"Add it all up and we’re at $1.3375T. Google’s enterprise value is only $1.823T. This leaves Search + Network being worth $486B."
Net cash is included both in the "add it all up number" and also in the EV. So I think you should've just deduct the $1.3T figure from $GOOG mkt cap, which would have given you about $600B valuation for Search & Network.
Also, not that it matters much to the conclusion, just that $GOOG doesn't have 100% of Waymo. Grok says between 70-80%.
Thanks again for the post, learning a lot from you!
Reposting from X since I think you're focusing on Substack:
A few questions and thoughts based on an initial read:
RE: Youtube - When you use Netflix's valuation, would it make more sense to value YT subscription revenue (~$14bn est.) at the $NFLX multiple (since its stickier, subscription based) and then apply some ad-supported multiple to the YT ad revenue (ie. a $META type-multiple)?
YT also has YoutubeTV which I've seen estimated at ~8 million users now paying $83 per month. The margins on this business are probably quite modest due to content cost (teens? but just guessing). That's ~$8 billion of subscription revenue that would attract a very low multiple, even though it makes YT better and scarier to incumbent cable TV.
RE: Waymo - Has anyone done a simple single car unit economics?
A car has a fixed cost, known depreciable life and operating cost per mile (no one wants to ride in a 150k mile taxi?) Add on the estimated cost of Google's sensors and assume modest operating (compute) costs.
Then just need to know the avg. Waymo fare-per-mile and the key variable becomes utilisation, but 100% provides the upper bound.
Also different possible models of WaymOS licensing (royalty) vs. Owned and operated (above), but would be clearer to understand the actual unit revenue potential, even if long-term margins are unclear.
RE: AI-sistant - Is it possible that Google makes the AndroidOS integrated voice-based Gemini Agent (Siri) so good, so fast, that it becomes a must have feature and $AAPL falls far enough behind that they have to license it or allow it on the iPhone?
RE:Valuation.
Agree SOTP shows a fair degree of Google $GOOG undervaluation.
Another framing is that given an undemanding starting multiple (<20x trailing P/E), it's possible to simply underwrite the EPS growth rate as an expected return without any revaluation.
Each of the dimensions you outline, ignoring Search ads (which are also probably still growing, but economically sensitive), especially Cloud, provide a high degree of certainty of a continued strong EPS growth rate, unless search ads truly dry-up and turn negative.
Would be interested to see you turn the analysis into a range of possible EPS growth rates / scenarios.
re: AI Assistant - definitely possible, though taking share from Apple isn't part of what I feel like is baked in - it definitely IS part of what I perceive as Google having lots of optionality. Apple is already working with Google on licensing their intelligence, though this may get the kibosh if Apple is forced to stop receiving money in return for Google running their browser; And I think your last point is great - returns should track EPS growth plus you get a call option on a multiple re-rate. After the capex cycle slows down they'll start buying more shares too which will be nice. Alternatively, maybe capex cycle doensn't slow because they see incredible opportunities to invest? I"m not put off by that scenario...
re: YouTube - Idk, you could be right. As I say at the end idk how useful an SOTP is in Google's case except insofar as showing how diversified it is. My main thought on YouTube vs. Netflix is that YT business model seems much safer and more diversified to me, with more optionality moving forward. I know I would rather hold YouTube over Netflix if their market caps were the same - no question. I don't feel like there's much signal breaking it down into the individual components - but I could envision it making sense to really try and estimate their margins to get a more direct compare - which I didn't do. Re: Waymo economics, two links: https://www.01core.com/p/driverless-car-costs-have-gotten <<that's mine, there's also this excellent post which has higher cost estimates than mine, https://www.2120insights.com/p/how-autonomous-vehicles-will-change?open=false#%C2%A7medium-term-is-the-price-right;
"Once ChatGPT is plugged into the internet and can manage the process of filtering through hotel options, rental car options, etc - it would become a serious threat."
ChatGPT is getting close. It can connect to the internet and ~do these things already. When I search for Airbnbs, it lists a few but the links are no good (just lead to a general page of the location). A matter of weeks/months, I guess?
Given the price of Lidar is plummeting, in a couple of years, every self driving car will have it.
This is my current view. Why not add something that provides super human capabilities if it isn't too expensive?
Should we also cover Chrome?
Yes, but post was already getting loooooooooong
Great post Ben!
Just a couple of technical points:
I think you may have double counted net cash:
"Add it all up and we’re at $1.3375T. Google’s enterprise value is only $1.823T. This leaves Search + Network being worth $486B."
Net cash is included both in the "add it all up number" and also in the EV. So I think you should've just deduct the $1.3T figure from $GOOG mkt cap, which would have given you about $600B valuation for Search & Network.
Also, not that it matters much to the conclusion, just that $GOOG doesn't have 100% of Waymo. Grok says between 70-80%.
Thanks again for the post, learning a lot from you!
Hah - great catch on both counts - will fix now. Ty brother!
Excellent post!
Thanks Nick!!!
Ben, great post! Thank you.
Reposting from X since I think you're focusing on Substack:
A few questions and thoughts based on an initial read:
RE: Youtube - When you use Netflix's valuation, would it make more sense to value YT subscription revenue (~$14bn est.) at the $NFLX multiple (since its stickier, subscription based) and then apply some ad-supported multiple to the YT ad revenue (ie. a $META type-multiple)?
YT also has YoutubeTV which I've seen estimated at ~8 million users now paying $83 per month. The margins on this business are probably quite modest due to content cost (teens? but just guessing). That's ~$8 billion of subscription revenue that would attract a very low multiple, even though it makes YT better and scarier to incumbent cable TV.
RE: Waymo - Has anyone done a simple single car unit economics?
A car has a fixed cost, known depreciable life and operating cost per mile (no one wants to ride in a 150k mile taxi?) Add on the estimated cost of Google's sensors and assume modest operating (compute) costs.
Then just need to know the avg. Waymo fare-per-mile and the key variable becomes utilisation, but 100% provides the upper bound.
Also different possible models of WaymOS licensing (royalty) vs. Owned and operated (above), but would be clearer to understand the actual unit revenue potential, even if long-term margins are unclear.
RE: AI-sistant - Is it possible that Google makes the AndroidOS integrated voice-based Gemini Agent (Siri) so good, so fast, that it becomes a must have feature and $AAPL falls far enough behind that they have to license it or allow it on the iPhone?
RE:Valuation.
Agree SOTP shows a fair degree of Google $GOOG undervaluation.
Another framing is that given an undemanding starting multiple (<20x trailing P/E), it's possible to simply underwrite the EPS growth rate as an expected return without any revaluation.
Each of the dimensions you outline, ignoring Search ads (which are also probably still growing, but economically sensitive), especially Cloud, provide a high degree of certainty of a continued strong EPS growth rate, unless search ads truly dry-up and turn negative.
Would be interested to see you turn the analysis into a range of possible EPS growth rates / scenarios.
Much appreciated!
re: AI Assistant - definitely possible, though taking share from Apple isn't part of what I feel like is baked in - it definitely IS part of what I perceive as Google having lots of optionality. Apple is already working with Google on licensing their intelligence, though this may get the kibosh if Apple is forced to stop receiving money in return for Google running their browser; And I think your last point is great - returns should track EPS growth plus you get a call option on a multiple re-rate. After the capex cycle slows down they'll start buying more shares too which will be nice. Alternatively, maybe capex cycle doensn't slow because they see incredible opportunities to invest? I"m not put off by that scenario...
re: YouTube - Idk, you could be right. As I say at the end idk how useful an SOTP is in Google's case except insofar as showing how diversified it is. My main thought on YouTube vs. Netflix is that YT business model seems much safer and more diversified to me, with more optionality moving forward. I know I would rather hold YouTube over Netflix if their market caps were the same - no question. I don't feel like there's much signal breaking it down into the individual components - but I could envision it making sense to really try and estimate their margins to get a more direct compare - which I didn't do. Re: Waymo economics, two links: https://www.01core.com/p/driverless-car-costs-have-gotten <<that's mine, there's also this excellent post which has higher cost estimates than mine, https://www.2120insights.com/p/how-autonomous-vehicles-will-change?open=false#%C2%A7medium-term-is-the-price-right;
will respond when i get some time!
"Once ChatGPT is plugged into the internet and can manage the process of filtering through hotel options, rental car options, etc - it would become a serious threat."
ChatGPT is getting close. It can connect to the internet and ~do these things already. When I search for Airbnbs, it lists a few but the links are no good (just lead to a general page of the location). A matter of weeks/months, I guess?