Bloomberg Tech

Chris Urmson on Scaling Driverless Trucks From Proof to Profit

Chris Urmson· Co-founder & CEO at Aurora Innovation
·~7 min·English·Bloomberg
RoboticsAI CompanyBusiness StrategyPolicy
TL;DR

Aurora co-founder and CEO Chris Urmson explains why, with the self-driving problem largely solved, scaling driverless trucks is now a story about manufacturing rigor, utilization economics, a two-step path to profit, and a driver shortage with rare bipartisan backing.

01How They Build

Two Paths to a Thousand Trucks

Aurora is scaling driverless trucks two ways at once — upfitting existing International trucks with its driver kit, and having OEMs like Volvo and PACCAR install that kit straight on the factory line.

Today, we have two paths to bringing vehicles to scale.

— Chris Urmson, Bloomberg Tech
Key Insight
Running both paths hedges the scaling risk: upfitting puts trucks on the road now without waiting on factory integration, while the OEM line is what makes a thousand-plus trucks a year actually manufacturable.
02How They Build

The Hard Part Is Now Manufacturing, Not Driving

With more than half a million driverless miles proven, the remaining challenge is industrial — standing up a production line with mass-manufacturing rigor, where a six-week supplier slip is a normal growing pain, not a setback to the plan.

It's a whole level of additional rigor that you have to put in to achieve this manufacturing goal.

— Chris Urmson, Bloomberg Tech
Key Insight
The problems Urmson names — supplier quality, material availability — are classic automotive manufacturing issues, not autonomy ones. That is the tell that the self-driving question is largely behind them and the execution risk is the kind every hardware company carries.
03How They Sell

Selling a Driver, Not a Truck

Rather than owning and operating trucks itself, Aurora is shifting to a driver-as-a-service model — customers keep their trucks, financing, and maintenance, and Aurora supplies the Aurora Driver, which can roughly double how much each truck is used.

they own and operate trucks. They understand the financials of that. They really know how to maintain them

— Chris Urmson, Bloomberg Tech
Key Insight
Splitting the truck from the driver lets each side do what it is best at: fleets keep the capital and operations they already understand, and Aurora sells only the capability no one else has, which is a far easier thing to scale than running a trucking company.
04The Economics

The Economics Come Down to Utilization

A driverless truck does not tire or need to live near the route, so it can run far more miles — Aurora targets about 225,000 a year versus the 100,000-125,000 a human manages, and in hard-to-staff routes like the Permian it can move five to six loads a day instead of three.

they can only really get three loads of fracking sand per day moved with a person driving the truck, and they can't get a whole lot more people to go there. We could do between five and six.

— Chris Urmson, Bloomberg Tech
Key Insight
Framing the gain as utilization, not labor savings, reframes autonomy from replacing drivers to getting more out of trucks that already exist — a larger and more politically durable story, because the binding constraint is miles, not wages.
05The Economics

Positive Per-Mile in 2027, Profitable in 2028

Aurora's second-generation hardware is the lever to reach positive unit economics — making money on each mile in 2027 — with company-level profitability targeted for 2028.

we could make money each mile that we drive. Right. We don't do that quite yet, but we expect to achieve that in '27. And then as a company, we expect to get to a scale where we're actually profitable as a company in '28.

— Chris Urmson, Bloomberg Tech
Key Insight
A per-mile margin a full year before company profit tells you fixed costs — the autonomy R&D and the new manufacturing line — still dominate. Profit waits on volume, not on the per-trip math, which is why the thousand-truck ramp matters more than any single route.
06The Policy

A Rare Bipartisan Tailwind

Aurora already runs driverless trucks across Arizona, New Mexico, and Texas, and Urmson reports support from both the administration and Congress, with the Department of Transportation taking concrete steps to enable the technology.

we're seeing bipartisan action to support automated trucking. We're seeing the Department of Transportation really take concrete steps to enable this technology to benefit Americans.

— Chris Urmson, Bloomberg Tech
Key Insight
Bipartisan backing is unusual for an automation story, and it holds here because the framing is a shortage rather than job loss. When the labor being replaced is labor no one can hire, the usual political opposition has little to grab onto.
07The Demand

The Pull Is a Shortage, Not a Bet on Demand

Urmson frames Aurora's market as filling a gap the industry simply cannot staff — an expected need for a million drivers over the next decade, deepened by policies that pull hundreds of thousands more out of the pool.

we used to expect to need a million drivers over the next decade.

— Chris Urmson, Bloomberg Tech
Key Insight
Pinning the market to an unfillable shortage makes autonomy a necessity rather than a cost-cutting choice — but it also ties Aurora's demand story to labor-policy numbers it does not control, like the nondomiciled-driver rules Urmson cites.