ReThinking Transportation: Summary
Costs and Speed of Adoption
Transport as a Service (TaaS)
The convergence of Electric Vehicles (EVs) with autonomous technology (AVs) and ride hailing will create an entirely new form of transportation; Transport-as-a-Service (or robo-taxis) which has the potential to transform our road transportation system. It will lead to the end of car ownership, as people access cars when needed, with no upfront costs, and pay a charge per mile (or a subscription), vastly lower than the cost of driving their own vehicle. These cars will be fleet owned and will be in use far more than vehicles currently – travelling 10x the distance of an individually-owned car each year (100,000 miles instead of 10,000 miles). The transition will happen far faster than expected, driven mainly by economics.
The key to understanding the cost improvements of TaaS vehicles is understanding vehicle degradation which affects both maintenance costs and vehicle lifetime (in miles). The main causes of vehicle degradation are all vastly reduced in an Autonomous Electric Vehicle (A-EV) due to:
- Reduced friction - 20 moving parts in an EV compared to 2,000 moving parts in a gasoline vehicle leads to far less friction and means there is less that can go wrong. EVs are essentially batteries on wheels.
- Reduced heat and vibration - the simplicity of the powertrain dramatically reduces heat and vibration in A-EVs, which reduces wear and tear on the vehicles.
- Reduced time-based degradation - under the TaaS model, cars no longer sit idle for 96% of the time (as they do in private ownership), but are instead in operation 10 times more, covering 100,000 miles per year instead of 10,000. This reduces degradation like corrosion or partial battery degradation (driven by car lifetime in YEARS).
Vehicles will last much longer in terms of vehicle miles (500,000-1M miles) and cost just a fraction to repair (conservatively estimated at 20% of gasoline vehicle repair costs).
Only TaaS will benefit from the increased lifetimes. In individual ownership, a 1 million mile EV would last 100 years. Given the increasing pace of fleet turnover, vehicles will be obsolete quickly in the individual ownership model, and so depreciation (or lease) payments will still be based on residual value calculations.
But in TaaS, the capacity to travel 1 million miles combined with fleet ownership, no secondary market, and high utilization means that depreciation will be calculated by spreading the costs over the lifetime miles evenly, leading to dramatic savings: each mile travelled by TaaS will cost just 1/1,000,000th of the upfront cost.
The upfront cost of an EV compared to a gasoline vehicle is the key focus of commentators, but it has far less impact on cost per mile than the improvement in lifetime miles.
- Insurance costs: 90% reduction as AI will be dramatically safer than human drivers and cars will be impossible to steal.
- Fuel costs: 70% reduction as electricity is cheaper than gasoline per unit of energy and EV motors are far more efficient than combustion engines.
- Finance costs: 90% cheaper per mile as the fixed annual interest cost is spread over 10 times as many miles.
Vehicle manufacturers will compete on cost per mile (rather than upfront cost of purchase or other attributes like performance, comfort, appearance, etc.). Currently, manufacturers make money from selling vehicles and maintaining them, creating incentives towards planned obsolescence.
In the future, they will compete to maximize vehicle lifetimes and minimize operating costs. Further gains are possible as vehicles are re-designed to strip out unnecessary accessories and platforms are created that can be easily maintained, with all parts capable of simple replacement when required. Labor in maintenance will be minimized through automation.
All together, these changes will deliver transport by TaaS or TaaS Pool at a cost per mile that is four-to-10 times cheaper than purchasing a new car AND two-to-four times cheaper than operating (maintenance, fuel, and insurance) an existing vehicle.
The result of these disruptions will be a much faster transition than is currently perceived by mainstream analysis, which focuses not on the disruption of car ownership, but only on the technology disruption of individually owned EVs replacing one-for-one ICE vehicles when they are purchased new – which leads to a multi-decadal replacement of the existing vehicle fleet. Existing owners will abandon their vehicles, stranding up to 100 million gasoline and electric vehicles in the United States.
This decrease in cost opens the way to free transportation (initially in the TaaS Pool model), where the low cost of transport (1-2 cents per mile) is covered by other revenue sources – from advertising, data monetization, entertainment, or product sales. Other revenue could come from grid stabilization and demand management, corporate sponsorship of fleets or government subsidies to provide access to transport. Indeed, the success of the winners might be defined by their ability to exploit these new opportunities more than their ability to produce vehicles.
All things being equal, one would expect that the scale of this cost reduction would lead to far faster adoption than is currently expected. However, “all else” is NOT equal. Once adoption of TaaS begins, ‘systems dynamics’ will drive it ever faster.
Given this new economic analysis, how quickly will the disruption occur? To understand this, we need to consider the drivers of decisions and actions by consumers (demand), businesses (supply), and policy makers (the regulatory environment). We also need to understand the non-linear impact of the accelerators and brakes that occur in technology disruptions – these are the feedback loops that make TaaS ever cheaper and better to use over time and individually owned ICE vehicles more expensive and harder to use. These are the network effects, tipping points, and market forces that change dynamically over time.
Consumers will be motivated by cost savings above all else – which amount to 10% of total household income (a minimum of $5,600 per year). The scale of the savings will override other factors like love of driving or ownership.
The value of additional free time to work or do other things will support this. Pent-up demand for TaaS from those excluded from transport (the elderly, disabled, or poor), the replacement of current taxis and ride-hailing networks, and early adopters will ensure TaaS demand gets a fast start.
Businesses will be motivated to seize market share by flooding local markets with vehicles, given the winners-take-all dynamic of the network effect for TaaS fleets (the more vehicles, the better the availability, the more customers), the size of the potential market, and the existential risk to their existing business.
This will ensure vehicle supply meets demand. Businesses will face a competitive market environment where price trends towards cost, meaning consumers benefit as cost savings are passed on in lower prices.
Policy makers will adopt a supportive approach to regulation, driven by the economic gains ($1 trillion in consumer disposable income and $1 trillion in increased GDP from time freed from driving), lower infrastructure spending (from reduction in vehicle numbers as one vehicle does the work of 10+), and a land bonanza as valuable urban land freed from roads and parking requirements becomes available.
Health benefits from cleaner air and fewer accidents, and environmental gains from vastly lower GHG emissions, will spur this supportive environment. The countries and regions that lead will benefit from gains in jobs, revenues, and influence that come with technology leadership driving a competitive regulatory framework in which regions compete to lead the disruption. Incumbent push back from oil companies or other losers will be offset by support from the vastly richer Silicon Valley. Some countries might resist but developments elsewhere will mean they are forced to play catch up.
TaaS will get a fast start in cities and then radiate outwards to suburban and rural areas. As adoption increases, a tipping point in availability will be reached early, when it becomes feasible for car owners to rely entirely on TaaS. As used car supplies increase and demand decreases, values will plummet. Those steadfastly sticking to individual ownership will be far more likely to buy used cars, driving down new car purchases dramatically and leading to a death spiral of reversing economies of scale, increasing costs and leading to supply chain breakdown.
Over time, another tipping point is reached when gas stations begin to close down, spare parts disappear, and maintenance becomes harder or costlier to find. ICE vehicles will become harder and more expensive to use. Consumer opinion moves from being a brake (scared of autonomous vehicles) to being an accelerator (seeing human drivers as reckless, dangerous, and slow), driving regulation that bans or restricts human driven vehicles.
Barriers perceived to slow adoption of EVs, like range or charging availability, are not relevant to TaaS trips, where cars will have sufficient charge for trips ordered, and car relays, battery swapping, and fast charging will enable longer trips.
We see no supply, demand, or policy constraints, meaning that within 10 years of regulatory approval of TaaS, the vast majority of passenger miles travelled (95%) in the U.S. will be with TaaS. Even in the rural areas, where wait times and cost (due to redundant trips) might be higher, lower incomes will motivate adoption, as will the increasing cost and difficulty of using a gasoline vehicle. The only hold-outs might be the very rich, the most rural, or the equivalent of those who still use horses for transport.
The key enabler of TaaS is autonomous technology. Its availability is based on both technological progress and regulation. Given our analysis of both, we have high confidence that it will begin in 2020-25, with 2021 as the most likely date. Even if the U.S. adopts a precautionary approach, other areas (China, Singapore, European cities) will push ahead, meaning the technology reaches readiness in this time frame.
AVs learn by doing, so the more cars on the road and the more miles they cover, the faster they will be ready. Technologies that allow cars to benefit from the learning expected from 100 million miles in just 1,000 miles are in development, offering a faster track to deployment. This means that by 2030, the disruption will be almost complete.
Regulation can drive this process even faster by hastening the development of AV technology through broad trials, removing barriers and accelerating adoption by ensuring universal access, (the utility model), special lanes or routes for AVs to allow faster travel and greater volume of traffic, or through restrictions on humans drivers in cities or during peak times.