savs, shared autonomous vehicles, SAVs, Google, automakers, ford, gm, personal vehicles, cost, industrial innovation, arkq, arkindu, ark invest, ark investment management

Shared Autonomous Vehicles: A Multi-Hundred-Billion Dollar Opportunity for Google

UPDATE (04/26/16): Today, most automakers are investing heavily in autonomous technology and sharing platforms, anticipating what is, in ARK’s view, an inevitable shift to shared autonomous vehicles ( SAVs ). By way of contrast, in 2012 when a GoogleGOOG team met with various automakers to discuss partnerships to build self-driving cars, the automakers felt that Google was “speaking a different language,” and disagreed on almost every issue.

Since then, many original equipment manufacturers (OEMs) have accelerated their efforts to adapt and invest in autonomous technology. Recently, ToyotaTM, FordF, and VolkswagenVLKAY each gave $300,000 to fund artificial intelligence research at the University of California Berkeley.1 General MotorsGM has been particularly aggressive this year with its $500M investment in Lyft, its acquisition of the assets of Sidecar, a failed ridesharing service, and its $1B acquisition of Cruise Automation, a startup that develops autonomous driving kits.

The graphic below2 illustrates why this may be happening. According to ARK’s research, a SAVs network could generate more than $600 billion in annual US revenue, depending on the price of autonomous travel.3

shared autonomous vehicles, SAVs, Cost of owning a car, autonomous cars

In the US, personal vehicle travel costs 70₵ per mile,4 totaling $2.1 trillion dollars per year. If Google SAVs were to account for 60% of miles driven in the US, and were priced at 35₵ per mile, the market size would be $630 billion.5

Each SAV will obviate the need for nine traditional vehicles,6 a significant threat to new car sales. Not surprisingly, automakers including innovative firms like TeslaTSLA have said that self-driving cars face a “regulatory minefield,” which may be an attempt to influence a slower integration process for autonomous vehicles. Google, however, has been working with lawmakers from the start. Its head of the Self-Driving Car Project, Chris Urmson, has said that he foresees no significant regulatory hurdles for autonomous vehicles, and has a goal of commercialization by 2020.7

Google seems to be having an easier time finding auto-manufacturing partners now than it had in the past. General Motors executives have made numerous comments that they would be open to a Google partnership,8 though rumor has it that Google has been in talks with Ford for some time.9 Ford may be the more likely partner with Google because former Ford CEO Alan Mullaly joined Google’s board in July 2014.

AvisCAR, HertzHTZ, Uber, Lyft, Getaround, and other sharing economy companies also will benefit from the evolving auto market. KPMG predicts that over half of households will only own one car in the next 25 years, compared to today’s average of two cars.10 Further, fewer teens want drivers licenses these days. The percentage of 18 year olds who have a driver’s license fell from 75% in the late 90s to 60% in 2014.11 ARK sees a a bright future for SAVs where many different shared mobility models will emerge, and these statistics show that Americans already are counting on them.

Extensive local knowledge and experience is likely to prove a crucial enabler for fully autonomous cars, presenting the possibility of geographic monopolies. Based on ARK’s research, Google, MobileyeMBLY, and Tesla are the players with the most “training” miles, or miles spent collecting data to improve autonomous driving:12

savs, shared autonomous vehicles, google, tesla, mobileye

Tesla has collected 780 million miles of data from owners using autopilot, or roughly 30 times the number of training miles that Mobileye has under its belt. Traditional automakers using their customers’ cars to collect local information could acquire data at a much faster rate than their technology counterparts.

That said, Google seems to be collecting the most detailed and extensive data per car as it uses LiDAR sensors while other players primarily use cameras. While 2.5 million miles may seem small in comparison to Tesla’s data collection points, Google is collecting more than three times the number of miles most people in the US will drive in a lifetime.13

Aside from its data-detail advantage, Google may have a better relationship with regulators than do its competitors. Apparently, Google has been working closely with the US Secretary of Transportation, Anthony Foxx. Further, the US National Highway Traffic Safety Administration recently responded to a letter from Google asking if an artificial intelligence system could be used in place of a human driver with a steering wheel, pedals and brakes.14 The answer was yes, raising the probability that fully autonomous cars could move ahead in the US without running into significant regulatory traffic jams.

  1.  Source: http://www.bloomberg.com/news/articles/2016-03-16/automakers-go-back-to-school-to-learn-to-build-self-driving-cars?utm_content=31260795&utm_medium=social&utm_source=twitter
  2.  Source: ARK original research. For sources for estimates for the cost of personal and future travel, please see our past blogs, which are linked to in the text above.
  3. Source: A previous version of this article on SAVs included ARK’s prior estimate for autonomous travel at $0.25 per mile, which implied a $450B market opportunity.
  4. Source:Assumes gas costs $2.252 per gallon as of 2/16/15.
  5. Source: Note that this combines ARK’s estimate for autonomous travel in the year 2020 ($0.35/mile) with a portion of today’s miles driven to arrive at $630 billion. 60% of today’s miles driven is roughly all urban miles. Miles driven may increase with autonomous vehicles and SAVs, and any change to the price of autonomous travel or the addressable miles driven will increase or decrease this figure.
  6. Source: As noted in the linked blog, studies show sharing services could increase vehicle utilization by eight- to twelve-fold. Also, ARK research shows that Zipcars have a 34% utilization rate, which is 8.5 times that of personal cars.
  7. Source: http://www.reuters.com/article/2015/01/14/us-autoshow-google-urmson-idUSKBN0KN29820150114
  8. Source: http://www.reuters.com/article/2015/01/12/autoshow-gm-google-idUSL1N0UR2MZ20150112
  9. Source: http://www.theverge.com/2016/1/11/10750524/ford-google-alphabet-self-driving-cars-wsj
  10.  Source: http://www.nbcnews.com/business/economy/wave-goodbye-two-car-family-n251036
  11.  Source: Office of Highway Policy Information; 75% in 1998, 60% in the most recently available datapoint (2014)
  12.  Notes: HERE, a mapping company owned by BMW, Audi, and Daimler, has also likely collected over 2 million miles of mapping data, but was left out of this analysis because this data is not training data. Also, autonomous car expert Brad Templeton has expressed doubts that HERE’s maps are as detailed as players like Google have prepared.  Sources:

    ARK Investment Management LLC, http://360.here.com/2015/03/11/official-everywhere/ , http://ir.mobileye.com/investor-relations/events-and-presentations/default.aspx,  https://static.googleusercontent.com/media/www.google.com/en//selfdrivingcar/files/reports/report-0216.pdf, https://www.technologyreview.com/s/601567/tesla-tests-self-driving-functions-with-secret-updates-to-its-customers-cars/

  13.  The average US 16 year old has 63 more years of life and the average US driver drives 13,500 miles per year. Sources:

    http://www.cdc.gov/nchs/data/nvsr/nvsr64/nvsr64_11.pdf

    http://cars.lovetoknow.com/about-cars/how-many-miles-do-americans-drive-per-year

  14.  Source: http://arstechnica.com/cars/2016/02/googles-self-driving-car-ai-can-be-the-vehicles-legal-driver-us-government-says/


ARK's statements are not an endorsement of any company or a recommendation to buy, sell or hold any security. For a list of all purchases and sales made by ARK for client accounts during the past year that could be considered by the SEC as recommendations, click here. It should not be assumed that recommendations made in the future will be profitable or will equal the performance of the securities in this list. For full disclosures, click here.