Part II: Where in the world will Level 4 Autonomous Vehicles land first?
With pilots underway , AV players are eager to assess which cities will be most open to the future of mobility and large-scale deployment? What does that mean for key AV features?
This is Part II of a 4-part series. Stay tuned for Part III to explore Waymo, Uber and Tesla’s emerging business models, and Part IV to understand how those may affect successful deployment. See Part I for more background on AV fundamentals.
There are some baseline criteria that make a city attractive for AVs. At the AUVSI Summit in March, nuTonomy CEO Karl Iagnemma identified those in order as willpower of state, weather conditions, infrastructure conditions, traffic and driving conditions. Other important indicators include existing public transit (as a complementary service), density, ridesharing uptake and possibly average income. These last two presume fleet ownership, whereby AVs are likely too expensive for individual ownership and thus the cost must be amortized over riders and savings from driver replacement. These also align well with the idea of geofencing and short-range rides.
Given nuTonomy’s success in Singapore and likely follow-on in Dubai, it’s no surprise they view regulatory openness as a competitive advantage for deployment. Under that lens, there are two types of markets in which AVs could flourish, and part of that rests on the impetus for change. Both markets assume a high baseline of density, income and infrastructure. In both scenarios, policymakers are weighing AV benefits against the risk of more Vehicle Miles Traveled (VMT), zombie cars and driver job loss. They are likely also thinking about where these benefits will accrue given the fundamental nature of transport.
Type 1: AVs “pulled” by centralized policymakers
This represents markets like Singapore, Dubai and China, where policymakers could determine AVs are necessary and rapidly roll out the technology with a few best in class players. Here, the impetus will likely be around the environment and/or a desire to be an innovation hub. As an example, Beijing’s recent pollution lockdown has shifted policy to push Electric Vehicle (EV) adoption more aggressively, with a new mandate to make all taxis electric in the coming years. India recently announced a similar 2030 EV mandate. At this juncture, if AVs can position themselves as environmentally-friendly, centralized policymakers will be inclined to ride the AV wave faster, “pulling” technology online.
If this happens, we will likely see AVs become synonymous with EVs. These markets could potentially support the necessary EV infrastructure investments, a major barrier in markets like the U.S. where the government has pushed that onus onto OEMs. This infrastructure need has been complicated further by a lack of battery standardization, which a preeminent AV/EV player could attempt to solve. Right now, Tesla is attacking this opportunity in China with 120 existing supercharger stations and a goal of over 800 by the end of the year, versus 370 in the U.S.
These markets are also more accustomed to large public-private partnerships (PPPs), which makes cost-sharing solutions more likely. Level 4 ownership and paths to monetization remain highly uncertain — who will own these asset-intensive fleets? Who will pay for indirect costs like servicing and maintenance? With Mobility-as-a-Service, will transportation become more regulated, similar to a utility? The questions go on. Thus, PPPs such as Singapore Land Transit Authority’s partnership with Delphi would allow AVs to hit scale economics and adoption while input costs remain high, without having to assume as much capital risk. Furthermore, PPPs make infrastructure alterations like designated lanes and sensor installments more viable, better highlighting the technology’s strengths with platooning and vehicle-to-vehicle (V2V)/ vehicle-to-infrastructure (V2I) communication possible.
Centralized policymakers could allow for testing with looser regulatory restriction in the near-term, allowing for shorter innovation cycles and possibly a faster convergence on regulations around safety, testing, etc. In that case, decentralized policymakers could simply wait and adapt proven models. However, such speed could lead to scaling AV technology prematurely, a risk which OEMs should weigh carefully in deployment.
Type 2: AVs “push” decentralized policymakers to adopt
This represents the U.S. and European markets, where policy likely spans a subset of a desired driving region. Patchwork regulation remains a toss-up. While some view patchwork as an innovation killer, Uber’s NHTSA comments posit that it may allow for simultaneous short innovation cycles. That is, if states are allowed to pursue their desired models, we will see many different models of infrastructure, policy and technology tested in the coming years.
Here, the aforementioned barriers for AVs will remain intact at the national level, but local markets may try to gain a competitive edge by easing testing regulations. As such, those markets stand to arbitrage innovation and investment, essentially replicating the model in “Type 1”. Note, these markets will be secondary cities (e.g. Portland, Detroit, Sacramento and Boston) as opposed to premier megacities like New York where investment is already plentiful and changing the transit status quo seems riskiest. Still, these markets will have to weigh how far to go with resource re-allocation given capital constraints. For example, do they forego other public infrastructure projects to commit to AV? Thus, this model is more likely to lead to localized winners and losers, with the associated companies sharing in that risk.
As such, it pays for players and regulators to move slowly in these markets, with congestion and inadequate public transportation as a forcing function for adoption. Most “Type 2” cities will probably avoid infrastructure or politically-charged issues like designated lanes initially so these cars will be driving in mixed-lane, more complex environments. Also, AV wins on the local level do not equate to AV acceptance at the national level. AV players will have to “push” or convince cities that public goods outweigh perceived public losses, especially around vehicle mileage and employment. Here, surge-pricing may be used to counter latent demand effects — the phenomenon whereby increased road capacity leads to increased driving.
VMT taxation will be more in focus here as well, as a means to compensate cities and mitigate the downsides of zombie cars — cars that simply drive around empty because the marginal cost is so low. VMT taxation has also been touted as a way to fund many of the necessary infrastructure upgrades, especially since American infrastructure needs have ballooned, gasoline taxes have stagnated for decades and AV will mean lost parking fees and traffic fines. Thus, ridesharing and public mobility solutions will be key to selling into “Type 2” markets, and making the economics/utilization work early on. Higher capacity vehicle players like Chariot and Via will seem more appealing, and data sharing will serve as a beachhead offering to cities, a possible explanation behind the launch of Uber Movement in January. Markets like Spain and Italy where ridesharing has struggled and labor relations are paramount will be hard to keep in the fold.
All of this assumes there are some generalizable learnings from each deployment — both around the software and the manufacturing costs. The more that holds true, the less important total addressable market (TAM) size is and the more likely “Type 1” is at the forefront of AV deployment favoring the speed factor. The less generalizable learning models are, the better for large TAM “Type 2” markets, as players will look to deploy in markets that present large data and monetization potential.
Personally, I think current model fragility and the location of most AV talent and innovation favors Type 2 markets, particularly in the U.S. While lack of infrastructure and disjointed policies will make the U.S. a hard market to broadly own, secondary markets like Pittsburgh and Phoenix have already moved to the forefront based on tech partnerships and an aligned political agenda. These markets are embracing AV in the context of ridesharing in mixed-use areas, and are beginning to work on socializing the technology with local commuters. These partnerships are in the early innings, but I expect to see more data-related terms and transit plug-ins as they evolve.
For companies, now is the time to develop your sales pitch for policymakers. Some players have already begun to develop forces targeting mobility solutions for cities. They’re trying to convince cities to be an active force in democratizing mobility. The question they’re getting ahead of: what will be the city’s impetus for change?
Finally for cities, it seems like every city’s move now. This is a chance to get ahead of what could be a privately-dominated shift in value and responsibility. As Portland’s Mayor Ted Wheeler said, “Either the technology will happen to us, or we are going to shape the playing field.”
My recommendation: start building a pilot with a partner of your choice, focused on solving your market’s specific transit and infrastructure needs. Avoid weak data agreements akin to Pittsburgh/Uber, perhaps with clearer legal requirements as seen by New York City’s TLC. If played correctly, cities could begin to regulate/open-source everything from mapping to data while also solving many first/last mile access problems, accruing more value to the general public rather than private players and pushing the technology forward faster. After all, ideas like the Hyperlane could circumvent years of expensive and messy transit construction, and cities could instead begin to think about accurately pricing these autonomous miles to fund any necessary infrastructure spend.