Marketing School - Digital Marketing and Online Marketing Tips

Google Is Burning Its Free Cash Flow To Win AI

22 min
Aug 3, 202623 days ago
Listen to Episode
Summary

The hosts discuss Google's unprecedented negative free cash flow as it aggressively invests in AI infrastructure, arguing the risk is justified given its $514B cloud backlog and diversified asset base. They explore Naval Ravikant's thesis on vertical software integration replacing horizontal SaaS, and introduce the concept of a 'revenue brain' — an AI-powered, vertically integrated system managing sales, marketing, and outbound. The episode closes with a deep dive into ChatGPT Voice as a personal command center, data privacy concerns with AI platforms, and the open-source alternative Hermes Voice.

Insights
  • Google's negative free cash flow is a calculated bet, not a crisis — its $514B Google Cloud backlog signals massive future revenue realization that justifies current infrastructure spend.
  • The shift from horizontal SaaS to vertically integrated AI 'brains' (revenue, finance, company) represents the next major software paradigm, where opinionated, data-rich vertical solutions outcompete generic tools.
  • Traditional agencies face existential pressure as AI agent stacks handle cold outbound, content creation, retargeting, and ABM — even Google Ads insiders are telling agency operators the old model is dead.
  • ChatGPT Voice combined with computer-use capabilities is emerging as a practical Jarvis-style command center, capable of setting up tools, managing recruiting flows, and publishing content autonomously.
  • Data trust is becoming a competitive differentiator for AI platforms — OpenAI's failure to fully delete user data is eroding enterprise confidence and pushing users toward on-premise open-source alternatives like Hermes.
Trends
Hyperscalers burning free cash flow on AI infrastructure as a strategic land-grab, not a financial distress signalVertical AI integration replacing horizontal best-of-breed SaaS stacks across every business functionAI 'brain' products (revenue, finance, legal) being productized by major SaaS players as retention and expansion toolsVoice-controlled AI agents becoming daily operational command centers for entrepreneurs and executivesOpen-source voice models (e.g., Hermes Voice) enabling on-premise AI assistants that bypass data privacy concernsTraditional marketing and media agencies being disrupted by AI agent stacks that automate core deliverablesRecruiting automation via AI agents integrating LinkedIn, ATS, and outreach tools into unified workflowsOpenAI and Anthropic expanding into high-margin verticals (legal, life sciences) using frontier modelsApple shifting AI partnership from OpenAI to Google following IP disputes with OpenAI/Jony Ive ventureEnterprise trust hierarchy emerging: Apple, Google, Microsoft, Amazon ranked above OpenAI and Anthropic for sensitive data
Topics
Companies
Google
Reported first-ever negative free cash flow as it aggressively invests in AI infrastructure to compete in the AI race.
Google Cloud
Has a $514B revenue backlog, cited as the primary justification for Google's massive AI infrastructure spending.
Microsoft
Reported strong earnings with Azure showing ~100% year-on-year growth, validating cloud AI investment thesis.
OpenAI
Discussed for ChatGPT Voice capabilities, data privacy concerns, and alleged failure to fully delete user data.
Anthropic
Mentioned as a competitor with shaky business model and new voice product; hosts expressed low trust for sensitive data.
Apple
Cited as most trusted company for data privacy; reportedly pivoting AI partnership from OpenAI to Google.
Amazon
Listed among the most trusted companies for enterprise data handling alongside Google, Microsoft, and Apple.
Nvidia
Briefly mentioned as a trusted company with no data privacy concerns raised by the hosts.
YouTube
Cited as one of Google's key distribution assets contributing to its competitive moat in the AI race.
Waymo
Mentioned as one of Google's strategic investments illustrating the breadth of its asset portfolio.
Google Ventures
Referenced as a strong investment arm that has delivered solid returns as part of Google's broader portfolio.
NetSuite
Used as an example of a major SaaS player likely to productize a vertical AI 'financial brain' offering.
Victor
Cited as an example of a vertically integrated AI brain product living inside Slack, doing well in the market.
Single Grain
Eric Siu's agency, referenced as evolving toward AI agent-based revenue services rather than traditional agency work.
LinkedIn
Used as an example platform that ChatGPT Voice with computer-use can log into autonomously for recruiting workflows.
People
Neil Patel
Co-host discussing Google's AI investment strategy, data privacy concerns, and ChatGPT Voice use cases.
Eric Siu
Co-host introducing Google's negative free cash flow news, the revenue brain concept, and ChatGPT Voice workflow.
Anthony Pompliano
Tweeted the Bloomberg chart about Google's negative free cash flow that sparked the episode's opening discussion.
Chamath Palihapitiya
Referenced from the All-In Podcast discussing Google's ~30% annual shareholder returns over the past 10-20 years.
Naval Ravikant
His tweet on vertical software integration replacing horizontal SaaS was used as the framework for the revenue brain ...
Jack Dorsey
Released Buzz, a Slack-for-agents platform that Eric demoed setting up autonomously via ChatGPT Voice.
Jony Ive
OpenAI reportedly acquired his hardware company, leading to an IP dispute with Apple over AI hardware development.
Quotes
"Google is willing to risk the entire company to win the AI race."
Eric SiuEarly segment
"People tend to forget businesses operate to maximize revenue growth and profitability. Their whole purpose is to keep making more money. People just don't invest for the sake of investing."
Neil PatelMid segment
"Now that software is cheap, value moves to vertically integrated businesses that deliver opinionated end-to-end experiences."
Eric SiuMid segment
"I fully deleted my data as an experiment... I'm telling you, it's still pulling stuff from the past. That just makes me not want to trust them."
Neil PatelLate segment
"The more time you spend on recruiting great people, the less time you spend on management. Great people don't need to be managed as much."
Eric SiuClosing segment
Full Transcript
2 Speakers
Speaker A

So Neil, I think Google is doing something right now that is shocking to me because this is the first time that they're going, they're using up their free cash flow to win the AI race. So Google is going negative on their free cash flow, right? And so by the way, they make a lot of money. It's not like they're starving for money. I want to show this little chart over here and then we can react to it. Anthony Popliano here tweets this. So Google is willing to risk the entire company to win the AI race. And then Google reports first quarter of negative cash flow. So this is free cash flow, right? This is reported by Bloomberg and you can see it for the first time. It's gone down. They started in 2004 and then they've continued to move up. Neil is so anyway, I think it's interesting because my first reaction to this is, are they betting on the right thing? Now let's think about everything that Google has, right? Google has Search, which is their money making machine. Google also has, they have their TPUs. Okay? Google also has YouTube as well. Google also has Google Cloud. And then Google also owns what else? They have Android. So Android Ecosystem, they have Chrome as well. And so they just have all the distribution in the world. They have Waymo as well. By the way, you think about some of the investments that they've made from Google Ventures, like they do pretty well there as well. And so, and so all this data center, you know, data center investments, build outs that they're doing right now, it's. Is it truly a risk at the end of the day to win this AI race? In my opinion, I don't think it's much of a risk because of how well they've done over the years. Like you can say they're a big conglomerate now. But you know, I think I was listening to the all in podcast. Did you listen to the recent one where Chamath was just talking about what the return on their investments has been in the last, I don't know, let's call it 10, 20 years or so. Do you know, I think he was

0:00

Speaker B

talking about the Google stock in which the average returns of your shareholders roughly 30% a year, not necessarily their revenue growth or their EBITDA growth. But what most people tend to forget is Google's not investing only for the sake of investing. They have a 514 billion billion backlog for its Google Cloud division to give everyone a quick perspective on the numbers and why they're spending all this money they want to actually take that backlog and realize the revenue from it.

1:57

Speaker A

Yep. So I think, and they, and I think they will. Right. You look at all the demand, by the way, you look at Microsoft, yesterday they reported earnings, right? Their stock has gone up quite a bit. You look at how Azure, which is their cloud provider, like how much it's gone up, I think it was like 100% year on year growth or something like that. So all that to say is they already have proven business models and in fact most people are going to be coming to them and they're not risking everything. You look at kind of how shaky things might be right now with like an anthropic or open AI with these, these, these open models. Right. And so I think their business models are going to work, but I think they're getting a lot of pressure in other areas versus Google has already established themselves. Like they have plenty of business models that are already working for them. So.

2:27

Speaker B

Dude, totally. And people tend to forget businesses operate to maximize revenue growth and profitability. Their whole purpose is to keep making more money. People just don't invest for the sake of investing, especially this amount of money without really realizing how much more money they can end up making. It's done intentionally. And they're looking at how much more growth potential they can get if they do some of these investments.

3:09

Speaker A

Yep. So one thing I want to talk about. So I was reading this, this, this tweet from Naval and I think this, this stuff all kind of ties in together. So you look at what Naval is tweeting over here. So he says when software was expensive, so thin, horizontal, best of breed software stacks extracted rents across every business. So what does that mean exactly? So we're talking about thin and horizontal. So horizontal means it can cover like a bunch of people. Right. And well, what he's saying now is that now that software is cheap. Keep in mind, before you needed a bunch of software engineers, you still do. But today anybody can vibe code something up. Right. And that's kind of how he's saying that software is cheap now. So now that software is cheap, Valley moves to vertically integrated businesses that deliver opinionated end to end experiences. Right. So what does that mean exactly? That means instead of trying to be everything to everyone, okay, you actually now are, let's say Neil and I are both in marketing. So marketing would be one vertical. Okay. So vertically integrated would be. And then let's say Neil has data from his companies that he works with. I have data on the stuff that I do. So Neil's gonna have his opinion on how something should be built, I'm gonna have my opinion as well. But now that software is more democratized, this is where the value is occurring now. And that's why I think everyone needs to build a revenue brain, right? So I've talked about single brain, but a revenue brain is more vertical, meaning that it covers sales, it covers marketing. That revenue brain can be something that integrates with your cold outbound system. So it's handling cold outbound for you, but maybe it's also integrated with your AEO SEO software suite too. So it's creating content, it's also looking at the customer calls that you're having. So it's taking all the connectors that you have and it's figuring out oh, where should we make retargeting ads from account based marketing standpoint, where should we figure out where we should attack more on the cold outbound or just outreach standpoint? So if you have a revenue brain that's going, we've talked about having your, your website be like a, like its own revenue agent. That is an example of a vertically integrated experience that you can deliver. Because what I'm seeing right now is when you look at companies out there like Victor, they're doing really well. They're like that's like a little like a, like a, like a brain that lives inside your slack. There's plenty of other competitors out there that's much more horizontal and that self serve. The way I look at it, I don't think agencies are going anywhere. I think agencies are going to evolve. But I think for us, I know for us it's more so hey, we're on the revenue side because that's what we understand and we have our opinions there. And I had a conversation with Google this week, Neil, a couple of people from the Google Ads team and they're like, oh man. They're like, we think traditional agencies are dead. I'm like, oh, it's interesting that you say that, right? Because I've talked to a handful of these traditional agencies, people that are trying to jump ship. They're like, yeah, that model doesn't work anymore. We're more interested in what you're doing on the agent side right now because we think what you're doing is going to, that that's going to affect how agency work is going to be done in the future. So that was the conversation that I was having and I think all this stuff kind of ties in together. So I think whether you need a finance brain or a revenue brain or a Company brain. Whatever brain you need, you need some, you need these vertically integrated brains. And there's gonna be plenty of companies offering this stuff. You know, Neil and I are probably more on the marketing brain, revenue brain side of things.

3:34

Speaker B

Not only are we seeing people build it themselves, we're seeing a lot of companies trying to offer solutions for it. But I think you're gonna start seeing a lot of this stuff just productized by some of the major SaaS players and in different verticals. So like the net suites of the world, My guess is over time they'll offer like a financial brain. You know, a lot of the marketing software solutions will do that on the marketing side, so forth, so on. And I believe that, you know, people can end up building it in house and I think a lot of people will, but majority of companies will just pay for a solution that already exists and they can just plug and play with everything else that's they're already doing. And I believe most of the software solutions that they're already paying for will start offering some of this stuff for free. It'll just be value added to keep retaining you.

6:38

Speaker A

Yep. And that's why I think right now, like the, the way we think about the single brain stuff, we're doing a lot of these, these free pilots. We're not even charging in, in, in, in, in many situations. Right. Because it very much is a situation where like how do you just get more logos in the door and how do you. Because it is, it is a retention and an expansion game for, for a lot of these, these companies. And you know, what I communicated a week or two ago was right now, like when we talk to the, one of these investors, like, oh, we talked to our portfolio company, like everyone's building this right now. And therefore, like, you know, we, they think that because everyone's building right now, it's not going to be a hot thing. I think in fact, if everyone's building it right now and they have the engineering resources, it means that more people are going to need this because you're going to need people that are managing the, who's going to manage it, who's going to optimize it, who's going to upgrade it. Um, so anyway, I think to Neil's point too, yeah, like a lot of these companies are just gonna, gonna, they're gonna offer this as well or they're probably gonna gobble up companies that are doing it too.

7:23

Speaker B

So. All right, so what's the next topic that you wanna move on to?

8:19

Speaker A

Yeah, so Neil, have you been using ChatGPT voice at all.

8:25

Speaker B

You know, I've never used the Voice feature inside ChatGPT, not once.

8:30

Speaker A

Okay, have you seen. Remember the. Do you remember watching Ironman?

8:35

Speaker B

Huh? Jarvis.

8:39

Speaker A

Jarvis. Okay, this is basically getting very close to a Jarvis. And so what I do in the morning now, I take out my phone, I'm like, hey, give me the daily command center, right? It's like, okay, here's what's happening right now on the recruiting side. Here's what's happening on sales side here. It's also evaluating all of the forms I have on my site to make sure nothing's broken. I just want to make sure everything's working right. And it's like, here's a recommended action for these things over here. Do you, do you approve or disapprove? And I'm like, approve, disapprove. I'm like, what about this thing that we're working on yesterday? Where's that at right now? Where's this at? Oh, by the way, not only is this operating in ChatGPT, it's connected to all my other harnesses because my CHAT GPT can use my computer, right? So yesterday I was setting up Buzz, which Jack Dorsey released. You know what I did to set that up? I was just talking to him. I was just talking to my ChatGPT voice. I'm like, hey, I need you to set up Buzz for me. I need you to invite these people over here. I need you to make sure that all the permissioning is set up correctly. And so now, by the way, it's hooked in, you can hook it in with, you know, you can have it sourced for you on the recruiting side. You can have it send partnership emails, you can have it do whatever you want, right? But ChatGPT voice has been great because ChatGPT has a browser and plus you can control the other apps that you have. And so my Hermes harness is available because my Hermes is much more powerful from like a connector standpoint. So all that to say is that, you know, you hear about these people right now going on these like four hour hikes and they're just talking to these things now, Neil. Now when I'm on a commute to go see my dad on the weekends, guess what? I'm just going to be talking. No more need to listen to podcasts. I'm just going to be talking to this thing for an hour and asking it to do things for me real quick. If you want to acquire customers faster and more efficiently this year with the latest strategies and tactics, then check out single grain.com that is my ad agency again www.singlegrain.com check it out and if it seems like a fit, we'll get in touch and help you with a free marketing plan.

8:40

Speaker B

Yeah, dude, and I think voice is going to get better and better. What I'm really waiting for is Apple to have a better Siri and it just being all integrated within, you know, your hardware that you have. And I believe they were supposed to partner with ChatGPT. It did not work out because Chat GPT or OpenAI supposedly building hardware with Johnny Ivey because they ended up buying his company. And I think Apple claims that they stole some intellectual property or something like that.

10:28

Speaker A

Definitely did.

10:56

Speaker B

And then OpenAI is claiming that Apple never integrating. I believe Apple's moving over to Google, but I'm waiting for their tech to get better and it just be integrated within the current systems and workflows.

10:58

Speaker A

Yeah, but Neil, why do you even need that when you're already paying for ChatGPT? You're paying for the $200 a month version. And so your, your home computer. Right. So the main thing is I have it set up on my home computer over here to desktop. Right. So it's like your command center. It's already hooked in with everything else. It already has all the context, plus you can control like your phone too. And so you already have all the context in the world. You want to just do this right now?

11:12

Speaker B

Honestly, I don't trust ChatGPT. I've deleted a lot of my backlog and history and it still pulls through it when I ask questions.

11:34

Speaker A

Oh wow, they don't fully delete your data.

11:42

Speaker B

I fully deleted my data as a experiment. And then people are like, give it a few days, give it a week. I'm telling you, it's still pulling stuff from the past.

11:44

Speaker A

Not everything.

11:51

Speaker B

So if I uploaded files and I was careful, I never uploaded files that I shouldn't have uploaded. But some of those files, yeah, you can get it from the pantry buddy. Some of those files, they don't actually have the file to reference, but they reference it in chats in the past. And the data that it referenced in the chats, that stuff is still there. When I ask it questions, it's still pulling from it. But if I had to ask something new and I had to pull from the original file, again, it's not doing that, but it's pulling from past chats where we analyze data and I just don't like it. So I hate to say it, I don't Trust it.

11:52

Speaker A

I still use it.

12:32

Speaker B

Just being honest, but I won't use it for anything sensitive. I just.

12:32

Speaker A

So here's a question for you. Which companies would you trust with your data? Apple.

12:36

Speaker B

I trust Apple. I trust Apple probably more than most. I also trust Google and Microsoft. Sure. Sit here and eat it, buddy. The reason I trust Google and Microsoft is they're so big, you know, everyone's like, oh, they're using their data to figure out what search results to show that kind of stuff. I don't mind if I.

12:41

Speaker A

You're so ocd. Just let William eat whatever he wants.

13:02

Speaker B

Well, I look to see my wife. He ate a little piece of chocolate, but the chocolate doesn't taste good. I'll just tell you that. So you can assume what's in it. Like, her brownies are made of sweet potato and dates and chocolate powder and vegetables, like her pancakes.

13:08

Speaker A

That's a scam. Chocolate.

13:25

Speaker B

Yeah. And I'm like, yeah, this stuff. She won't do dark chocolate because of caffeine. But still, like, it's. I've tried it. And I'm like, what the heck is this?

13:28

Speaker A

Yeah.

13:36

Speaker B

So even, like, before we get back to it, last night my wife was gone. She had a family emergency. So I'm taking care of the kids by myself. And there's like, mom's not here. We want a popsicle. So I. I gave him one of their popsicles, and there was three in there. So I had one because the kids were like, eat one with me. It's like, no added sugar. It's made of pure few. It tastes so nasty. And the kids were like, this is delicious. Because they don't know what the real stuff tastes like. You know, like the stuff we ate as kids. I took one bite and I threw it away. They're like, it's delicious.

13:37

Speaker A

You don't like. I'm like, no, you don't know what delicious is. But see, that's good marketing. That's good marketing.

14:09

Speaker B

That's.

14:16

Speaker A

The stuff that your wife has consumed is healthy. Right. But it's all stuff that she's probably seeing on Instagram or YouTube.

14:16

Speaker B

Yeah. But going back to Google and Microsoft, if I delete my data from Google and Microsoft, I'm not talking about search results or whatever, but I believe Google would genuinely delete my data and not reference it in the past for questions that I asked. And ChatGPT can say, whatever. I swear I'm asking questions. And it's pulling from the past on stuff that already deleted.

14:24

Speaker A

I would tell you. Okay, let me tell you, who I trust first and then what you would do if you're a little patient with this, this is more for the audience. So I trust Amazon, I trust Microsoft, I trust Google, I trust Apple. Right. I trust Nvidia as well. Like, you know, never had any issues there.

14:48

Speaker B

Right.

15:02

Speaker A

But those are, those are kind of the main companies that I trust. Now what I'll say is this. I don't necessarily trust Anthropic either. So I don't trust Anthropic or, or, or OpenAI. And so I wouldn't go like with, with super sensitive stuff ultimately. Here's what I would say if you want to run this, because Hermes Voice just came out yesterday, which is open source, by the way. So Hermes, you can run it on your own box, you can run it on your own systems and so you can talk to it, you can activate it. Just saying, hey Hermes. Kind of like, like, hey Siri. Right. So that Neil, you can actually set up. And because my Hermes harness is, is very powerful, I could just run it through that. And you can run it through open weights, like open source models. Right. So you can definitely do that. Now if you're impatient right now, you want the ChatGPT voice experience like that. It's unparalleled for the moment. I would say that Claude just came out with their own voice as well. Just keep in mind that they have their own incentives ultimately. And one more thing I'll say is that I believe that OpenAI and Anthropic, they're probably just going to build more and more applications and attack more of the profitable niches that they see their customers like legal, they're getting into legal, they're getting into life sciences, they're getting into. I think they're going to get into more and more stuff because they have these frontier models. Right. And you know, you know, I think there's also good reason to not trust both companies because of how both of the, let's just say the executives, how they operate. And I believe that how you do anything is how you do everything. So that being said, I still take advantage of it.

15:03

Speaker B

Yeah. So right now you're using voice. Okay. On ChatGPT. Are you asking it anything sensitive relating to data that it shouldn't have or you're worried that it could end up using in the future?

16:23

Speaker A

No, I don't, I don't do very sensitive stuff like financial analysis and like that type of stuff using, using these.

16:39

Speaker B

So yeah, when we do financial analysis, we do it in a controlled environment. We do not upload it to ChatGPT, we never have for stuff like that. I've used it for, like, personal stuff, like just asking like, hey, analyze my blood test results.

16:46

Speaker A

Right?

17:01

Speaker B

If that went out into the public, I know people are like, oh, health information and HIPAA and stuff. Honestly, if my blood test information went out into the public, I really wouldn't care. Right? But the fact that I can still ask it questions related to that pisses me off. And it still shows it. It just, to me, it just makes me not want to trust them. If you're asking questions related to your business and getting stuff done, you're saying with chatgpt voice because none of it's really sensitive, you're okay with just pushing it out there?

17:02

Speaker A

Yeah. And I'll give you some examples. Okay, so some things I use ChatGPT voice for right now are, oh, hey, look at the latest marketing school YouTube channel. I want you to create trial reels for. I want you to edit the videos, put an overlay, and then create six trial reels and then publish them. Right. And that can do it. Right. So because it has like the metrical API, that's one thing I ask it to do, because then it just controls my Hermes. My Hermes has all the connectors. Another thing I asked it to do is I yesterday, I didn't want to set up Jack Dorsey's new buzz, right? Which is like their slack for agents. I didn't want to set. It just set up all the stuff for me. It set up all the harnesses and it invited the right people and it put together all the channels. That was very nice for me. Right. The other thing too, as I was just testing this out because on our talent side, we're trying to recruit more people right now. But I was just testing out like a couple of these recruiting flows. What it did was it logged into my. It logged into LinkedIn, it logged into our applicant tracking system, it logged into our juice box, which is our, like, recruiting agents, and it figured out a way for it to reach out to these people in a tasteful way and figure out how to combine all the things. Right. And so that, to me is helpful because now I can hand that off to other people on my team and then they will work like that. And that to me is a very high leverage thing. Because in my opinion, the more time you spend on recruiting, the more. The less time you spend on management.

17:32

Speaker B

The le. The more time you're saying the more time you spend on recruiting, the less time you spend on management.

18:48

Speaker A

Yeah. If you spend more time on recruiting.

18:53

Speaker B

Great people.

18:54

Speaker A

Great people don't need to be managed as much.

18:55

Speaker B

Yeah, I thought you meant the more time you spend, like just recruiting. Like as an entrepreneur should spend all their time recruiting. Oh, no, you just mean if you recruit the right people, you don't have to deal with management.

18:59

Speaker A

That is it for today, and we will see you tomorrow.

19:10