Did you use hot water to wash them?
Probably yes. Last time I did the math making them out of plastic saved energy vs heating water to wash glass (it was a close call, it depended on if you used a dishwasher vs sink washing).
So it might seem environmental friendly, but if you run the numbers it's not necessarily so.
Power markets are very liquid, so I tend to favour calculating your personal footprint based on the average co2 output of the grid, not the particular energy production method you've chosen. This makes it clear that even if you're paying for wind power, it's still better not to waste electricity.
Did your calculations include things like the extraction and distillation of the petroleum?
Beyond the environmental impact, there's also just the stupid, wasted effort that goes into pointlessly recreating the same thing over and over when you don't have to. In this case, the bottle goes back on the return trip from delivering new milk, which had to happen anyway. The milk ends up being the only thing delivered, on net. It's just strikingly elegant in its lack of wasted effort and material. It's not clear if it's a net win, though, given the delivery part, unless there are others on the same street signed up for the same day.
If anyone's in the Bay Area and is interested in delivery sort of like this, which also tries to set up regular delivery routes to minimize the marginal impact of a customer, I recommend checking out Farmstead (https://www.farmsteadapp.com/). They provide a similar milk delivery service by picking up Straus milk bottles and refunding the deposit. They also solved my biggest issue with most of the delivery services by picking up their reusable bags and having minimal packaging inside, but COVID put a stop to the bag pickup, at least temporarily.
The issue isn't getting bottles hot enough to kill germs. It's doing so in a way that doesn't aerosolize the bacteria and preventing all the other ways cross contaimination can occur.