For the past couple of years, I’ve considered appropriate parallels to the rollout of automated vehicles (AVs), so I can better frame how the technology may evolve and scale. Of course no analogue will be perfect, but qualitative comparisons can balance a strictly data-driven measure of economic factors. I’ve seen several different technologies proffered, with much of the recent industry literature and other commentary settling on relatively recent ICT innovations (An example from Deloitte). Household internet. The cellphone. The smartphone.
The adoption rate of transformational technologies has accelerated over time, so it isn’t entirely unreasonable to place AVs in that sequence. In its own AV impacts analysis, real estate asset management firm Brookfield used the following chart to illustrate this progression:
Source: Brookfield
However, I see much of this comparison driven by a recency effect, with analysts overeager to use case studies they and their audiences can personally recall. I’d wager some of it also has to do with the desire to create a compelling narrative that rapid transformation is imminent.
Alternatively, I’ve seen one popular analysis treat vehicle automation as a price premium over current market offerings, and compares it to past automotive innovations. Another likens AV adoption to autopilot on commercial airplanes. And while they offer interesting parallels, neither addresses the shifting relationship between vehicle and consumer.
We will likely see automated vehicles introduced over the next decade in a variety of modes: 1. personal ownership 2. fleets providing on-demand mobility (in some cases shared) 3. fixed route and on-demand public transportation 4. freight and delivery. None of the aforementioned innovations address the variability of mobility and logistics services we will see shaped by AVs.
Instead, there are three revolutionary mobility technologies I’m studying as models for AV adoption, and none were introduced in the past century:
- The railroad
- The automobile
- The telephone
Again, no analogue is perfect, but we can glean insight from these technologies’ initial rollout and evolution over several decades. For instance, Christian Wolmar noted in the The Great Railroad Revolution that the American mentality in the first decades of rail amounted to: “get the track laid and the locomotives built, and start running trains as quickly as possible to generate income, even if that means cutting corners that drive up operating costs. Improvements could be made later, once the line started making a profit.” Cost-saving measures including iron-on-wood rails and no fencing increased operating expenses through rapid asset depreciation and derailments; steel rails and cattle movers only solved these issues later.
Source: Library of Congress
Similarly, we’re watching fleet-based AVs pushed towards rollout before the technology can deliver on expectations, with firms like GM Cruise and Uber chasing first mover advantages. Gen 1 & 2 AVs will likely disappoint customers for a number of reasons (quality control, geofencing, less-pleasant-than-advertised ride experience), and will quickly go obsolete as these firms and others deploy better iterations.
Railroads faced additional operational challenges addressing inadequate supply chains, lack of interoperability between competing lines (also a problem in early telephony), technical safety, and security from train robbers. It doesn’t take a leap of the imagination to recognize how AVs will face Information Age variants of these threats. If AVs go fully electric and manufacturing demand skyrockets, sourcing raw materials may produce manufacturing headwinds. Software reliability and cybersecurity are underappreciated cost and safety factors.
While trains couldn’t operate without new infrastructure, early automobiles drove on the same roads as horse-drawn carriages, and were hardly more useful. Yet over a series of decades, technical improvements and the buildout of road and energy infrastructure tailored to support them increased their attractiveness. Their spread forced policymakers to reconsider spending priorities, as Americans’ love of the technology fundamentally changed their relationship with space. Like trains, they reshaped cities and relegated legacy transportation modes to obsolescence.
Source: Philip Van Doren Stern
And finally, telephones introduced the first ever real-time telecommunications, marking a substantial improvement over the telegraph. But as Tim Wu argues in The Master Switch, Western Union was at a tremendous disadvantage when it deployed telephone networks to take on Alexander Graham Bell’s. While it avoided uses that would cannibalize its telegraph revenue, Bell had no such constraints. Will OEMs – with billions of dollars invested in manufacturing assets, unionized labor, and extensive dealership networks to support volume production – successfully balance these institutional forces against growing service lines if the demand for personal ownership collapses?
Eventually, strong network effects and increasing network complexity at scale forced regulators to grant AT&T monopoly power in 1913. That story shares significant similarities with the operations of transportation network companies (TNCs) today, and the urban AV fleets that will succeed them. (I may address this at greater length in a subsequent piece)
The path to a driverless future will be iterative, uneven across geographies, and destructive to many incumbent players. A deeper dive into prior disruptions of our transportation and telecommunications networks would better shape our expectations for the tumult of our next few decades.
Nathaniel Horadam 