[dmix]

Waymo is talking about scaling up operations globally and the market is competitive, the cost 100% does matter.

They need large Chinese production lines for lidar, integration kits for cars plus the in car computing, repair pipelines for both sensors and cars, real estate to park cars, the infrastructure/processes to clean and charge them quickly, teams of remote drivers, insurance policies, etc. Then they need to compete with mature decentralized Uber and taxi fleets who push their car/maintenance costs onto drivers, while Waymo grows adoption of their mobile app where prices will matter if they aren't as perfectly reliable and low risk as hiring a human. The self driving novelty effect won't last forever

All of that requires large capital expenditure and careful business models

[spacebanana7]

Google is capable of burning truly huge amounts of money on projects that look exciting and have long term prospects (e.g Youtube). They could lose $10-20 billion a year on Waymo for a decade if needed.

You can't just cancel Sergey's favourite pet project, regardless of economics.

[dmix]

That's easier to stay that when it's an R&D project doing pilot runs in a small set of cities. When you need tell shareholders you want to run a fleet of 100k cars then those numbers start becoming very serious.

Waymo also took $11B from outside investors, so it's just not Alphabet taking the risk

[cyberax]

They don't actually _need_ to do any of that. They can just license the technology to automakers and local operators.