Can Google Afford to Launch AI Mode?

The new Google AI mode looks slick and is a great user experience... but does the math add up?

Is AI Mode the future for Google? 🤖

Last month Google announced it would start testing “AI Mode” — a completely new way of searching that allows users to ask more free-form and exploratory questions. Much like Perplexity or ChatGPT, while AI Mode still surfaces web links, the bulk of the focus is in the chat experience.

At the moment it’s only out as a limited test to US beta users, but it goes without saying Google needs to tread very carefully. In 2024 nearly 70% of its revenue came from search and display advertising, so any changes here could obviously have big consequences.

But what if AI Mode tests amazingly from the UX perspective, increases search usage and fends off advances from the ChatGPT cutting into their market share? Can they afford to scale to 100%? Let’s take a look at some back of the envelope math, please feel to @ me with any assumptions that don’t make sense.

Can they afford it? 💸

There are a lot of factors at play, but the most obvious question is could Google survive the damage of:

Reduced search advertising
Less traffic sent to websites

The first being their revenue juggernaut, but website traffic to publishers is also essential as Google is the biggest provider of ads there as well. All this doesn’t even factor in the fact that serving LLM results is much more expensive than traditional search.

Hypothetical Scenario Time 🤔

Let’s consider that a combination better user experience (you get the answer directly) and fewer links in AI Mode decreases CTR on links by 25% — I think this is pretty conservative given the new UI, one study about AI Search showing a 91%! decrease. On the flip side let’s assume the better UX leads to +5% more searches.

With $200B in search revenue let’s assume 25% drop in CTR leads to a 15% decrease in revenue (CTR drop my skew more towards organic clicks) which would be a loss of $30B, but that’s not it. If websites then see 25% fewer visits Google Ad Network revenue may take a drop of $8B as well (from a base of $35B / year).

The revenue declines would offset this somewhat with +$10B from the 5% increase in searches, but I think this is asking a lot as there are already 14 billion searches per day. In this hypothetical the net revenue loss is $28B per year! Certainly not great, but only 8% in percentage terms for a revenue machine like Google.

Area	Impact Estimate per Year
Search revenue (2024)	$200B
CTR drop in AI Mode	25%
Revenue decrease (CTR)	-$30B
Fewer visits (Ad Network)	-$8B
More searches (+5%)	+$10B
Net revenue loss	-$28B

To be revenue net neutral, AI Mode would need to increase searches by 19%! Never count Google out, but this feels to me like a tall task given its search market share declined last year below 90% for the first time in 20 years.

How this all plays out remains to be seen, but there’s no doubt Google’s has some difficult decisions ahead of it. We're in for the new UI and White Lotus analysis though :)