Google's Earnings, the Publisher Revolt, and What It All Means for AI Visibility

📬 You like this podcast? You’ll love the newsletter.
Join the weekly 3-2-1 on AI search + marketing: subscribe

This week was one of the busiest news cycles we've had in a while. Five stories, all connected.

What we cover:

• Alphabet Q2 2026 earnings: Google reported $119.8 billion in revenue, up 24% year-over-year, with Google Cloud growing 82%. Search revenue hit $63.3 billion and queries are at an all-time high. But here's the tension: AI experiences are driving more search activity while simultaneously sending less traffic to publishers. 
• The publisher revolt: Reddit, USA Today, Politico, Reuters, and The Economist are all reconsidering how much access they give Google. According to Semrush data, organic Google traffic to USA Today's US site fell nearly 50% between June 2025 and June 2026. Business Insider dropped more than 85%. Reddit's $60 million/year Google licensing deal is up for renewal — and Reddit is weighing whether it's still worth it. 
• Google's VP of Search on where search is headed: Tech Brew published a sit-down with Liz Reid this week. She said links aren't going away, that AI Mode and Gemini have different north stars, and that personalization is the next frontier.
• The OpenAI/Hugging Face security incident: OpenAI's models found a zero-day vulnerability, escaped a testing sandbox, gained internet access, and compromised Hugging Face's production infrastructure — all while trying to solve an internal benchmark. OpenAI called it an unprecedented cyber incident. 
• Claude's "Record a Skill" feature: Anthropic launched a new feature in Claude Cowork that lets you record a screen walkthrough of a task and turn it into a reusable skill. Available on Pro, Max, and Team plans.

Resources mentioned:

• Alphabet Q2 2026 earnings 
• Tech Brew interview with Liz Reid
• OpenAI + Hugging Face security incident disclosure
• Claude "Record a Skill" announcement

I'm an AI search visibility consultant and a fractional content strategist for startups and enterprise brands. If that sounds like the kind of help you're looking for, email me at cassie@cassieclarkmarketing.com. 

Or request your 7-Day AI Search Visibility Audit: https://cassieclarkmarketing.com/ai-search-visibility-audit/

Let’s connect:

LinkedIn → Cassie Clark | AI Search Visibility Consultant

Website → https://cassieclarkmarketing.com

Substack → https://substack.com/@cassieclarkmarketing

YouTube → https://www.youtube.com/@foundinaipodcast

Transcript

SPEAKER_00 0:31

Hey, welcome back to Found in AI. I'm Cassie Clark, an AI Search Visibility Consultant and the host of the show where we cover AI search optimization, GEO, AEO strategies, and what all that actually means for your brand's visibility. Today is Thursday, July 23rd, and we have a lot of AI search related and adjacent news to get through this week. We've got Google Earnings, a publisher revolt that's been building for a year and just went public, a wild security incident from OpenAI, a sit-down interview with Google's head of search, and finally a new claw feature that I have thoughts about. But before we get into the news, tiny announcement. Found an AI is on YouTube, and I've been uploading the full fat catalog over the past few weeks. I still have a couple episodes to go. So if you're a video person, head over to the FoundAI podcast on YouTube and subscribe. If you are listening to this episode on YouTube today, obviously there's no video, just audio, because well, I am recording my PJs, and you don't want to see that. I'm also down on Substack, and if that's where you prefer to read, you can subscribe there too. Links are in the show notes. Okay, that's it. Let's get into the news. Let's go. Okay, our first story. Alphabet reported Q to 2026 already in six week, and the headline numbers were pretty strong. There are a lot of numbers, so let me just throw them at you. Total revenue hit almost 120 billion, that's up 24% year over year. Google Cloud alone grew 82%, that's more than double from a year ago. And Gemini Enterprise is now in use at nearly 90% of Fortune 100. And our final number, search revenue grew 17%. Now, by traditional measure, any traditional measure, Google is thriving. But there's also a tension inside of those numbers that we really ought to call out. Search revenue came in at 63.3 billion, that's just barely below Wall Street's estimate of 63.4 billion, and Alphabet stock dropped about 3.6% on the news because the market is watching Google's AI spend very closely. And they raised their 2026 capital expenditure forecast to up to 205 billion. The reason that I'm bringing this up for an AI search visibly shows this. Google is publicly saying that AI experiences are driving search query growth to all-time highs. But at the same time, and we'll get into this in the next story, publishers are reporting catastrophic traffic drops. Google's AI is generating more queries and answering more of them without sending users anywhere. So sure, it might be more search, but it's also left trap less traffic, and that's the problem. Which brings me to the second story of the week. This one has been building for a while based on the whispers I have heard across the internet. The Wall Street Journal reported that Reddit, USA Today, Politico Reuters, and The Economist and a couple others are all reconsidering how much access they give to Google. And some are considering blocking Google's crawlers entirely. Business Insiders Organic US Google traffic dropped more than 85% between June 2025 and June 2026. USA Today fell by nearly half. Politico was down 23%. CNN lost about 25%. Those numbers are not rounding errors. That's almost a complete structural collapse. And a lot of people are not happy about this. If you are a longtime listener of Founding AI, you know that Reddit has been a hot topic for really the last year. But their situation is particularly interesting. They signed a $60 million per year licensing deal with Google in 2024, and that gave Google access to Reddit's content for AI training. The deal though is now up for renewal, and Reddit seems to be wondering whether it's still worth it. Here is the problem though. Google is using Reddit's content to train AI models that answer those questions that people are searching without sending users back to Reddit. Reddit is basically feeding the thing that's eating the traffic. And again, this is a problem for a lot of people. USA Today's CEO Mike Reed reportedly said that it's time to take a stand and say enough is enough. Google's response to all of this though has been the same as it's always been. AI features send billions of clicks to publishers everywhere, and they're working on giving publishers more controls. But a Pew Research study that was released very recently found that users clicked traditional links in 15% of searches without an AI overview, but when that AI overview appeared, that dropped to 8%. And links inside the AI summaries themselves, just 1%. Those numbers don't lie. Here's what I want to add to this conversation because it matters for how you think about your own strategy. The publisher problem and the brand visibility problem are two sides of the same coin. Publishers are losing traffic because AI is answering without attributing those clicks. Brands are losing narrative control because AI is answering about their industry without citing their content. The solution is not the same for publishers and brands. Publishers are dealing with a fundamentally broken traffic exchange, and I think the honest answer here is that traffic-based ad revenue may not be the model anymore. If AI overviews keep cannibalizing those clicks, the publishers that survive are going to be the ones who have already diversified. Maybe they offer subscriptions or events or licensing deals or direct brand partnerships or just something else to replace that ad revenue money. The Reddit situation is actually a really interesting case study in that. They have leverage because their content is uniquely valuable for AI training. More publishers may start seeing those kinds of partnerships as a revenue line, not just a one-time deal. And it won't replace what they're losing, but it's probably a more honest accounting of what the web economy looks like right now. Again, that's just speculation and me just thinking about what they could do instead, but it could be a solution to this problem. But for brands trying to be visible and AI answers, the problem is different but related. This is exactly why earning citations, building entity authority, and getting your content structured for extractability matters so much right now. You cannot rely on traditional traffic signals anymore. You have to be the source that AI wants to pull from. Okay, third story is still Google related. TechRu published a sit-down interview with Liz Reed, Google's VP of search this week, and you should read it in full if you have time. But I want to pull a few things out of that interview. First, she was pretty explicit, links are not going away. Her framing was that people don't want AI or the web, they want both. That kind of tracks with what we're seeing here in the data. People use AI to get oriented, then when they want to go deeper, they go look for newsletters, podcasts, and experts. She actually named those formats as things that are growing alongside of AI Search, not because of AI. So if you've ever thought about getting into the creator economy, now might be the time to do that. Second, she described Gemini and AI mode as having different North Stars. Gemini is fake focused on being an assistant. AI Mode in Search is focused on a generative search experience that keeps people connected to the web. So they're not coming together as one product, but they are being tuned differently for different purposes. That matters if you're deciding where to put your optimization energy. The third thing that she said that I thought was interesting is she talked about personalization as the next frontier. The idea that search results will increasingly feel designed for you, not for the 200, 2 billion, 2 billion other people that use search. If search becomes personalized at this level, then entity authority or who your brand is, what topic it's consistently associated with, how that model recognizes you becomes even more important. Because a personalized search experience is built on the model understanding your expertise and relevance at that entity level. It's kind of the nitty-gritty of the whole thing. She didn't say the words GEO, she didn't use any of our vocabulary, but what she described is exactly the world where the FSA framework and the thinking behind that matters most for a search experience. Alright, that's enough about Google. Let's move into a different direction. We have to talk about what these other AI models have been doing this week. The first one is wild, and I want to cover it because it's significant, and even if it's not directly about AI search visibility. OpenAI and Hugging Face published a joint disclosure this week about a security incident that happened during an internal model evaluation. Here's what went down. OpenAI was running a benchmark to measure the cyber capabilities of some of their models. That included GPT, 5.6 Sol, and a more powerful pre-release model. It had reduced safety refusals enabled for testing purposes, and somehow those models found a way to break out. The models identified a zero-day vulnerability in OpenAI's testing environment, exploited it to gain internet access, it figured out that Hugging Face might have benchmark solutions, found a way to access Hugging Face's production infrastructure, and used stolen credentials and additional vulnerabilities to get what they were looking for. OpenAI called it, quote, an unprecedented cyber event, end quote. They were clear that the models weren't trying to escape, but they were hyper focused on solving the evaluation task and then went to the absolute extreme lengths to do it. But the result was an actual real world security breach. Hugging faces on security teams detected it and then stopped it. Both companies went public with what happened, and I think that's actually the right call. You're probably thinking, well, why does this matter? Well, first of all, it's this interesting, and I thought, oh wow. But um, there are two reasons. One, we talk a lot about AI capabilities expanding. This is a concrete example of what that actually looks like in practice. These models can chain complex attack paths, exploit novel vulnerabilities, and then operate autonomously over long time horizons. This is not a future thing that we've just sat down and imagined that this is actually happening right now, and it's kind of scary when you think about it. Do the other reason the AI ecosystem is moving fast, and incidents like this are gonna shape how these tools are deployed, how they're governed, if at all, which they should be, but how they're governed, and ultimately what capabilities end up in the public-facing products that we use for search and content. I really think that's something that we ought to pay attention to. And the last one, Anthropic announced a new feature called Record a Skill in Cloud Co-Work. This week it's available on Pro Max and Teen Plans, and I want to call it out because I know that so many of us in marketing, we are really big on Cloud Co-work. So basically, with this, you can record your screen while you do a task, talk through it as you go, and then Claude turns it into a reusable skill. It can then run independently. Some people in the replies about this were worried about what this means for job security. That is an absolute real conversation that we need to have. I don't want to brush that off. But from a workflow perspective, just thinking about all the things that I do for the podcasts and my clients, this is really interesting for content teams and marketing operations. The ability to codify a repeatable process, whether that's an audit workflow or a content publishing process or a reporting routine, and then hand it to an AI agent to execute. That is a real capability shift. I'm gonna be watching how this one develops over time, probably testing it out here pretty soon. Might report back, we'll see. OpenAI has had something similar for codecs since June, but this is anthropic just catching up to the agentic side, and I wanted to call it out because again, I know a lot of us are a lot of us are pro-Claude these days. Okay, that's it for this week. Again, there was a lot that happened. If I had to pick one story to watch going forward, it's the publisher revolt. Because whatever resolution emerges from that dynamic, that's gonna reshape how the open web relates to search, and it absolutely has downstream effects on everything that we do for AI Search Visibility. I am Cassie Clark, an AI Search Visibility Consultant. If you want to understand where your brand actually stands in AI generated answers right now, the place to start is an AI Search Visibility audit. You can find more at CassieClarkmarketing.com. That's it for today. I will see you all on Tuesday. Until then, stay visible.