Consider the general research - in all, it doesn't eliminate people, but let's say it shakes out to speeding up developers 10% over all tasks. (That includes creating tickets, writing documentation, unblocking bugs, writing scripts, building proof of concepts, and more rote refactoring, but does not solve the harder problems or stop us from doing the hard work of software engineering that doesn't involve lines of code.)

That means that it's worth up to 10% of a developer's salary as a tool. And more importantly, smaller teams go faster, so it might be worth that full 10%.

Now, assume other domains end up similar - some less, some more. So, that's a large TAM.

Those that aren't willing to pay for it directly, can still use it for free, but will just have to tolerate product placement.

It very much does not assume that, only that some fraction will have become accustomed to using it to the point of not giving it up. In fact, they could probably remain profitable without a single new customer, given the number of subscribers they already have.