The Signal · Issue 01 · Physical Intelligence
The robotaxi business learns to borrow
Waymo raised its first debt, began driverless testing in Detroit and announced Singapore for 2028. Autonomy is being financed and rolled out like infrastructure, city by city.
Facts are attributed and linked to their sources. The analysis and opinions are the author's own.
What it is
On 8 October Waymo CFO Steve Fieler wrote that the company had closed a $5 billion term loan, its first debt financing, describing it as a step in Waymo's evolution into a scaling commercial enterprise. He noted an earlier $16 billion equity investment this year and that Waymo had launched service in its fifteenth U.S. city the previous month.
Two days earlier, Waymo began fully autonomous driving in Detroit, initially for employees only while it establishes a safety framework, with public rides planned once the technology is properly validated, as reported by WDIV. The same day, electrive reported that Waymo will begin operations in Singapore with manually driven testing in 2027 and aims to offer fully autonomous public rides in 2028, working with the Ministry of Transport and the Land Transport Authority.
Why it matters
Debt is a statement about predictability. Lenders lend against cash flows they can forecast. Raising a large term loan says the robotaxi business is now seen, at least by these lenders, as something closer to an operating business than a research bet.
Lenders lend against cash flows they can forecast. That is the real news in a robotaxi term loan.
The rollout pattern is equally telling: familiarise, test with employees, validate, then open to the public, one city at a time, with local authorities involved. That is staged deployment, the physical-world version of the gated model releases we covered in Issue 01.
There is a useful contrast with humanoids. Robotaxis operate in one well-defined task, moving people safely from A to B, with years of logged miles behind them, which is why they can be financed and expanded city by city. Narrow scope and long measurement history are what make a physical AI system bankable. General-purpose robots have neither yet, and the gap between the two industries is largely a gap in evidence.
From a credit perspective, the interesting question is what the lenders underwrote. Autonomous fleets are capital-intensive, with vehicles, sensors, depots and remote assistance. The repayment case rests on utilisation, pricing and safety performance city by city. Those are measurable, and the loan implies someone believes the measurements are stable enough to lend against.
The employee-first phase is a good governance pattern and an underrated one. It gives real-world exposure with a population that can be briefed, surveyed and supported, before the general public is involved. Enterprise AI deployments can copy it directly: internal users first, structured feedback, then customers.
The international expansion raises a data question as well. Each city has its own road layouts, weather and driving norms; the system has to be validated locally rather than assumed to transfer. That is the physical version of the sim-to-real gap, and it is why the rollout is slow by design.
What I'd watch
- What 'properly validated' means in each city, and whether criteria are published.
- Performance in variable climates; northern cities are a harder test than sunny ones.
- How regulators like Singapore's LTA define and audit safety requirements before public service.
Sources
- Waymo Closes $5 Billion Debt Financing. Waymo, 8 Oct 2026.
- Waymo begins fully autonomous driving in Detroit. ClickOnDetroit (WDIV), 6 Oct 2026.
- Waymo to launch operations in Singapore in 2028. electrive, 6 Oct 2026.