Tuesday, October 17, 2017

Brookings: Autonomous Vehicle Investments Hit $80 Billion

From the Brookings Institution, October 16:

Gauging investment in self-driving cars
A CNN story late last year declared 2016 “a tipping point for excitement in self-driving cars.” In fact, the year saw reports of major investments by Ford, Mercedes, Apple, Intel, Delphi Automotive, and venture capitalists, among numerous players jockeying for position in the emerging market for autonomous vehicles. There may be debate about the time it will take the technology to adapt enough for wide deployment, but there is consensus that this time will come in a matter of years rather than decades.

To date, there are no public estimates of how large this surge in investment in autonomous vehicle technology is in the aggregate. PWC’s 2016 Connected Car Study, its fourth annual report on the sector, says the top five original equipment manufacturers spent $46 billion in research and development in 2015, and there are numerous reports that catalogue investments, acquisitions, and other activities in the growing ecosystem that supports self-driving cars. We set out to estimate the aggregate investment across this entire area.

The inquiry provides a useful window onto the state of play in development of autonomous vehicle technology, as entities from the major auto manufacturers to startups scramble to take the lead. It is also a measure of what it takes to reach a tipping point in development of sophisticated artificial (or augmented) intelligence. As Qi Lu, the star Microsoft engineer who moved to Beijing to become Baidu’s chief operation officer, recently put it, “In autonomous systems, the car is the first major commercial application that is going to land.”

Investment in self-driving cars appears to be the leading edge for AI development. Indeed, as progress accelerates on vehicles interest is increasing in broader artificial intelligence. A Forrester Research report predicted that 2017 will the year that “Artificial intelligence (AKA cognitive computing) technologies will be rapidly assimilated into analytics practices,” driving new insights from big data, and investment in artificial intelligence has been touted by popular investor news sites like The Motley Fool and a column on Fortune.com. If 2016 was a tipping point for investment in self-driving cars, 2017 could turn out to be a tipping point for investment in artificial intelligence more generally.

The Autonomous Vehicle Landscape 
The ecosystem for self-driving cars has numerous layers. It encompasses not just the machine-learning that operates the vehicles, but also the array of sensor and navigation technology needed, from refinements to the advanced braking or lane-keeping assistance common in today’s vehicles, advances in more granular and adaptive mapping, and vehicle-to-vehicle and vehicle-to-infrastructure communications systems.

Because lives and safety are at stake, autonomous vehicles require a high degree of reliability and need to adapt to widely diverse contexts. In Austin, for example, one of Google’s autonomous vehicles was befuddled by a cyclist rocking back and forth in a track stand at a red light and, in Australia, Volvo engineers found it a challenge to predict the bounding of kangaroos. As a result, in order to train self-driving artificial intelligence systems, the players involved are stretching to accumulate vehicle miles and investing untold hours of engineering time.

Every major car manufacturer in the world wants to be an early mover – or at least to avoid competitive disadvantage. They are joined by automotive suppliers like Bosch and Delphi Automotive. The technology-intensive aspect has drawn in tech companies that see their core competencies involved; Google’s self-driving cars have been the poster children of the sector for several years, but Apple, Microsoft, Alibaba, and Baidu are increasingly involved and players such as Intel, NVIDIA, and Qualcomm are investing in making the microprocessors required. Fleet operators are also involved, including rideshare and logistics companies that are likely to be the earliest adopters and have a ready supply of vehicle miles to supply data for machine learning.

In addition to these established players, startups are playing an increasing role: Nutonomy, a Massachusetts outgrowth of MIT, received a $20 million in investments investment and many of the investments by established players have involved the acquisitions or stakes in startups, such as Ford’s $1 billion investment in ArgoAI. This startup activity has attracted the attention of venture capitalists; venture leader Andreesen Horowitz announced investments in advanced mapping, driver assistance systems, and delivery robots, and several funds have been set up to invest in autonomous vehicle technologies. The venture firm Comet Labs mapped startups working on pieces of the autonomy puzzle and found some 263 firms involved. ...MUCH MORE