Dear Reader,
As I was telling a friend the other day, we’re all winners in this election.
We don’t get any ads anymore.
Woohoo.
Today I want to talk about Google versus AI, and specifically Google versus Anthropic and OpenAI.
I’ve learned that some college kids who take business classes watch this channel.
When I talked about the incumbent’s dilemma, I got emails and private messages asking me to explain how to think about it.
The Dilemma
Put yourself in Google’s seat.
You watch ChatGPT and Claude come along, and you think, oh my goodness, they can take my search business away.
You’re the incumbent.
How do you protect your business, and how do you attack your opponent at the same time?
The question people asked was whether these chatbots would take all of search away from Google.
Once you dig into Alphabet’s numbers, Google is winning the usage fight.
They aren’t losing their search business.
But their answer to the threat, their AI counterattack, is starting to cost them money.
That’s the case they’re going to teach in business school.
The dilemma is no longer about losing users.
It’s about earning less on each user who uses AI Mode instead of regular Google search.
What’s Working
Usage is growing faster than it has in years.
Search and other grew 12% last quarter, and year-over-year query growth was the highest in five years.
The Gemini app reached a billion users, which management called the fastest-growing product in Google’s history.
So the chatbots haven’t hollowed out the number of questions people ask Google, whether it’s through classic search or through Gemini.
Margins are holding too.
Google Services operating margin rose to 40.7% from 38.5%.
Traffic acquisition costs, which is what Google pays partners to send it traffic, held at roughly 20% of ad revenue.
And ads in AI are going global, with AI Mode expanding to international markets for the first time.
What’s Cracking
Search growth is decelerating.
It slowed from 19% in the first quarter to 17% in the second to 12% in the third.
The market was expecting something closer to 14% or 15%.
That miss is what sold the stock off so hard.
Here’s the reason.
More and more people are using the AI product, and Google gets paid less every time they do.
Let’s say, hypothetically, that every time someone searches on Google, Google makes a dollar.
And every time someone uses AI Mode, Google makes 70 cents.
For Google to be fine, it has to bring the yield on an AI query up close to what it earns on a search.
That’s the big challenge.
How do you monetize AI as well as you monetized search?
Google doesn’t disclose revenue per query for AI Mode, and it isn’t about to volunteer it.
So we don’t know yet.
Once they get comfortable with the numbers, they’ll probably start releasing them.
For now, total queries, search and AI together, grew.
But AI inside Google is growing fast and earns less, and that’s what’s putting pressure on the results.
Why I’m Still Bullish
Full disclosure: I own Alphabet stock.
I’ve owned it for a while, and I’m not selling it.
The business case isn’t written yet.
But I did a back-of-the-envelope calculation this week.
I think Google Services is worth about $2.7 trillion.
Google Cloud, about $1.1 trillion.
Waymo, about $100 billion.
Add roughly $140 billion in net cash and about $135 billion in private stakes.
Back out the cost of the AI spending at the Alphabet level, and I get a business worth about $3.8 trillion.
That works out to a base case of $312 a share, a bull case of $400, and a bear case of about $270.
The stock is trading at $278.
I think it’s worth more than that.
And I think all of the Mag Seven are trading at a discount right now.
There’s a good reason for the discount.
Several of them aren’t generating free cash flow at the moment.
But everything I see tells me they’re following the return on investment, and it’s very rare to be able to buy powerful companies like this so easily.
The Risk
Google carries more risk than Amazon, and in my view more than Microsoft.
The question is whether it can monetize AI at the same rate it monetizes search, or very close to it, to defend its margins.
If it can, this is going to look like a cheap buy.
If I’m wrong, it slowly declines over time.
I don’t see that happening.
I’m not as pessimistic as some people are.
Now compare it to Meta.
Meta’s core business is going crazy.
It’s using AI to hyper-target ads.
Who would have thought social media would be touched less by the AI revolution, because it isn’t in the search business?
Social media is nonsense for society, but it’s great for an investor.
For the College Kids
Here’s the lesson in one picture.
You’re the Roman Empire.
The barbarians are at the gate.
How do you defend your territory, and how do you counterattack?
Google’s counterattack is putting AI Mode into its search results.
The problem is that AI Mode doesn’t earn as much yet.
My personal theory is that as long as people keep using it, and they are, Google will figure out how to monetize it.
Google’s dilemma is that its own AI counterattack earns less than the search business it’s defending.
That’s what an incumbent under pressure looks like.
A moat doesn’t have to be software.
Sometimes it’s paperwork.
One American gold company spent roughly 15 years winning its federal permits and beating legal challenges, and construction is now underway.
A competitor can’t copy that in a year… or in ten.
A federal bank has also approved close to $3 billion in financing for the project.
Have a wonderful day.
I’ll see you tomorrow.
“The Buck Stops Here,”
Found this helpful? Share it with others.
Written by Dylan Jovine