Juniper Research: taxi sector to lead self-driving market to >22M consumer vehicles on the road by 2025
23 November 2016
New findings from Juniper Research project that the annual production of self-driving cars will reach 14.5 million in 2025, up significantly from only a few thousands in 2020, resulting in a global installed base of more than 22 million consumer vehicles by 2025.
The new research, Autonomous Vehicles & ADAS: Adoption, Regulation & Business Models 2016-2025, found that the market adoption of AV (Autonomous Vehicle) technology is set to accelerate over the next few years, driven by increasingly stringent vehicle safety specifications; environmental pressures; and rapid technological developments.
The research suggests that driverless vehicles will have a disruptive impact on transportation around the world and will ultimately lead to millions of professional drivers being made redundant. Juniper predicts that city-based taxi services will be one of the key early adopters of driverless vehicles.
The introduction of driverless cars will result in fundamental changes to the automotive world and society in general; and it is clear that the boundaries between private vehicle ownership, car sharing and rental fleets will increasingly become blurred.
—research author Gareth Owen
However, the research cautioned that following the first fatality in an AV vehicle—the recent Tesla S accident in Florida—the AV industry must convince the public that their vehicles are completely safe.
Juniper found that a number of major OEMs including BMW, Toyota and GM are accelerating their AV development and testing programmes and now have firm plans to launch production vehicles. As a result, Juniper forecasts that driverless vehicles will start to become widespread in the 2020-2025 timeframe, although they will initially be confined to city centers or key routes due to the need for extensive V2X (Vehicle-To-Everything) infrastructure.
Juniper believes that Level 4 vehicles will not be deployed in any significant numbers until well into the second half of the forecast period (2020-2025), although some OEMs expect to launch some vehicles before 2020.
Juniper expects that the Far East & China region will either lead, or at least be on a par, with North America and West Europe in terms of the number of Level 4 vehicles deployed at the end of 2020.
Juniper expects the Far East & China region to extend its market position thereafter so that by the end of 2025, it will dominate the market with a market share of 34%, followed by North America and West Europe.
Unsurprisingly, the lowest adoption rates will be seen in the less affluent regions of the Indian Subcontinent, Rest of Asia Pacific and Africa & Middle East.
ABI Research: OEMs Target 2021 for rollout of Levels 4 and 5 of autonomous driving. A separate study by ABI Research projects that semi-autonomous systems will continue to dominate the market over the next decade, with SAE level 2 and 3 systems accounting for 86% of autonomous vehicles shipping in 2026. Higher levels of autonomy will gain traction quickly, representing just under one-third of autonomous vehicles shipping in 2030.
Driverless cars will transform the way mobility is consumed, bringing environmental, societal and convenience advantages to the end user. It also represents a fundamental disruption to the business model that dominated the automotive market for almost one century. OEMs have much to gain in pursuing semi-autonomous operation, maintaining the importance of the driving experience. However, recent announcements from BMW, Ford, Renault-Nissan, and Tesla signal that OEMs are not only looking to introduce higher levels of autonomy by 2021, but are also actively planning to transition from vehicle sellers to mobility providers.
—James Hodgson, Industry Analyst at ABI Research
Both Ford and Renault-Nissan launched smart mobility divisions to build on the existing trend of OEM/rideshare partnerships and investments. The divisions also provide a platform for these brands to research and implement autonomous and connected technologies.
Meanwhile, the recently announced Tesla Network details how the brand intends to facilitate peer-to-peer autonomous car sharing, and how participation will impact consumer car ownership costs. ABI Research finds Tesla’s decision to withdraw the level 2 Autopilot system in favor of Enhanced Autopilot, and eventually deep learning-based autonomous functionality, consistent with SAE level 4, or even level 5, to be the most concrete example of the shift in industry attitudes toward low level semi-autonomous driving.
The spread of low-speed Traffic Jam Assist systems to more of the mass market, in tandem with the increasing combination of longitudinal and lateral assistance on highways, will see semi-autonomous vehicles retaining their dominant market share for some years. The quasi-universal 2021 target date for the rollout of more highly automated system nonetheless represents a significant acceleration in the autonomous technology market.
—James Hodgson
Resources
On Track with Self-Driving Vehicles 2.0 (Juniper Research whitepaper)
They have obviously not read the news that Tesla is already selling over 8000 fully self-driving capable vehicles per month.
Order one here fx https://www.tesla.com/models
And also see the self-driving Tesla video if you don’t believe it.
By dec. 31. 2016 there will be 20.000 fully self-driving capable Tesla vehicles on the road.
By dec. 31. 2017 there will be 170.000 fully self-driving capable Tesla vehicles on the road.
By dec. 31. 2018 there will be 670.000 fully self-driving Tesla vehicles on the road most of which will be used on the Tesla Network for further revenue generation.
By dec. 31. 2019 there will be about 1.300.000 fully self-driving Tesla vehicles on the road most of which will be used on the Tesla Network for further revenue generation.
By dec. 31. 2020 there will be about 2.300.000 fully self-driving Tesla vehicles on the road most of which will be used on the Tesla Network for further revenue generation.
Each Tesla could make 100.000 USD per year in taxi services by driving 100.000 miles for 1 USD per mile.
Consequently, Tesla will be highly profitable and Tesla owners will also make lot of money. I expect investment companies will start buying Tesla cars in 2018 solely in order to make them 100% available for profit making on the Tesla Network.
It is beyond me that there are still people in the auto industry that has not heard the news that fully self-driving cars are already here. The Tesla cars just need a software upload coming by dec. 2017 and a few legislative changes and approvals (coming sometime in 2018) to become an on demand global taxi service.
Posted by: Account Deleted | 23 November 2016 at 02:35 AM
Moving more people individually in ADVs (e-taxis) will increase traffic jams unless people learn to share larger (10 to 15 passenger) AD vehicles.
I hope that the first ADV taxi licenses will be given to larger ADVs with 10+ passenger capabilities to effectively reduce the number of vehicles on our street, roads and highways.
Can this be done is signifiant numbers in most States by 2020 remains to be seen.
Posted by: HarveyD | 23 November 2016 at 11:08 AM
ADVs can either increase or decrease vehicle miles traveled (VMT). It's absolutely clear that they will reduce parking space requirements. If these are also used as first mile and last mile connections to mass transit, it could increase the use of mass transit. However, city planners need to step up and incorporate ADVs. Check out the Bloomberg Aspen initiative which is taking the on.
Posted by: Guy_hall | 23 November 2016 at 11:47 AM
Henrik's forecast seems excessively rosy to me.
Posted by: Brent Jatko | 23 November 2016 at 12:27 PM
IMO in order not to clog streets and increase traffic speed separate infrastructure would be needed. Back to old PRT paradigm. But it's nothing wrong with that since it will specific to light duty self driving vehicles therefore many times lighter and cheaper of existing bridges or subways and providing the same speed as underground train. The advantage of upcoming self driving taxi service against PRT will be capability utilizing roads we have now where no traffic jams expected and carry us to the doorstep. I thing city planners should start planing now both self driving taxi charging and jumping infrastructure.
Posted by: Darius | 23 November 2016 at 05:07 PM
Brent they are based on actual production rate in 2016 and Musks prediction on delivering 500.000 cars in 2018 and 1.0000.000 in 2020. Then you just accumulate and extrapolate. Musk could be too optimistic and unexpected things may happen that slow things down.
One thing I am almost sure about is that Tesla will have a fully self driving software ready by the end of 2017. Look at that video https://www.tesla.com/models
It is almost done. It does not seem to lack much in terms of functionality just some debugging.
Posted by: Account Deleted | 23 November 2016 at 11:29 PM
Darius Tesla does not plan to launch Tesla Network with aid from public charging or other infrastructure. They build it all by themselves. A few years back Tesla tried to set up chargers in cooperation with public money but they demanded that the chargers be build in places where Tesla owners did not use them (like public libraries and museums) so Tesla dropped the idea. Tesla's self-driving cars can function 100% without public infrastructure. However, they need the Tesla cloud services to do it efficiently avoiding traffic jams, pot holes, etc.
Posted by: Account Deleted | 23 November 2016 at 11:44 PM
Uber is available now. It apparently will cost less. In my city 75 cents/mile an you get a human to converse with and that can offer suggestions. Purchase an Uber ride and you support a strong network of job creation. The service is very handy. That would be my first choice as a consumer. The customer base and business model will actually facilitate or develops human relationships and friendship, something we need more of now days. Also, they support ride share, which helps defray costs and traffic. The multi-passenger minivan is popular.
Posted by: Trees | 24 November 2016 at 01:07 AM
75 cents / mile uber is during peak hours. Tesla Network can do it cheaper if they have to. As I said 20 cents/mile is all a fully self driving Model 3 needs to drive profitable as a taxi. Uber will be blown out of the market when Tesla Network boots.
Posted by: Account Deleted | 24 November 2016 at 02:33 AM
Henrik,
City taxi service with self driving cars would require charging inside city and taxi parking as well. I do not imagine not involving public authorities. Currently all Tesla charging infrustructure is intercity.
Posted by: Darius | 24 November 2016 at 06:53 AM
Darius you are right. There will be coordination and talk between Tesla and city planners each time Tesla deploy the Tesla Network to a large city. And as you say the Tesla Network will require supercharger stations inside the large cities. Some of these stations will be built on municipal property and that will require a deal to be made on who pays what and to what extent.
Tesla will also need many more supercharger stations per car to service the Tesla Network than they currently provide per car. When Tesla’s car becomes fully self-driving it opens up a new market for Tesla with people living in cities that have no access to charging where they live. They will use the self-driving feature to have their cars drive out at night to a nearby supercharger station to charge. In other words, these cars will do all of their charging on Tesla’s charging network not just long distance travel that the current network of intercity superchargers take care of. That requires many more chargers per car.
The next 24 months where Tesla will roll out its fully self-driving software and boot the Tesla Network is going to be really interesting to follow. It will be a historic moment for mankind. The world will never be the same after that. Tesla’s success in this area will force the old auto industry to stop doing what they currently do and go all in on self-driving BEVs. Then the real change will start to happen.
Posted by: Account Deleted | 25 November 2016 at 12:34 AM
Henrik,
Teslas are NOT fully self driving, more than 90% of the owners KNOW that they require driver intervention.
Posted by: SJC | 26 November 2016 at 06:58 AM
SJC Tesla's new cars are fully self driving capable. Read for yourself at Tesla online ordering page. They just need the final software to be uploaded and the authorities to approve it. The first will happen at the end of next year and the latter will happen sometime in 2018. All Tesla cars in production since oktober 2016 are fully self-driving capable. See video of cars with fully self-driving beta software here For a video showing all there feature in action on beta software see https://vimeo.com/192179726
and for Tesla ordering page here https://www.tesla.com/modelx/design
SJC you need to pay better attention to the news about Tesla. They are ahead of everyone so any news about them deserve the outmost attention.
Posted by: Account Deleted | 27 November 2016 at 02:06 AM
Self driving CAPABLE!?
What the heck is THAT suppose to mean?
Posted by: SJC | 27 November 2016 at 11:51 AM
Musk just tweeted that the fully self-driving capable Teslas will get their first autopilot software in 3 weeks in time for Christmas and will thereafter receive monthly updates that should add abilities until they reaches full self-driving ability by the end of next year.
Electrek has the story https://electrek.co/2016/11/27/tesla-enhanced-autopilot-release/
Posted by: Account Deleted | 27 November 2016 at 12:26 PM
These fully autonomous driving cars will never exist. It is impossible to drive better than a human and you have to park the car yourself, wake -up this is another investment scam. Only crooked investor devellop these systems to grab subsidies from the state, again like they do for batteries, biofuels , arms, poisons, windmills, security, solar panels. 90% of state money go toward corrupted products, jobs and services. It is also a putch from silicon valley mafia and it gonna cost a lot.
Posted by: gorr | 27 November 2016 at 02:42 PM
Extended range fully ADVs (BEVs and/or FCEVs) will be around and authorized in many places by 2020/2025. TESLA will become ONE of the gang.
Refill and charging facilities could easily be safely installed under city parks in large-medium-small cities and/or under current parking lots, in shopping centers etc.
Underground facilities should be the solution of choice and could be built quickly at little extra cost if any.
Posted by: HarveyD | 29 November 2016 at 09:12 AM
The Vancouver 18.5 Km (16 stations) Canada Line subway has been operating for 5 years without drivers. The Canada Line was built by SNC using Hyundai 300 passenger trains composed of 2 cars each. Frequency is one per 3 minutes and 120,000 passengers. Cost was close to CAN $2B.
The new Montreal + South Shore + Airport + West Island REM line (67 Km & 27 stations) will use the same system without drivers. Frequency will be variable and 180,000 passengers. Cost will be close to CAN $6B.
It is difficult to understand why other subways (Metro) have drivers? They create more avoidable problems.
Posted by: HarveyD | 30 November 2016 at 10:17 AM