(201) 305-0055
Tim Peter Thinks

Google Won’t Win By Selling AI. Neither Will Anyone Else. (Digital Reset Episode 507)

Screenshot of Big Tech's current stock prices to illustrate why Google won't win by selling AI. And neither will anyone else.

Podcast: Play in new window | Download

Subscribe: Apple Podcasts | Spotify | Amazon Music | RSS

Nobody made much money selling "the internet." The biggest network companies, WorldCom and Global Crossing, built the plumbing and assumed they’d extract a toll from everyone building on top of it. Instead, they and ended up holding the bag when the dot-com bubble burst.

The companies that won — you might have heard of them, like Google, Amazon, Microsoft, and Meta — built something valuable on top of the infrastructure, including services like search, commerce, productivity tools, and social networks.

That pattern is playing out again right now. And Big Tech’s Q2 2026 earnings calls make it incredibly clear.

Anthropic posted its first-ever operating profit: $559 million in Q2 on a $65 billion annualized revenue run rate. They did this not by selling access to foundation models, but from selling coding tools built on top of those models. On the other hand, OpenAI is forecasting $100 billion in ad revenue by 2030, despite the fact that eMarketer projects the entire US chatbot ad market will be under $1 billion this year.

Google made $63 billion in search revenue last quarter using AI to make search better, not by selling AI access. And both Mark Zuckerberg and Satya Nadella used their earnings calls to explicitly describe their ambitions to own the front door between you and your customers.

Just like the internet era, infrastructure is becoming a commodity. The real value is in what gets built on top of it. And the biggest players are racing to own that layer… and to charge you a toll every time you use it to reach your customers.

In Episode 507 of Digital Reset, Tim Peter breaks down what Q2 2026 Big Tech earnings actually reveal about where the value in AI will land, and what every marketing leader and business owner should do about it before gatekeepers stand up their the next set of gates.

What You’ll Learn in This Episode

  • The “AI is the internet” analogy, with a very specific meaning: Almost nobody made money providing the plumbing that powered the internet, and the same pattern seems to be playing out with AI right now
  • Who’s making money and why: Anthropic’s $559M operating profit, Google’s $63B search quarter, Amazon’s doubled “Alexa for Shopping” users… and what they all have in common
  • Who’s not making money and why: Meta’s soft quarter despite record revenues, and OpenAI’s ad revenue forecast versus the actual forecasted size of the chatbot ad market
  • The "super app" moment: Why Satya Nadella’s use of that phrase on an earnings call might be the most important thing Tim heard and why it mirrors the direction he predicted Big Tech will take in May
  • The one significant difference between the AI era and the internet era: Why it makes the analogy an analogy rather than a perfect repeat
  • The "two kids in a garage" risk: The biggest threat every major AI player faces right now
  • Three things to do with all of this”. And how you can get started right now

Key Data Points From This Episode

  • Anthropic Q2 2026: Operating profit of $559M; annualized revenue run rate of $65B. But its source is tools (Claude Code and Cowork), not foundation model access
  • Google Search Q2 2026: Revenue over $63 billion for the quarter
  • Amazon: Alexa for Shopping active users "close to doubling"; interactions up over 5x year-over-year
  • Meta AI: 60% increase in daily interactions, but lowest profits in over a year despite outstanding revenues… all due to AI CapEx spending
  • OpenAI’s optimistic ad revenue forecast: $100M ARR (March 2026), with $2.4B projected for 2026 and $100B projected by 2030, despite eMarketer’s projection that the entire US chatbot ad market will end up under $1B this year
  • Combined Big Tech AI capex 2026: Approximately $725 billion
  • OpenAI capex through 2030: Approximately $750 billion, 25% more than estimated earlier this year
  • Meta Q2 capex: Over $31 billion; full-year guidance $130 to $145 billion
  • Ben Thompson on Meta: "Expenses increased 55% while revenue increased 28%, and a lot of the company’s CapEx hasn’t started depreciating yet. I came away from the call a bit alarmed."

The Three Takeaways

  • AI is a tool, not the end game.


    • You almost certainly don’t need access to foundation models to create better products, services, and experiences for your customers right now. Core to your success is still providing value to your customers at a profit. As Mark Schaefer says: the most human company wins.

  • Big Tech wants to own the front door between you and your customers.


    • Mark Zuckerberg described agents working 24/7 on your behalf. Satya Nadella called Copilot’s TAM structurally larger than Office. Google showed its everything-app ambitions at I/O in May. They are not being subtle. Build your direct customer relationships, including email, SMS, and direct navigation, before gatekeepers put new tolls in place… or raise existing ones even higher.

  • Any of these players could win or lose.


    • AT&T predicted the future in 1993. They had cool commercials. They had an amazing tagline. And then… they didn’t deliver. Google has scale advantages. But AT&T had advantages too. Make sure your brand is the one customers ask for by name, just in case the company you’re renting land from ends up as an also-ran.

    Resources Mentioned

    Related Episodes

    Buy the Book — Digital Reset: Driving Marketing and Customer Acquisition Beyond Big Tech

    Tim Peter has written a new book called Digital Reset: Driving Marketing Beyond Big Tech. You can learn more about it here on the site. Or buy your copy on Amazon.com today.

    See Tim Peter in Action

    Watch Tim Peter break down what Big Tech’s Q2 2026 earnings actually reveal about where the value in AI will land, and why the infrastructure builders are not the ones who will win.

    Free Downloads

    We have some free downloads for you to help you navigate the current situation, which you can find right here:

    • A Modern Content Marketing Checklist. Want to ensure that each piece of content works for your business? Download our latest checklist to help put your content marketing to work for you.

    Subscribe to Digital Reset

    Contact information for the podcast: podcast@timpeter.com

    Technical Details for Digital Reset

    Recorded using a Shure SM7B Vocal Dynamic Microphone and a Focusrite Scarlett 4i4 (3rd Gen) USB Audio Interface.

    Running time: 22:20

    Transcript: Google Won’t Win By Selling AI. Neither Will Anyone Else. (Digital Reset Episode 507)

    Big Tech defines the digital economy. Their actions and reactions tell us a ton about the state of digital, and those behaviors provide incredibly useful information if you’re a marketer or a business owner.

    Even better, their quarterly earnings calls and earnings reports explicitly require them to tell us what’s going on in their worlds, where they’re investing, what’s working… and what isn’t. You know the expression, "Put your money where your mouth is"? Well, Big Tech’s earnings statements and their earnings calls literally show us where they put their money, not just tell us where their mouth is.

    And this quarter’s Big Tech earnings calls showed us a lot.

    Here’s my big takeaway from everything they showed us. AI is a lot like the internet. Yes, you’ve heard people say that before, sure. But I mean this in a very specific sense. I mean that going back to the birth of the internet, almost nobody made any money by providing the plumbing that powered the internet. The folks who made money built value on top of the internet. And with one significant difference, we’re seeing the same pattern play out right now.

    What is that pattern? What’s the significant difference? What does that mean for you? And what can you do about it?

    This is episode 507 of the Digital Reset Podcast. I’m Tim Peter. Today, we’re looking into Big Tech’s earnings and what they tell us about why Google won’t win by selling AI… and neither will anyone else. Let’s dive in.

    Today, I’m going to talk about what Big Tech’s earnings tell us about the business value of AI. And my core thesis is that just like the internet before it, nobody makes money on the infrastructure, the undifferentiated model access. Nobody can make money. Doesn’t matter how good it is. It ultimately is just commodity plumbing.

    Instead, the folks who will win using AI will make their money on products that provide real value to customers. Sure, they’ll need the plumbing, but it’s a commodity. There’s no long-term value there. In fact, at least at the moment, it kinda looks like it’s a money pit. More on that later.

    Think about the internet giants that exist today: Google, Amazon, Meta/Facebook, Microsoft, JD.com, Alibaba, and Tencent in China. None of them got big by selling access to the internet, not really. They got big by offering value on top of the internet: web search, and email, and e-commerce, and productivity tools, and connecting with friends and family.

    Even in the early days, eBay and Amazon were just commerce companies. Yahoo and Google offered directories, search, and email. Microsoft had Hotmail. Yes, there was AOL, which did provide access to the internet, but it was all the other stuff, email and news and stock market information and chat rooms oh my God, the chat rooms, that kept people coming back every day, at least if you could log on once the kids got home from school. And sure, some new folks came along like Netflix and TikTok and such, but none of them are selling "the internet." They’re selling utility on top of the internet.

    Essentially, none of the big internet connectivity companies exist in any meaningful sense any longer. WorldCom got folded into Verizon twenty years ago. Global Crossing got bought by CenturyLink a few years after that. The internet, it’s just plumbing. Nobody makes any real money selling it to customers.

    AI in this context is the same as the internet. It’s rapidly becoming the plumbing. Look at the performance of the frontier models offered by OpenAI, Anthropic, Google, and Meta. They’re all… fine. You know, unless you have very specific technical needs and knowledge, the models themselves are fairly undifferentiated. I don’t mean the front-end features and functions, which I’ll come back to in a moment, or even things like cloud hosting and security and the guardrails around them. Those absolutely have value. I mean the models underneath those front-end features and functions. Sure, any one of them can pull ahead for a few weeks or months and claim to be "the best," but most of us, we can’t tell the difference — we don’t need to care about the difference — for most of our needs.

    Let’s talk about Anthropic and Claude. In May, The Wall Street Journal reported that Anthropic was, quote, "Set to turn an operating profit of five hundred fifty-nine million dollars in the June quarter." And a couple of days ago, they also reported an annualized run rate of $65 billion in revenue. To be clear, they’re making that money, as the Journal said, quote, "As enterprises across the world race to adopt its popular set of coding tools."

    In short, they’re not making their profits from access to the core models. They’re making them from the coding capabilities provided by Claude Code and Cowork, value adds built on top of the core models. Think about it. They couldn’t even offer access to Fable 5 and Mythos 5, their latest, most powerful models, for a few weeks in June because of the, you know, some stuff going on with the Trump administration. That certainly suggests that the value isn’t in the foundation model. It’s in the services built on top of those foundation models that matter to customers. Most people didn’t need Fable and Mythos. They needed the coding services. That’s what Anthropic’s customers are willing to pay for.

    Now let’s talk about Big Tech themselves, almost all of whom had monster quarters. Amazon had a great quarter powered by e-commerce and AWS. Microsoft had a great quarter powered by software sales and its Azure cloud. Google had a fantastic quarter powered by ad sales. Their CEOs, to a person, touted the role AI is playing during their earnings calls in making those core products more effective and more efficient.

    Sundar Pichai said that, and this is a quote, "AI-powered features are driving increased Search usage." And Philipp Schindler followed that up by saying, "Strong performance in Search and YouTube underscores how our investments in AI translate into measurable value for our users and advertisers." Speaking specifically of Search, he said, "Revenues reached over $63 billion for the quarter." He said that "Sundar highlighted the momentum we see in Search, which directly impacts our advertising business." Right? Where are they making the money? Ads.

    Amazon did something similar. They combined their Rufus and Alexa+ assistants into something they’re now calling "Alexa for Shopping." And they say that "active users are close to doubling," and interactions are, quote, "Up over 5x year over year."

    Mark Zuckerberg said that the company’s rebuilt Meta AI drove, quote, "A sixty percent increase in the number of people interacting with the assistant each day," and they called WhatsApp, "the leading surface where people engage with Meta AI." Meta, by the way, didn’t have the quarter Wall Street was looking for. I’m gonna come back to that in a minute.

    Notice, though, that each of these CEOs were explaining to Wall Street how AI is helping their businesses make money. Not making money by selling AI access to people, but using it to enhance their value-add products and services.

    Mark Zuckerberg and Microsoft’s Satya Nadella each went a step further. Zuckerberg said that, quote, "We will have agents that can work twenty-four by seven on your behalf to help achieve your goals and improve your life, your health, your relationships, your finances." And Nadella described how they’re consolidating, quote, "Chat, Cowork, Autopilots, including code in one ‘super app’ spanning both consumer and commercial experiences." He went on to say that Copilot’s addressable market is structurally larger than Office, their core product, ever was. He said, "This is the first time where you have really an enterprise-wide tool which has both per-seat and usage-based pricing. The total addressable market is much more expansive."

    This is the first time I know of that a Big Tech leader used the phrase "super app" in recent memory. I explicitly predicted this was coming following Google’s I/O 2026 in May, referring to "Google’s Everything App Trap." I am increasingly confident that this is how Big Tech leaders think about these tools. They’re looking to own the front door between you and your customers and to charge you a toll every time you connect with those customers. At some point, one or more of these companies will need to own the agentic commerce layer. No one is there today, but I feel very comfortable saying it’s coming. I think Mark Zuckerberg’s agents and Satya Nadella’s "super app" comments are about as explicit on that point as can be.

    So that’s who’s making money. Well, sort of, we’ll get to this. Let’s talk about who’s not, or at least not the kind of money Wall Street wants them to.

    To me, there are two poster children for "not making money" right now, and I think for the same reason. Those companies are Meta and OpenAI. Meta had a soft quarter. They paired outstanding revenues with the lowest profits they’ve made in at least a year. And as a recent Business Insider article stated, "OpenAI had $100 million in annual recurring revenue as of March 2026," but was still forecasting "$2.4 billion for the year and $100 billion by 2030" from ads. Meanwhile, eMarketer projects that the entire US chatbot ad market will be under $1 billion this year. eMarketer’s Nate Elliott called OpenAI’s assumptions "unrealistic." I think that’s super charitable. I’d call them total bull… stuff.

    Each of these companies is plowing obscene amounts of money into their AI build-outs. Meta spent over $31 billion last quarter on capital expenditures, most of that for AI, and will spend between $130 and $145 billion in total… this year. They’re also unwilling to say, they will not give a number, for how much they plan to spend next year. But I bet it’s a lot.

    OpenAI is saying it will spend roughly $750 billion between now and 2030. According to TechCrunch, that’s "…twenty-five percent more than it estimated earlier this year." Yikes. Go figure that Wall Street isn’t showing either one of those companies much love right now.

    Ben Thompson of Stratechery said, speaking about Meta, quote, "Expenses increased 55%, while revenue increased 28%. And a lot of the company’s CapEx hasn’t started depreciating yet. I came away from the call," he said, "a bit alarmed." Yeah, I’ll bet he did.

    Now, to be fair, all of Big Tech is investing in AI. Combined, they’re on target to spend somewhere in the ballpark of $725 billion this year. That’s… a lot. The difference is that Wall Street seems to believe that Google and Microsoft and Amazon have a plan for making money back. Meta said it’s going to start selling excess computing power, kind of like Google Cloud and Microsoft Azure and Amazon AWS, I guess. And all of those are solid and thriving businesses, really. So there’s logic there. But it seems like Wall Street is asking, "if you, Meta, if you, Mr. Zuckerberg, actually have excess computing capacity to sell, why are you investing so much money and not focusing more on where customers find value today?" That’s kind of the whole question, isn’t it?

    To get back to my core point, though, these all look like signs that the big wins aren’t coming from selling access to foundation AI models. Instead, access to foundation models looks like a cost of doing business for Big Tech.

    Now, I want to be fair. There is one big difference between my "AI is the internet analogy" and the actual internet build-out of thirty years ago. That is that the big internet providers back then — WorldCom and Global Crossing come immediately to mind — weren’t building the services on top of the internet. They assumed that they would extract a toll from everyone building services on top of their commodity product… and essentially print money forever. They ended up mostly left holding the bag when the dot-com bubble burst, and then their assets were picked up mostly on the cheap by other telecom companies.

    The biggest difference though, between the internet era and the AI era, as I see it, and the thing that makes the analogy an analogy and not precisely the same thing? AI requires three major components to work properly. One, you need computing power. Two, you need lots of data. And three, you need amazing talent to teach that computing power how to use that data. Then… you need to do that again. Then you need to do that again, right?

    There’s still a whole lot of work to be done before we reach the kinds of capabilities that all of these companies are pursuing, and that Zuckerberg and Nadella made explicit on their most recent earnings call. The "everything app" doesn’t work until customers can just set it and forget it, and we’re not there yet.

    Internet connectivity, by contrast, pretty much required physical wires in the ground. Once enough wires were in place, you could just send any kind of data you needed across those wires without much further investment beyond general maintenance. And without any meaningful way for company A to differentiate its wires from company B… the very definition of a commodity.

    AI, though, requires continual investment. It’s the gift that keeps on taking. That continued investment, and especially the upfront cost, is a huge problem. Again, Big Tech is investing roughly three-quarters of a trillion dollars this year to make all of this work, plus however much investment they’ll need next year… and the year after that.

    Big Tech could be in big trouble if the foundation model capabilities they’re working on can be replicated with lower cost, less power-hungry models in the future. We might not be having nearly so many arguments about data centers and people hating AI if some smart folks, the proverbial two kids in a garage, figure out how to run models like these on your iPhone. That’s the biggest threat all of the big players face right now.

    Returning to history, Google and Microsoft and Amazon were among the companies who provided the value add back then. I think they’re simply applying lessons they learned back in the ’90s that they need to provide the value add on top of the infrastructure, and their success as gatekeepers has taught them that it’s best if they control the infrastructure too.

    If WorldCom or Global Crossing or some other telecom had successfully provided the value add, we might live in a different world right now. Ironically, AT&T had a series of commercials in the 1990s, the famous "You Will" campaign, that showed what the world would look like. They showed concepts we take for granted today, like Zoom meetings and instant translation and distance learning. Their tagline? "And the company that’ll bring it to you: AT&T."

    Why AT&T wasn’t the company that brought it to you is far beyond the scope of today’s episode. I’m sure someone has written the definitive history of why AT&T fumbled the ball, even though they clearly had a vision for what was possible. But my core point remains. Infrastructure and good ideas aren’t enough. You’ve got to deliver the value adds, too.

    All right, I think I’ve made my point. The real question, the real thing you care about is what do you do with this information?

    First, and most importantly, recognize that AI is a tool, not the end game. You almost certainly don’t need access to foundation models to create better products, services, and experiences for your customers right now. Sure, you can use AI to help you define, design, and deliver those products, services, and experiences, but core to your success is still about providing value to your customers at a profit. In fact, given the customer backlash against data centers and the folks actively opposed to AI, you might be better served not using AI some of the time. Heresy, I know. But as my friend Mark Schaefer likes to say, "The most human company wins." Rand Fishkin has a fun take on that same idea, noting that there’s still a place for what he calls "human and weird." I think they’re both a hundred percent right.

    Second, recognize that Big Tech wants to be the front door between you and your customers. "Gatekeepers gonna gate" isn’t just a pithy phrase I thought up. It’s a fundamental law of the digital economy. In their earnings calls, Mark Zuckerberg and Satya Nadella explicitly stated their ambitions to make that happen. Google did the same thing at I/O this year. These folks want to be "the everything app." They’re not content with being plumbing. They plan to offer the value add for many services… and to be the gatekeeper between all the rest. Your job now is to build direct connections with your customers, grow your email and SMS list, grow your physical mailing list, connect with customers directly before gatekeepers cut off your access or raise the tolls you’ll pay to reach them.

    Finally, keep in mind that any of these players could win… or lose. Yes, they’ve all got great plans. They’ve all more or less got access to computing power, data, talent, and the money to keep those things going. I’ve gone on record saying that Google’s scale in those areas gives it a huge advantage. But you know what else? AT&T had a great plan, too. They didn’t succeed in delivering on it.

    You know the saying about don’t build your brand on rented land? Well, one risk of building your brand on rented land is that the landowner you’re renting from might fail. You could have the best strategy for succeeding on Facebook or Instagram or YouTube or SEO or Amazon or app stores, and the company providing that platform could get upended by another member of Big Tech, an aspiring member of Big Tech, or by two kids in a garage. Make sure you’re building value adds your customers are willing to ask for by name, are willing to use your brand as the prompt. That way, you’re in a great position no matter what. Your business won’t depend on a given company winning. It will lead your company to win in every case.

    Now, if this episode gave you a clearer read on where Big Tech is actually heading and what that means for your business, do me a favor. Send it to a colleague who’s trying to figure out their AI strategy right now. It might save them from building their brand on an expensive piece of rented land.

    You can find the show notes for today’s episode, including all of the research and the sources that I cited, at timpeter.com/podcasts.

    And if you’re ready to go deeper on building a brand that customers will ask for by name, one that helps you bypass gatekeepers, my book, Digital Reset: Driving Marketing and Customer Acquisition Beyond Big Tech, is the roadmap you need. You’ll find it on amazon.com and bookshop.org, and the links are in the show notes.

    Thank you so much for listening today. I genuinely, genuinely appreciate you. Until next time, please be well, be safe, and be excellent to each other. I’ll see you soon.

    Take Your Next Step Toward a Digital Reset

    "Digital Reset with Tim Peter" helps you look beyond the "shiny objects" to build a business that lasts. How can we help you today?

    • The Brief: Get the weekly email that turns these strategic ideas into actionable demand. Subscribe to The Digital Reset Brief
    • The Book: Master the framework with Digital Reset: Driving Marketing and Customer Acquisition Beyond Big Tech. Buy the Book
    • The Experience: Need a bespoke digital strategy for your hotel, resort, SaaS firm, or financial services firm? Tim Peter & Associates can help you. Work with Tim

    Tim Peter is the founder and president of Tim Peter & Associates. You can learn more about our company's strategy and digital marketing consulting services here or about Tim here.

    Back To Top
    Search