Yeah, this is what I surmise too, but I do have doubts about how lucrative this model will be after the eventual transition to agents. Basically my thinking is this:

1. Google has monopolized the current AdTech stack from all sides. Not only does it dominate the place where most lucrative ads are displayed (search), it dominates the tools used by publishers to sell ads, the tools used by advertisers to buy ads, and the exchange that matches them. And as the AdTech antitrust lawsuit found, they've been pulling some shenanigans in the auctions to pad their margins by eating into from the the publishers' margins.

2. The search UI is simultaneously limited real estate but also easy to stuff with ads, which is why it has been so valuable. (Google also monetizes a chunk of the display ad space across websites, but that is very clearly dead as plummeting traffic numbers show.)

3. The entire serving stack has been crazy hyper-optmized for this model.

However, with agents this whole stack gets bypassed. The conversational UX is very different. A user will not do N searches for a product with M ads on each SERP, resulting in NxM impressions and ~N clicks as they explore the product page and reviews to come to a conclusion.

They will instead have a conversation with an agent over N rounds of back and forth and only the final round will matter somewhat, ideally resulting in a final, very valuable click.

So now the ad impression volume (which is apparently more important than CPC these days?! https://www.marketingdive.com/news/google-parent-shifts-basi...) has gone down by a factor of NxM. Oh, eventually they will start displaying ads at each turn of the conversation, of course. But due to the user doing all the product exploration in the chat itself, it seems highly unlikely they'll click on those, and so those may not be very valuable at all.

Now layer on top of that the fact that LLMs are probably more expensive to serve than search or ads. So there's pressure on the top- as well as bottom-lines.

Furthermore, this could shift ad-selling and buying patterns, so all those shenanigans padding their margins may also vanish. (This will be a problem for Amazon too BTW.)

It is unclear that the value increase in that final click will compensate for that NxM display and margin-padding loss. Concerningly, even if N is 1, M these days is on the order of 10+.

Google will still be a ridiculously lucrative company, but its search + ad revenue could be a fraction of what it is today. That cloud revenue better pick up fast.