A strategic hypothesis for how Uber can navigate the transition to highly automated vehicles and maintain its leadership position. 

Originally prepared in October 2016 by Suguru Azegami, John Deniston, Stephen Lewis, and Damien Scott as a long-form strategy paper for a Stanford Graduate School of Business course, STRAMGT574 by Prof. Robert Burgelman and Sven Beiker.Adapted for online publication in February 2017 by Stephen Lewis.


Despite not yet being profitable on a global basis, Uber’s position as a leader in today’s U.S. ride hailing market is clear. It has the largest network of drivers, making its service the fastest and most convenient.  But, how can Uber be able to maintain its lead in a future where its supply of drivers is no longer an advantage, because swarms of driverless vehicles (Highly Automated Vehicles, or “HAVs”) roam the streets? 

Our hypothesis is that Uber will be most successful if it can convert its drivers from a network of laborers into a network of capital asset owners/providers, thereby allowing Uber to remain asset light and minimize supplier power, two characteristics of its success to date. However, in order to navigate the decade-or-longer period before HAVs are available for consumer purchase at a reasonable price, Uber must carefully partner with OEMs to advance the development of HAVs and prove their effectiveness in initial markets.

Evaluating Uber’s Optimal Future-state Market Structure

One of the biggest questions facing the TNCs is “who will own the vehicles?”  There are several possible answers, each with different implications for Uber’s market position and profitability.

Uber owns the fleet – This approach would give Uber complete control over its vehicles, allowing it to optimize routing, maintenance, and customer experience. However, ownership and management of a fleet would be a radically new competency Uber would have to develop.  Additionally, it is unlikely that Uber could convince investors to finance asset purchases of this scale.  Finally, purchasing vehicles from OEMs subjects Uber to drastically increased supplier power as compared to today’s industry structure, in which individuals provide the vehicles. 

OEMs own the fleet – This approach gets the vehicles off of Uber’s balance sheet. However, it assumes that OEMs are willing to serve this role, which is not a given.  OEM’s do not traditionally own and operate vehicles eithers.  And, the OEMs that are already actively pursuing competing routes into the TNC business would be unlikely to partner with Uber (e.g., GM with its investment in Lyft, or VW with its creation of Moia). Assuming Uber could find willing partners, the limited number of OEMs would mean Uber would be subject to increased supplier power.     

Individuals own the fleet – This structure would be most favorable to Uber.  In this scenario, individual owners bid their personally owned HAVs into the Uber network when not in use.  This allows Uber to remain “asset light” and extend its current model of decentralized ownership into the HAV future. Critically, this keeps the supplier power facing Uber very low.  Uber would have a more difficult time controlling routing and customer experience, but those are challenges it is experienced in managing. 

Guiding the Market toward Uber’s Preferred State

Consumer choice among TNCs is primarily driven by convenience (vehicle availability/proximity) and price. As such, the first player to deploy a HAV fleet with enough scale to achieve convenience and affordability in a given market will likely be the market leader in the long run.  If Uber waits on individual owners to build its HAV fleet, it will lose the race to critical mass to a competitor that directly owns and deploys its HAV fleet.

Therefore, Uber’s strategic actions for navigating this transition period should include accelerating research and development of mass produced HAVs, demonstrating business model effectiveness with the initial HAVs through targeted ownership of HAVs or partnership with OEMs, and finally, inclusion of personally owned HAVs in its fleet.

Immediate Action

Uber’s current unit economics are not favorable and in many markets it is loss making, largely because acquiring and paying human drivers is expensive. The faster that HAVs are available, the faster Uber can become profitable at a unit level. While the company has raised substantial capital to date, this may become harder if a path to profitability remains distant and uncertain. Therefore, Uber must accelerate the industry toward mass adoption of HAVs, as it has done to date with its Pittsburgh-based Advanced Technology Center.

However Uber should do more to support the efforts of other firms pursuing the development of highly automated vehicle technologies, recognizing that it stands to benefit from the field’s faster movement.  Given proper legal structures, such actions would also have the effect of establishing vital partner relationships with OEMs.  This could involve open sourcing its real and synthetic training scenario databases, creating partnerships to facilitate trials of technology from other companies and sharing high-resolution mapping data.  Uber could require its drivers to install a free device that has a front facing camera and OBD2 interface into the vehicle, to capture driving scenarios and how drivers react to them. This would allow Uber to leverage its driver base to undertake fleet learning on a massive scale.

Transition to HAVs

As soon as HAV technology has been sufficiently tested and approved by regulatory bodies for commercial deployment, Uber should own and operate its own small fleet of HAVs in key launch locations. Initially, it must maintain control over the vehicles, including their deployment and fleet operations, so that it can learn and adapt rapidly in response to customer feedback and experiences. It should use vehicles equipped with its own technology and also proactively secure partnerships to deploy HAVs produced and even owned by other companies.

Large well capitalized OEMs that are actively working on competing ride hailing products are likely to reject Uber’s invitation to deploy their vehicles (imagine GM in partnership with Lyft, or BMW with its DriveNow / ReachNow services). Uber should target smaller OEMs that do not have the resources or intent to pursue their own vertically integrated offering (e.g., Volvo or Chrysler). Uber should aim to partner with three to five different OEMs during the transition period to provide scale and limit the partners’ power.

Scaling the HAV Fleet           

As the cost of AV technology comes down, it will become more feasible for individuals to purchase their own HAVs.  As this happens, it is critical that Uber maximizes the number of individually owned HAVs in its fleet by ensuring OEMs allow individuals to bid their vehicles into the Uber network.  OEMs that are directly pursuing the TNC market will be unwilling to make their vehicles compatible with Uber’s network.  However, Uber should again look to partner with OEMs that do not have their own TNC or automated vehicle strategy in order to access those OEMs’ vehicles, if an owner opts to contract with Uber.

Once broad vehicle compatibility is available, Uber can leverage its strength in recruiting and incentivizing drivers to convert its drivers (who contribute their labor) into asset owners (who finance Uber’s HAVs) – this could be an unparalleled long-term competitive advantage. Such an approach could be positioned as a highly positive step in managing the technological obsolescence of its drivers putting the company in favor with the public, regulators and customers. From a societal perspective such a transition would help to navigate the relative decrease in the role of labor compared to capital to the overall economy.

To increase individual’s access to HAVs and expedite the deployment of individually owned HAVs in Uber’s fleet, Uber should work with the finance industry to create shared ownership financing structures. On a small scale, this might look similar to shared ownership of taxi medallions. On a larger scale, this could look more like asset securitization or real estate investment trusts.

As Uber’s fleet of HAVs grows, so will its competitive advantage over its competitors, which might include OEMs attempting to operate their own TNC services.  Eventually, because vehicles can be more lucrative when in larger, more efficient fleets, car buyers will prefer that their new vehicle is “Uber compatible”.  OEMs that are still restricting their customers from accessing Uber’s network in an attempt to support their proprietary TNC networks may begin to see vehicle sales decline, and eventually, may be compelled to provide Uber access. At that point, Uber’s leadership position will be secure.


Author: Stephen Lewis
MBA Candidate, 2017, at Stanford Graduate School of Business

LinkedIn