#1: Fully self-driving ride-hailing
What you’ll see: Ride-hailing services powered by self-driving vehicles without safety drivers. After 10 years and more than $1.1 billion spent on building infrastructure and core technology, Waymo is finally rolling out pilots in Phoenix and San Francisco. More self-driving operators will follow.
What it means: If successful, this achievement will mark as big of a step forward for humankind as the transition from horses and buggies to automobiles. It will mean technical validation and vindication for the self-driving industry and investors, who have committed more than $80 billion to bring this technology to market — Waymo’s expected valuation is $175 billion (more than Uber and Lyft combined), and this is before they have generated any meaningful revenue. Note that there are likely to be hiccups along the way, with starts and stops, and years before meaningful scale in the US. To some extent, the journey is just starting.
Following successful pilots, there will be a land grab to scale and finalize key infrastructure and technology. Scaling HD map-making and addressing geographical differences — that is, automating the cars to work in Tokyo and Paris as well as they do in Phoenix and San Francisco — will take time, talent, and lots of money. Regulators will need to step up and move faster and companies will need to find over $1 trillion of capital to get the first 10 million self-driving vehicles on the road.
#2: China leading the world
What you’ll see: Hundreds, then thousands of live, fully autonomous vehicles all over China, while the rest of the world continues to grapple with regulatory hurdles.
East Asia, led by China and Singapore, rivals Silicon Valley’s appetite for self-driving vehicles. And despite starting several years after their US counterparts, Chinese companies are closing the technology gap with each day that passes. Regulatory advantages, smart infrastructure, a systems approach, and progressive consumers create a clear path towards scale, despite the unit economics not being as favorable for China given a much cheaper cost of labor.
Baidu has attracted dozens of F500 companies onto its open platform (in the space of one week, it announced deals with Volvo, Ford, and China’s First Auto Works); Tencent is building a team of self-driving experts; SAIC received first batches of licenses to test their self-driving vehicles on public roads; both Singapore and China are building towns from scratch with infrastructure dedicated to self-driving vehicles. Since 2012, the Chinese government (with a budget of $965 billion) has been working with industry stakeholders to develop the world’s longest and most utilized high-speed railway network.
By 2030, China plans to have 30 million self-driving vehicles on the road — more than any other country.
What it means: Self-driving tech in the US will be slower to scale because of unfavorable policies and higher technological barriers due to outdated infrastructure. In order to minimize losses, other governments all over the world need to move faster to adopt policies like China’s around the development of regulations and infrastructure for self-driving vehicles.
However, even if policy is not able to catch up with technological progress, controlling city bodies will likely still purchase and operate millions of self-driving services in order to support economic prosperity.
#3: Consolidation of self-driving companies
What you’ll see: Over the last few years, dozens of companies developing next-generation sensors, HD maps, and core self-driving vehicle technology have launched and raised funding — there are over 70 LiDAR startups globally. But in 2019, the players that didn’t raise enough to scale their efforts quickly and claim a piece of the pie will have difficulty growing enough to stay in the game.
Building up the data and talent needed to support in-house mapping, infrastructure, and core autonomy algorithms is expensive, requiring hundreds of millions of dollars of capital. Smaller players will get swallowed by larger efforts or refocus on smaller deployments.
Some companies are already experiencing the crunch and others are still seeking more cash — Cruise raised $2 billion from Honda and another $2.25 billion from Softbank in 2018. Volkswagen tried to buy a $13 billion stake in Waymo. Delphi acquired NuTonomy for $450 million. Ford invested $1 billion in Argo AI.
What it means: This same rapid growth and subsequent crunch happened in the airline industry, providing precedent for the evolution we expect to see at the supply chain, OEM, and customer levels.
In order to survive, companies developing core self-driving technology without competitive financial backing will need to focus on operating in niche, controlled environments with smaller geo-fences (e.g., campuses, select cities). This also opens up opportunities for companies to focus on the other types of technology that will be key to bringing self-driving vehicles to market — mapping, sensing, etc. technologies — and sell them to larger players.
#4: A total transportation system transformation
What you’ll see: Self-driving vehicles will disrupt all aspects of the transportation industry, including our roads, car ownership, real estate, public transport systems, and existing human-driven ride-hailing networks.
This is already starting to happen — more and more households are going car-free. As self-driving vehicles become more publicly available in 2019, expect to see more of the world’s population participate in the shift.
What it means: Consumers will benefit from safer, cheaper, and faster transportation.
Car accidents and other vehicle-related injuries will become more rare since self-driving vehicles are much better at driving than humans — of the 1.3 million fatalities from car accidents each year, 94% of them are due to human error.
Although the path towards a zero-fatality future will include rough patches along the way (e.g., Uber’s crash in Arizona, Tesla’s autopilot fatality), more self-driving vehicles on the road means more safe rides and more efficiency. The industry will need to use a safety-first approach to testing in order to convince the public of the benefits of coordinated self-driving vehicles that do not get distracted by phones or food.
The price of a trip will drop dramatically over the long run because self-driving-powered transportation companies, whether public buses or ride-hailing services, won’t need to compensate drivers and there will be a surplus of competing sensor technology, computing systems, and software. We can expect services like ride-hailing to cost up to ten times less than they do now (imagine getting anywhere in San Francisco for $1.50, or even $0.75 for a pooled trip).
Once the industry figures out a way to allow for real-time communication between all types and modes of transport operating in a given vicinity, we’ll see a steep decline in traffic congestion. Traffic control centers will manage the flow of cars and other vehicles with efficient coordinated algorithms in a centralized system.
#5: Severe growing pains while scaling
What you’ll see: As multiple self-driving fleets each made up of tens of thousands of units come into the world at the same time, things will get messy.
Autonomous vehicle companies will encounter a new series of problems: How will the cars know where to go? How will fleets coordinate dispatching? Unfortunately, the maps and ride-hailing services available today are only equipped to power human-driven vehicles — Waze relies on drivers following its navigation instructions and Uber dispatches drivers according to human-centric standards.
While companies and regulators figure out these crucial details, we’ll continue to see congestion and accidents.
What it means: Maps and ride-hailing services need to be reinvented for self-driving vehicles, which have different strengths and limitations than human drivers — e.g., self-driving vehicles may not initially be able to safely make unprotected left turns during certain times of day, drive in certain high-density pedestrian areas or around construction, or operate safely on difficult terrain or in harsh weather conditions. Safety-critical data needs to be shared from one fleet to another for the safety of passengers and other motorists.
Routing systems will need to change fundamentally to accommodate the capabilities of robot cars and dispatch services will need new practices and algorithms to create efficient marketplaces of riders and autonomous vehicles. We’ll need a next-gen traffic control system to fully realize the benefits of self-driving at scale.
