Google Q2 Revenue Hits $119.8B Record as EU Fines It €890M

Google's Q2 2026 earnings delivered its biggest quarter in company history: overall revenue hit $119.8B (+24% YoY), with Search ad revenue rising 14.5% to $81.6B. However, as Search Engine Journal notes, Search-specific revenue grew 17% to $63.27B — the first easing of acceleration after four quarters of consecutive expansion, with Google crediting retail and Gemini integration as the main drivers.

Behind the headline numbers, Alphabet posted a $5.85 billion negative free cash flow, driven by massive infrastructure spending on AI chips and data centers. As Marie Haynes highlights from the earnings call, Google Network (AdSense) revenues actually dipped 1%, while Cloud revenues — fueled by AI infrastructure and Gemini model usage — surged 87% year-over-year. That gap signals where Google's strategic bets are being placed.

CEO Sundar Pichai acknowledged on the call that Google needs Gemini 4 to remain competitive at the AI frontier, singling out coding as a key weakness while Gemini 3.5 Pro remains delayed. Search Engine Journal separately flags a credibility issue from the earnings call: while Google reported Search revenue to the decimal, its claims about web traffic driven by AI search were described in vague, unverifiable terms — widening the transparency gap that publishers and SEOs are increasingly frustrated by.

The European Union issued a €890 million (~$1 billion) fine against Google this week for breaching the Digital Markets Act through anti-competitive behavior in Google Search and Google Play — a landmark regulatory penalty. Separately, the EU issued binding rules requiring Google to share anonymized search data with rival search engines and AI chatbots, a structural intervention that could reshape the competitive landscape over time.

The DMA pressure is already producing visible product changes: Google appears to be rolling out updates to its hotel search results in the EU to comply with European Commission requirements, despite publicly opposing the mandate. Search Engine Journal's weekly SEO Pulse roundup frames these developments alongside slowing Search revenue growth as converging signals of institutional pressure on Google's search dominance.

Search Engine Roundtable reports that several large publishers are strongly considering blocking Google Search entirely, driven by a sustained decline in referral clicks that has made the content-for-traffic exchange no longer economically viable. The implied logic: if Google consumes their content for AI Overviews and AI Mode without delivering meaningful click-throughs in return, removing access may be preferable to giving content away for free.

This development sits within a broader context captured in SparkToro's newsletter: 68% of Google searches now end without a click, with user journeys fragmenting across Google, ChatGPT, YouTube, Reddit, social feeds, and newsletters. The publisher revolt — if it escalates — could fundamentally alter what content Google can crawl and surface, creating a structural tension between Google's AI ambitions and the health of the open web it depends on.

During its Q2 earnings call, Google responded with a claim that AI search sends billions of clicks to websites every week, but as Search Engine Journal notes, these figures were presented without verifiable methodology — making them difficult for publishers to act on or trust.

A U.S. federal court dismissed Google's lawsuit against SerpApi, which alleged the data aggregation company illegally scraped Google Search results. U.S. District Judge Yvonne Gonzalez Rogers dismissed the DMCA claims, ruling that blocking scrapers from accessing public search results does not constitute copyright circumvention on its own. However, Google was granted 21 days to refile if it can demonstrate authorization from copyright owners — keeping the door open for a revised legal challenge.

The case has significant implications for the SEO tooling ecosystem that relies on SERP data providers, and signals that the legal framework around search data access remains unsettled.

Following our June 29th warning that Cloudflare's AI crawler management system could inadvertently block Googlebot depending on configuration — and the note that default settings were scheduled to activate September 15 — TechSEO Vitals now reports the situation has escalated significantly. Cloudflare's most restrictive rule is now applying across all crawlers, including Googlebot, and critically, the enforcement is happening at the firewall level before Googlebot can even read the site's robots.txt file. This means affected sites may be inadvertently de-indexed or starved of crawl without any warning in Search Console. Any site owner who activated the single-click AI-blocker at any point over the past year should audit their Cloudflare settings immediately — the September 15 default deadline is no longer the only risk window.

Google updated its official Crawl Budget documentation this week, adding a key clarification: every site starts with the same default, conservative crawl capacity limit, regardless of size or authority. The update also improves terminology consistency throughout the document — a signal that Google is formalizing guidance that was previously implicit.

Separately, Google's John Mueller clarified a robots.txt edge case where Googlebot may ignore crawl directives in ways that could negatively impact SEO. Mueller's explanation confirms there are specific conditions — not widely documented — under which Googlebot proceeds despite robots.txt restrictions. SEOs managing complex site architectures should investigate this quirk.

Search Engine Roundtable reported signs of an unconfirmed Google search ranking update around July 18th, with third-party tracking tools showing notable volatility and elevated community chatter over the weekend. Google has not confirmed the update. SEOs seeing unusual weekend shifts in rankings or analytics should flag this period when investigating anomalies.

Two significant data findings emerged this week about Google's own AI Mode. First, Google itself is the second most-cited domain in AI Mode responses, driven by Google-hosted Business Profile cards and Product Knowledge Panels — a finding with direct implications for local SEO and brand visibility strategy. Second, a new analysis found that ads appear on nearly one in three commercial keywords in AI Mode, but traditional advertisers were rarely among the cited organic sources. This creates a split dynamic: paid visibility is accessible but organic citation in AI Mode follows different logic than conventional SEO.

Meanwhile, Google is adding more AI search data to Search Console and Merchant Center, but clicks, individual query data, and key reporting details remain absent, leaving SEOs without the metrics needed to properly measure AI search performance. The data gaps are widening the distance between what Google reports and what practitioners can independently verify.

Liz Reid's public comments on AI search continue to generate coverage — following her June 29th remarks about publisher visibility and personalization, this week Reid discussed 'persistent autonomous search': a mode where search operates continuously in the background, proactively surfacing information based on user context rather than responding to discrete queries. A related Google patent describes six specific triggers for this always-on behavior, signaling active development. On the infrastructure side, Google confirmed that Gemini 3.5 Flash-Lite is now rolling out in Google Search for agentic search experiences — described as Google's fastest and most cost-effective 3.5-class model at 350 output tokens per second. Google also released Gemini 3.6 Flash and 3.5 Flash Cyber alongside it.

Building on our June 29th coverage of the Top Stories carousel launching inside AI Overviews, new data this week shows the broader scale of AI Overviews' local search dominance: AI Overviews are now outpacing local packs across 15+ markets, making them the primary local search surface for many queries. Location pages seeking visibility in AI-generated answers must address proximity signals, structured hours and services data, and consistent NAP information across third-party sources. Google is also rolling out Top Stories carousels within AI Overviews for US mobile users — extending the feature beyond the initial rollout documented last month. On citation UI, Search Engine Roundtable observes that both Google AI Mode and ChatGPT are testing bottom-anchored citation card formats — Google anchors users to citation cards at the bottom of answers, while ChatGPT loads citation overlays on click.

Google made a significant public clarification this week: many Search Console 'error' reports are not actual problems, and the tool was never designed as a checklist of fixes. This statement, reported by Search Engine Journal, has immediate implications for how SEO teams prioritize technical work and communicate with clients. The framing suggests practitioners have been over-indexing on Search Console status messages, potentially burning dev resources on issues that carry no meaningful ranking impact.

This guidance inverts a common SEO workflow where any Search Console error triggers a remediation ticket. Google's position is that practitioners should use judgment about whether a flagged issue actually affects user experience or crawl efficiency — not treat every report as an urgent fix.

Steve Toth's SEO Notebook newsletter highlights research from Koray Tuğberk Gübür suggesting that page layout, UI components, and visual structure now factor into Google's ranking signals — not just textual content. A programmatic site reportedly jumped from 3.47 million to 4.53 million clicks after repositioning a calculator into the centerpiece of the page, framing it as the primary interactive element.

Separately, a free tool was highlighted that flags discrepancies between rendered HTML and source HTML across word count, links, schema markup, and meta tags — a practical utility for diagnosing rendering gaps that might cause Google to evaluate a page differently than intended.

A Semrush analysis of 1,094 ChatGPT topic categories found that only 15.2% had a clear brand owner — defined as a brand appearing in at least four of five related prompts and leading its runner-up by a meaningful margin. That means 85% of topic categories show no dominant brand presence in AI responses, and appearing in one prompt does not predict consistent appearance across related questions in the same category.

Kevin Indig's Growth Memo investigates whether topical authority translates to AI search, using real brand examples including Forbes, HubSpot, and ClickUp. His finding: broad topical focus works for a few mega-brands like Forbes, but most brands achieve better AI visibility by concentrating content on their core topics. The implication is that topical authority still matters in AI search — but the mechanism differs from traditional Google ranking, operating through training data density and category association rather than PageRank.

Search Engine Land adds a complementary angle: category framing in third-party content — how external sources associate a brand with specific query categories — drives AI recommendation more than entity strength or schema markup alone. Brands need to align their category associations with the language their customers actually use.

Following the June 29th coverage of Google Business Profile reviews disappearing and July 6th's confirmation it was a dashboard display bug, a new related issue emerged this week: review replies are not displaying for many Google Business Profiles across all accounts — a separate, widespread bug that local SEOs should proactively communicate to clients. Beyond the technical issue, new research reported by Search Engine Journal reveals that AI answers about business locations are frequently inaccurate, citing wrong postcodes, false closure claims, and invented services — a finding that reframes the GBP reliability question beyond just reviews. Amsive adds that a five-star Google Business Profile is no longer the single source of truth for AI search: AI systems evaluate brand reputation across review sites, forums, social platforms, videos, and third-party sources when framing brand presence and guiding consumer decisions.

Search Engine Journal argues that simply counting AI citations is a flawed proxy for AI search visibility — citations indicate a source was referenced, but not that a brand was actively recommended. The gap between these two outcomes is growing as AI systems become more nuanced in how they present information versus what they actively endorse.

A Search Engine Journal opinion piece draws a sharp parallel to Google's 2011 'not provided' keyword data removal, urging SEOs to build measurement frameworks they own rather than waiting for platforms to provide ROI proof. The argument: just as SEOs adapted to keyword data opacity after 2011, they must now develop independent measurement approaches for AI traffic — because AI platforms are showing the same reluctance to provide actionable attribution data.

New research highlighted by Search Engine Journal finds that 70% of top retailers are effectively invisible to AI-powered agentic commerce systems — the emerging class of AI agents that can search, evaluate, and complete purchases autonomously. The core finding: future e-commerce SEO success will hinge less on content quality and more on structured data and machine-readable product information that agents can parse without a visual rendering layer.

Search Engine Journal separately argues that AI browsers like OpenAI's cancelled Atlas project were always a workaround, and that the correct architectural response is building websites that AI agents can read directly — through APIs, structured data, and clean machine-readable markup — rather than retrofitting visual interfaces for AI navigation.

Google has officially confirmed that Local Services Ads (LSA) are transitioning into the main Google Ads platform, replacing the standalone LSA interface with a new specialized Performance Max campaign type optimized for pay-per-lead goals. Changes begin rolling out from August 2026. This announcement coincides with a previously reported change also taking effect August 31: Local Inventory Ads (LIA) will be enabled by default for all Shopping campaigns — covered in our July 13th issue — with the additional detail that the ability to disable LIA through the Google Ads API is also being removed. Local businesses and agencies should treat both deadlines as concurrent priorities when reviewing their Google Ads configurations.

Following last week's coverage of ChatGPT Ads' addition of Sales ROAS metrics and geo-location controls, OpenAI continued expanding its advertising platform this week with conversion value tooltips in the ChatGPT Ads Manager, breaking attribution data into View Through Attribution (VTA) and Click Through Attribution (CTA) columns — further progress toward standard performance measurement frameworks. Separately, ChatGPT has licensed Yelp reviews, giving OpenAI access to a significant corpus of real-world business review data with direct implications for how AI search surfaces local and business recommendations — a development particularly relevant for local SEO practitioners monitoring brand presence across AI systems.

The Similarweb finding that 68% of Google searches end without a click — first highlighted in our July 6th issue alongside the AI traffic misattribution problem — is now the organizing frame for SparkToro's newsletter this week, which extends the story beyond the raw statistic. The new framing shifts from measurement to customer journey mechanics: users now move fluidly across Google, ChatGPT, YouTube, Reddit, social feeds, newsletters, and branded searches — sometimes clicking, sometimes absorbing information and returning days later with intent already formed. The practical implication for SEO strategy is to expand beyond ranking to include presence in the environments where awareness forms, even when those touchpoints don't generate direct attribution.

Search Engine Journal makes the case that YouTube has become a meaningful source of AI Overview citations, making creator partnerships more strategically valuable than traditional view-count metrics suggest. The argument: the downstream value of YouTube content — traffic, AI citations, and conversions — is not captured in standard video analytics dashboards, leading teams to underinvest in a channel that increasingly influences AI-generated search results.

Moz reinforces a related structural shift, arguing that search is adopting creator-platform logic akin to TikTok, where content virality and creator authority are increasingly displacing traditional ranking signals as traffic drivers. This explains the disconnect many SEOs report: stable rankings but declining organic traffic.

Several Google Ads platform changes landed this week. Google released version 25 of the Google Ads API — a major release covering campaigns, planning, reporting, and conversions, with both breaking and non-breaking changes. Google Ads also moved 'Missed Growth Opportunities' out of Labs and into the main Recommendations tab, renaming it 'Missed Opportunities' and making estimated missed clicks, conversions, and revenue more prominently visible alongside standard optimization recommendations.

On policy: Google Ads introduced a 6-month time limit on policy violation appeals, effective July 21, 2026. Advertisers can no longer appeal decisions made more than six months prior directly from their accounts. Additionally, Google has officially retired the Related Search format for AdSense auto ads effective August 6, 2026, with publishers already notified by email.