The first self-driving taxis available through Uber in London arrived
Thursday with a small technical complication: they still have
drivers.
Fifteen Wayve-equipped vehicles can now be matched with passengers
through the Uber app, but a licensed human remains in the front seat,
ready to take control. Fully driverless operation needs separate
regulatory approval, which is not expected immediately. In a city with
more than 100,000 private-hire vehicles, this is less a robotic takeover
than a very cautious introduction.
It is also an unusually clear picture of Uber’s future.
Uber does not need to manufacture the best autonomous car, write
every line of its driving software or even remove the human from the
front seat this afternoon. It needs to remain the place where people
press a button when they want to go somewhere.
That sounds almost modest until you consider what the button
controls.
The company
between the passenger and the car
Uber once tried to build autonomous vehicles itself. It sold
that unit to Aurora in 2020 after years of heavy spending and a
fatal 2018 testing crash in Arizona. The retreat looked, at the time,
like surrendering the most important technology in its own industry.
Its newer approach is less romantic and potentially much more
powerful: let other companies endure the expensive, uncertain work of
teaching cars to drive, then offer them passengers, maps, payments and
local operations when they are ready.
Uber describes its Autonomous Solutions
business as infrastructure for autonomous-vehicle developers. The
pitch includes mapping, fleet operations, regulatory reach and the
street-level knowledge collected through billions of trips. Its
data-collection fleets and dashcam network, the company says, gather
more than 100,000 hours of driving footage across the United States and
Europe.
In other words, Uber is trying to become the department store of
automated transportation. The shelves may hold vehicles from Wayve,
Waymo, Rivian, Nuro or somebody not yet famous enough to have a booth at
a technology conference. The customer relationship still belongs to the
store.
Thursday’s London launch shows why that matters. According to The
Guardian, roughly 100,000 people registered for the autonomous
rides. Wayve gets access to willing passengers in one of the world’s
messiest driving environments. Uber gets to teach those passengers that
a robotaxi is not a separate product requiring a separate habit. It is
simply another car that might appear after the usual tap.
[AD PLACEMENT]
The arrangement also gives Uber a hedge against picking the wrong
technical winner. Building one autonomous system means betting on one
stack of sensors, software and manufacturing assumptions. A marketplace
can carry several.
Uber has been assembling those relationships aggressively. A March
agreement with Rivian, for example, calls for an initial 10,000
autonomous R2 vehicles, with deployments planned to begin in San
Francisco and Miami in 2028. The companies say the program could expand
to 25 cities and up to 50,000 vehicles, although those are
forward-looking targets dependent on technical milestones and
regulation—not cars already waiting in a parking lot.
That distinction is important because the robotaxi business has
always produced more confident timelines than functioning taxis.
A platform is not a
neutral passenger
Uber’s strategy has a tidy logic, but it should not be confused with
merely helping clever car companies find customers.
The platform that matches rides can influence which fleets receive
demand, how prices are presented, what happens when a vehicle fails and
whether human drivers are treated as partners, backups or temporary
scaffolding. If several robotaxi operators depend on Uber for
passengers, the app gains leverage over companies that supposedly own
the transformational technology.
The same week offered an uncomfortable reminder that “platform
efficiency” always reaches people eventually.
On September 2, CEO Dara Khosrowshahi announced that Uber would
reduce its workforce by about 10 per cent. His message to
employees said the company was removing management layers, combining
teams and creating capacity to invest in growth and “the autonomous
future.” TechCrunch
reported that the reduction amounted to roughly 3,300 jobs.
The layoffs do not mean robots directly replaced those workers;
Uber’s stated reasons centre on organizational complexity, coordination
and duplicated structures. Linking every lost office job to a driverless
car would be a cheap trick.
Still, the timing reveals the company’s priorities. Uber is
simplifying the organization while building a broad network of
autonomous partners. It is preparing for a future in which its most
valuable asset may be neither employees in an office nor drivers in
cars, but its position between transportation suppliers and demand.
[AD PLACEMENT]
There is a credible counterargument. Autonomous vehicles remain
expensive, regulators remain cautious, and even Uber’s
autonomous-mobility head told The Guardian that nobody currently
operates them more cheaply than conventional human-driven rides.
London’s cars still require a person at the wheel. For years to come,
human drivers will likely remain essential in many markets, weather
conditions and awkward pickup zones that confuse both maps and
machines.
Uber itself says it expects human-driven and autonomous vehicles to
share the platform. That may be sensible forecasting. It is also the
position that preserves Uber’s relevance no matter how quickly the
technology advances.
If robotaxis stall, Uber still has drivers. If they succeed, Uber has
partners. If different companies win in different cities, the same app
can display all of them.
The great robotaxi contest is usually described as a race to build a
car that can drive itself. Uber is making a quieter bet: the winning car
will still need someone to summon it, somewhere to find passengers and a
system that knows where the airport pickup line begins.
The future may arrive without a driver. It will probably still ask
where you would like to go.
