p ride hailing services such as Uber and Lyft bringing down the cost of taking a taxi. In fact, driverless cars will make that new sector obsolete, in its present form, almost as quickly as it arrived.
As with any big technology shift, driverless cars are likely to decimate some parts of the economy and necessitate a possibly painful transition for those who directly and indirectly rely on newly-obsolete industries and occupations. It is hoped the lifestyle benefits and economic efficiencies that result from the hours freed up will be far greater in the long run.
Economic analysts and futurists believe that driverless cars will have an influence on the way we live and the economy that we earn our living from that will rank among the most profound changes in modern history. Time will tell but driverless cars could well be up there with railways, the process of the privately owned car replacing the horse and cart and the advent of affordable international air travel.
Enthusiasts talk of the effective disappearance of the traffic accidents that the WHO estimates kill 1.25 million globally every year. With driverless cars being manufactured as electric vehicles, emissions should be minimised, massively improving the air quality of urban areas, especially in big cities. Driverless cars using AI to optimise routes, car sharing, a huge reduction in urban parking requirements and cars being able to drive closer to each other are all hoped to mean congestion may also largely become a thing of the past.
However, if the impending changes brought about by the mainstreaming of autonomous vehicles will genuinely improve our quality of life long term or not remains to be seen. The Jevons Paradox occurs when efficiencies of new technology spur demand that increases consumption. This leads to an end result of more of a resource actually being used than was the case before the new technology improved efficiency. When cars replaced the horse and carriage in London they were seen as a far faster and cleaner mode of transport. It takes roughly the same amount of time to navigate central London by car today as it did by horse and carriage at the end of the 19th Century and the impact of petrol-fuelled cars on the environment is well known.
The Driverless-Car Market Value
The point is, it’s far from a certainty that we’ll all really have more free time and expendable income as a result of driverless cars. Computers didn’t quite result in a 4-hour working day. What is certain is that it will be a hugely valuable market. Estimates vary but UBS puts the 2030 value of the self-driving technology sector at $2.8 trillion. Intel thinks a more conservative $800 billion, rising quickly to $8 trillion between 2030 and 2050. Goldman Sachs predict the pie for robo-taxis and TaaS (transport-as-service) sector will be worth $285 billion by the same year.
Intel believes by 2050, driverless ride services will be worth $4 trillion and delivery and logistics $3 trillion. It’s also almost certain new use cases will provide significant revenue streams we aren’t yet predicting. Google Maps technology leading to ride hailing apps and now driverless cars is a perfect example of how new use cases develop out of new technology.
Car Makers vs. Tech Specialists
Driverless cars will inevitably change patterns of car ownership and the business model for car sales. This will lead to a huge industry shake-up and the battle for first survival and then dominance is already underway between traditional auto manufacturers.
However, with driverless cars, the ‘brain’ of the vehicle is more important than its body. The efficiency and cost of the array of sensory equipment, AI software and the data that empowers that AI and makes driverless vehicles a possible is what will define the success of the actors vying for supremacy. The consequence of this reality is that many analysts believe it is tech specialists, not car manufacturers, who will dominate the new driverless car economy. Google-parent Alphabet’s Waymo unit, ride hailing app company Uber and Elon Musk’s Tesla, a relatively new car maker focused on electric vehicles and driverless tech, have invested heavily in the power struggle for the driverless car market. Other Big Tech companies such as Amazon and Apple are also in the mix, albeit quietly for now and don’t discount survivors emerging from the start-ups.
Car Manufacturing Know-How vs. Tech Smarts
Success in developing reliable, safe driverless vehicles that are able to secure permission to operate and, even more importantly, the trust of passengers will, at least in the early phase of the market, be far more about the technology system than the manufacturing of the vehicle. While strong early market share will not guarantee companies long term dominance or even survival, it will certainly give them a much better chance of achieving both.
Car manufacturers had a head start in developing some elements of the technology integral to driverless cars.
Cruise control systems of gradually advancing capabilities and automated parking were the first steps towards driverless cars and have been around for a while – particularly the former. However, there is a significant leap to the kind of sensory system, AI software and database required for fully automated autonomous vehicles. That’s the kind of cutting edge tech expertise many industry analysts believe give technology specialists the edge.
It’s a lot easier to buy cars made by a car manufacturing specialist and integrate a driverless system than it is to develop a driverless system. However, car manufacturing specialists should not be discounted from the race.
There is an argument that building a vehicle from the ground up to include an autonomous sensor set allows for a better handle on how production should look and how everything can be optimised. Traditional car makers already have efficient production lines and if they can develop an operating sensory system and software, that could be a winning combination.
That, however, is a big if. It is forecast that a few will manage to achieve this combination on time and many others will not. Those who don’t will be forced to either strike strong partnerships with tech partners who have the system or face obscurity. Analysts believe that the fractured international car manufacturing market will see many casualties and consolidation. Another trend predicted is car manufacturers forced to drop their own brands to produce white label driverless fleets for tech companies and ride-hailing TaaS brands.
Datasets Are Key
It is widely accepted that at this stage, the resource most important to any company with ambitions in the driverless car/automated vehicle race is data. Data is the raw ingredient the AI that will react to the sensory input from the surrounding environment needs to take decisions. The AI takes decisions based on the outcome of tens of thousands to millions of comparable situations recorded in the database. The greater the wealth of the data resource, the lower the statistical chance of an unanticipated outcome. With enough data that chance should, in theory, be completely statistically eliminated even in circumstances of a highly unusual combinations of factors.
Companies developing the driverless car systems they hope will win early approval to operate commercially are in a race to gather that data. They are doing that in two ways – road tests involving Level 4 driverless cars that are already autonomous but still have a driver that can take the controls in an emergency and advanced digital simulations. Some companies have now conducted Level 5 road tests in controlled zones along strictly defined routes.
Let the Battle Commence
Let’s take a look at how some of the early leaders, dark horses and less-known start-ups in the space are currently shaping up and positioning themselves for the battle ahead. With a $7 trillion prize set to be divided among those who can carve themselves out defendable territory, there is a lot at stake.
Waymo – Alphabet
Widely considered to be setting the early pace, leading the pack with the most aggressive early ambitions for the driverless car economy is Waymo. A unit of Google’s parent company Alphabet, Waymo has both huge financial resources and many of the world’s leading technology experts at its disposal. Crucially, this contender also has a significant head start in terms of the data it has already gathered.
Waymo began working on an end-to-end driverless system at least 5 years before the big car manufacturers and has completed over 5 million miles of Level 4 test driving as well as 5 billion virtually simulated miles.
The tangible result of this head start is that Waymo is also almost certain to be the pacesetter when it comes to beginning to operate a driverless commercial service. A pilot programme for paying customers will begin operation later this year in Phoenix, Arizona, with the intention to quickly expand from there as the necessary permits are secured for other locations. Not only will this allow Waymo to start building its brand as an operational service ahead of rivals, it will also allow it to keep building its invaluable data set, a virtuous circle the company will hope keeps it ahead of the chasing pack.
Waymo has recently placed an order with Jaguar Land Rover for 20,000 I-Pace electric vehicles to be fitted with its system, ramping up real test drive miles and, presumably, to subsequently be used across new commercial operations once confirmed.
For now, Waymo’s approach looks similar to that of how its parent company approached the smartphone market. Android is the dominant smartphone OS worldwide, with Apple’s iOS its only genuine rival. Other pretenders such as Windows mobile, even with the might of Microsoft behind the OS, have fallen by the wayside. Waymo is also a system that will be theoretically compatible with different vehicles and the strategy is presumably for car manufacturers to license it for integration in their own cars, as well as own-brand positioning in the TaaS/robo-taxi market.
The extent of Waymo’s head start is such that a recent UBS report forecast the company will own 60% of the self-driving market by 2030. While nothing is guaranteed nor set in stone, Waymo’s role in this market looks set to mirror that of Android in the smartphone sector.
There is far more debate around which company follows Waymo in the ‘best of the rest’ race. However, capital markets still appear to have confidence in Elon Musk’s Tesla. It’s the only thing that can really explain why the company has a market capitalisation of just under $50 billion companied to Ford’s $44 billion. Ford’s 2017 revenues were $156.7 billion and Tesla’s $11.8.
While it nominally started out as an electric car manufacturer, Tesla’s real goal has always been the driverless car market and the company has been relentlessly pursuing that in the background. It is widely thought that Waymo currently has the superior end-to-end driverless tech but Tesla may just have a secret weapon up its sleeve when it comes to the all-important data set at its disposal.
Tesla already has over 300,000 cars on the road, all over the USA and also in many other parts of the world. Each of those vehicles is fitted with the company’s Autopilot system, a semi-autonomous driving system. The company’s privacy agreement allows Tesla to collect data when its vehicles are using the Autopilot mode. More importantly still, Autopilot also has a ‘shadow mode’. When drivers are not using the software it can still log what is happening and instances when Autopilot would have taken an action.
Last year Musk told IEEE Spectrum that the data gathered added up to 3 million miles a day being logged. As of mid-2017, the Tesla fleet had clocked a combined 5 billion real road miles. Tesla’s real road miles, as logged and gathered by its semi-autonomous Autopilot functionality, activated and in ‘shadow mode’, is more than every other rival test-driving autonomous vehicles combined. Adam Jones, a Morgan Stanley analyst, believes that data set is likely to be more valuable to the company than the cars it currently sells.
The other joker that Tesla hopes to be able to play is the alternative sensory system the company is pursuing. All of the other players in the self-driving car market use LIDAR sensors. Tesla doesn’t. LIDAR sends out millions of laser light signals. It builds a picture of the surrounding environment based on how long it takes for them to bounce back and from what angle. It essentially works in a very similar way to radar but using laser signals rather than radio waves. LIDAR is most important when a car is driving in the dark and cameras are less effective. Radar and ultrasound technologies currently lack the same precision as LIDAR.
All of Tesla’s competitors believe LIDAR is essential to a safe self-driving system but Musk disagrees, calling it a ‘crutch’. Tesla are attempting to develop their system without it. It’s looks like a high risk gambit but if Musk’s team manage to pull it off, they will have a huge advantage. LIDAR is expensive, bulky and involves moving mechanical parts. Taking it out of the equation will hugely reduce the cost of Tesla’s driverless vehicles and allow for sleeker design.
In summary, the strength of Tesla’s position will rely heavily on two unknown factors. Firstly, how much data the company has generated from its existing fleet’s semi-autonomous Autopilot mode and how deep that data is. Secondly, can it pull off a reliable self-driving system that manages to bypass LIDAR?
Of the traditional automakers, US giant GM appears to be furthest ahead in its self-driving vehicle trials. Like Tesla, GM also already has a semi-autonomous product already on the roads. However, its Super Cruise functionality is only built into one model of Cadillac and isn’t thought to send back the same level of data as Tesla’s vehicles.
GM has pole-vaulted itself into the self-driving race through acquisitions. The bought San Francisco-based Cruise Automation for its self-driving system technology and then LIDAR maker Strobe, which makes a far smaller, cheaper and more effective version of the technology than what GM was using before. The LIDAR General Motors was integrating into its Chevrolet Bolt electric cars was costing $30,000. Strobe’s tech reduces that cost to hundreds of dollar per vehicle.
GM’s other strength is that it has a stake in Lyft, the popular P2P taxi-hailing app. That offers it the opportunity of ready-made market share in the robo-taxi sector that is expected to be the most lucrative part of the self-driving market.
In terms of road tests, GM has had a fleet of similar size to Waymo’s out on the streets. Many analysts believe the company has missed a trick in not replicating Tesla’s approach of including semi-autonomous technology in more of the cars it already sells – much more than Tesla. GM’s road tests have seen the highest rate of accidents among the companies test driving – 22 of the 27 involving driverless vehicles in the California and 5 of the 7 so far this year. However, they have mostly been minor, not the fault of the GM vehicles and can also largely be attributed to the fact tests are being conducted on San Francisco’s busy, often narrow, streets. It can be argued this could, in fact, be a long term advantage for GM.
Finally, GM have the manufacturing experience and capacity to quickly produce hundreds of thousands of driverless electric Chevrolet Bolts when the time comes.
Until relatively recently, Uber were considered likely to be one of the early pacesetters in the market. The company already has strong market share of the P2P taxi-hailing business in the US and internationally through the popular Uber app. That’s potentially a powerful springboard.
However, the company has hit a rough patch and seen its license to operate suspended in important markets such as London and Paris as a result of scandals, mismanagement and the traditional taxi industry lobby.
Uber had conducted 2 million miles of real life road tests by last November despite earlier being caught doing test drives with modified semi-autonomous Volvos in San Francisco without having secured the necessary permits. Testing has also been suspended following a fatal incident involving an Uber car in Arizona during March.
While we certainly haven’t heard the last of Uber in the context of the early stages of the driverless cars market, the company’s sometimes overly aggressive tech start-up ‘growth hacking’ approach, may have damaged its prospects by going too far and crossing the line towards reckless. How it bounces back over the coming months will be crucial to long term prospects.
Germany’s automaker giants have recently forged an alliance in their attempts to create a European player in the new autonomous vehicles market. The two companies recently announced a joint venture which will combine their car-sharing and ride-hailing subsidiary companies. Daimler’s Car2Go and BMW’s DriveNow are subscription services that mean users can use pooled cars with Moovel and MyTaxi their respective taxi-hailing apps. All four have been moved to newly-formed and jointly-owned company. There is speculation it could be listed on the Frankfurt Stock Exchange.
Daimler Chief Exec. Dieter Zetsche set a bullish tone when the announcement was made. He made it clear that the German car makers don’t plan on sitting back and letting big tech take over their turf with the dawn of driverless cars:
“As pioneers in automotive engineering, we will not leave the task of shaping future urban mobility to others.” His BMW counterpart, CEO Harald Krüger added “this alliance will make it easier for our customers to discover the emission-free mobility of the future”.
Both companies are conducting extensive road tests of their autonomous vehicle technology in Germany. Navigant Research, a research company that studies emerging technologies and the business models that will shape the future of transport ranked Daimler as third, behind Waymo and GM, ahead of Tesla, in its league table forecasting the companies that will dominate the new sector.
Another tech specialist, Nvidia is a US computer hardware and software specialist at the cutting edge of AI and VR. A dark horse in the race, like Waymo, Nvidia is focused on systems rather than manufacturing vehicles and will look to build market share through partnerships rather than going it alone. Its Drive Constellation simulation technology powers the research of many of the major players in the self-driving space, including Tesla.
Simulation has some limitations compared to real road testing but also strengths. Danny Shapiro, the senior director for automotive at Nvidia, comments:
“..in the real world, you only have a few minutes every day to drive a particular road as the sun goes down. In simulation? “We can drive every road 24 hours a day at sunset, and stage all kinds of [other] potential hazards.”
Nvidia may become the industry standard for self-driving simulation. The problem regulators have when deciding if a company has enough data to justify receiving test-drive and commercial permits is that much of it comes from simulations. And when these simulations don’t adhere to one standard it is difficult to judge their quality. As such, a company that can provide an accepted industry standard will be in a strong position.
Nvidia’s partnerships are already numerous and as well as Tesla include VW and start-up self-driving car manufacturer Aurora.
Apple’s ‘Project Titan’ autonomous vehicle programme is one of the most mysterious and the company is generally not talked about as a likely early leader of the industry. However, write the iPhone maker off from making a late surge at your peril.
Like many of the other tech giants, Apple is working on a technology ‘stack’ that can be integrated into the cars of other manufacturers to give them self-driving capabilities rather than planning on manufacturing vehicles itself. It’s been awarded numerous patents for technology in the autonomous vehicle space and has been test-driving modified Lexus SUVs in California.
Sitting on the largest cash pile of any company in the world, Apple’s most likely play is through the acquisition of one of the most promising start-ups in the space. It has the resources to supercharge the development of any acquisition and catapult itself into the race. And then, why not add one of the more popular taxi-hailing apps into the mix?
Chinese tech giants Baidu, Alibaba and Tencent, collectively referred to as BAT, will also certainly figure and are likely to dominate their domestic market. That could then potentially lead to wider international ambitions. All three are now throwing resources into driverless cars technology and a recent McKinsey report forecast that the Chinese market for autonomous vehicles and TaaS will be the most valuable in the world, worth more than $500 billion by 2030.
Baidu, China’s Google equivalent, is furthest advanced. Like US peer Waymo, it has five years of research and development under its belt and has produce an open-source self-driving software called Apollo. This allows any car manufacturer, or even individual, to use Apollo, test it, collect and share the data. Baidu has a 2021 target to produce Level 4 autonomous cars, that allow the intervention of a driver, in partnership with Chinese carmaker BAIC Group.
While we can see how things are shaping up and make predictions based on the evidence available, how the battle for the self-driving car market will eventually play out can’t be predicted with any real certainty. There are also numerous prospective start-ups in the space having money thrown at them by investors. Most will either fail or do just enough to be gobbled up by the bigger fish, their most valuable innovations incorporated into the tech stacks of the major players. However, as tends to be the case in new markets, a new unicorn may well also rise to forge its own independent kingdom. Whichever way the tides turn, and they are likely to ebb and flow over the coming decades, it will be fascinating to watch this new world order unfold.