If you only caught the headlines from Alphabet's April 29 earnings call, you saw a clean beat. The coverage led with the growth rates (they're all in the first table below) and the 2 billion monthly AI Overviews users. The consumer-tech write-ups stopped there, and most SEOs nodded and moved on. I think you should take a second look.
The transcript and the slide deck say a lot more: where Google is taking its AI strategy, how it will make money from AI answers, which verticals already pay, and how tight compute looks going into 2027. I count ten signals in there that matter if you do SEO or AEO work. Each one is sourced to the primary transcript or the Alphabet press release, and each comes with the change I'd make to your content plan for the second half.
A first-quarter call is usually where a direction that runs over several quarters becomes readable. This one says more than most, because the AI-first Search transition is now producing revenue you can measure at scale. Every section below ends with a tactical move.
The consolidated numbers
| Line | Q1 2026 result, change year over year |
|---|---|
| Consolidated revenue | $109.9B, up 22% |
| Search & Other Advertising | $60.4B, up 19% |
| YouTube ads | $9.9B, up 11% |
| Google Cloud | $20B, up 63%, its first quarter above $20B |
| Subscriptions and devices | $12.4B, up 19% |
| Net income | Up 81% |
Revenue came to $109.9 billion, per Alphabet's first-quarter CEO letter. The 22% growth rate is Google's fastest in a single quarter in roughly four years, and it arrived at a size where most analysts had stopped expecting any acceleration.
Look hardest at Search & Other, since that's where cannibalization would show up. If AI Overviews and AI Mode were eating Search revenue, the line would shrink. It grew 19%. Across the business, AI answers are expanding revenue, so what looked like Google defending Search is now its growth story.
Put that 19% next to 2 billion-plus AI Overviews users and you get a company that has made AI answers add to its ad revenue. I don't see that reversing. If your plan assumes AI answers will retreat once the ad model breaks, you're planning for a scenario that just failed to show up in the largest quarter Google has ever reported.
The Network revenue decline SEOs missed
The line I most want you to remember sat in the press release. Google Network revenue, what Google earns placing ads on third-party publishers through AdSense, AdMob and Ad Manager, fell 4% year over year to $6.97 billion.
Alphabet frames its release for investors, which is exactly why the one line that looks bad deserves your attention. Publishers should have it on a whiteboard, and so should agencies and any SaaS whose growth runs on content.
I call it a two-ledger quarter. Google's own Search grew 19% while the open-web network Google sits in the middle of shrank, so on Google's ledger AI answers bring money in, and on the open web's ledger they take it out. AI Overviews is squeezing open-web ad revenue at the exact moment Google's owned monetization speeds up.
That's the mechanism behind the Chartbeat traffic decline covered by Axios, where the smaller the publisher, the harder the fall:
| Publisher size | Google search traffic, twelve months ending December 2025 |
|---|---|
| Small publishers | -60% |
| Medium publishers | -47% |
| Large publishers | -22% |
Publisher-side monetization is contracting, and Google's side is capturing the difference.
For a SaaS founder, the "publish more content on your blog to capture traffic" playbook now runs against pressure that won't reverse. Ad dollars that used to flow through third-party publisher inventory increasingly land inside AI Overviews answers, and the first-quarter numbers effectively confirm it. So redefine content ROI for the second half. Your blog's traffic is the receding asset, and citation share inside AI Overviews answers is the rising one. You still need the articles, since they're what gets cited, but budget against the second number.
Capex: $35.7B in a quarter, and 2027 will "significantly increase"
Spending tells you how long Google expects this to last. That $35.7 billion was capex for the first quarter alone. Google then raised its full-year range, and CFO Anat Ashkenazi told analysts that 2027 capex will "significantly increase" from that baseline.
| Capex signal | Figure |
|---|---|
| Capex, Q1 2026 | $35.7B |
| Full-year 2026 guidance, previous | $175-185B |
| Full-year 2026 guidance, raised to | $180-190B |
| 2027 capex | Will "significantly increase" (CFO Anat Ashkenazi) |
| TPU 8i | 80% better performance per dollar than the prior generation |
| TPU 8t | 3x the processing power of Ironwood |
You're looking at a compute-constrained company signaling a build that runs for years. With the new TPUs, Google is deliberately widening its lead on inference cost, and every Gemini answer for the rest of the decade will run on infrastructure that's under construction right now.
For your plan, that makes AI Overviews and AI Mode permanent products, scaled aggressively, with unit economics that only improve. Model your pipeline on AI answers retreating and you're betting against the largest infrastructure investment in Google's history.
In this single fiscal year, Alphabet is committing more capital to compute than Meta and Microsoft combined put into their AI infrastructure programs in 2024. That boxes in OpenAI, Anthropic and Perplexity, who now face a compute cost gap that hardware alone can't close. Over the next 12 months they'll have to differentiate on data or workflow, or through verticalization, since raw model capability won't do it.
Lily Ray called out the user-count framing during Pichai's presentation:
I think a big part of it was to show that "search is evolving with AI" and to say "we have the most daily users of AI products" (well yeah, if you group AI Overviews into Search with no way to opt out)
Lily Ray 😏@lilyraynycJul 23, 2025She has a point. "AI Overviews users" is a somewhat manufactured number, because AI Overviews sits in Search results with no opt-out, and I'd discount the 2 billion for it. None of that changes the compute being built behind it, or the ads already shipping inside AI Overviews answers on U.S. mobile.
Cloud backlog doubled to $462B
| Cloud metric | What Google reported |
|---|---|
| Cloud backlog | $462B, up from roughly $240B a quarter earlier |
| Cloud operating margin | 32.9%, against 17.8% the year before |
| Cloud GenAI revenue | Up approximately 800% year over year |
Why should you care about Google's cloud business? Enterprise adoption of Gemini through Vertex, Gemini Enterprise and BigQuery is what keeps Google's AI answer surface going. With Cloud GenAI revenue up about 800% and enterprise seats compounding at that rate, the index AI Overviews and AI Mode retrieve from keeps growing in coverage and quality.
At the first-quarter run rate, the $462B backlog is roughly six years of Cloud revenue, which is what a compute-constrained inference market looks like from Google's side. The demand exists and the supply has to be built. That's why I believe the 2027 capex ramp, and why you should expect Google's AI answers to cover more and respond faster through the rest of the decade, with more room for ads.
Gemini Enterprise: 8M+ seats across 2,800+ companies
Pichai and Ashkenazi got specific about enterprise Gemini, right down to customer names.
| Enterprise signal | Figure |
|---|---|
| Paid enterprise Gemini seats | 8 million+ |
| Companies behind those seats | 2,800+ |
| Named on the call | Bosch, Citi Wealth, Merck, Mars, American Express, Vodafone |
| $100M-$1B deals | Doubled year over year |
| Cloud customers processing 1T+ tokens annually | 330 |
| Cloud customers that hit 10T | 35 |
This under-the-radar B2B moat feeds straight back into Search. Enterprise usage creates the retrieval and grounding signals that improve Gemini, a better Gemini drives AI Overviews adoption, and that adoption drives Search revenue. It's a flywheel whose shape you can now see in disclosed numbers, and it ends in the answers your buyers read.
Ad monetization inside AI Overviews and AI Mode
The quote SEOs missed was short. Google is running ads inside AI Overviews for U.S. mobile and actively testing them inside AI Mode. Pichai added that Search & Other growth came primarily from retail and finance, with health also contributing.
So if you still see AEO advice calling AI answers ad-free, it's out of date. Retail, finance and health are where ad units are landing best in the second and third quarters, which lines up with BrightEdge's data on how often AI Overviews appear:
| Vertical | Queries showing AI Overviews (BrightEdge) |
|---|---|
| Healthcare | 88% |
| Education | 83% |
| B2B tech | 82% |
One caution: BrightEdge counted where AI Overviews appear, which is a separate question from where the ads go, so treat the match as a hint.
If your SaaS sells into those verticals, your citation inside an AI answer now competes directly with paid placements. Getting cited becomes your defense against a paid slot showing your competitor's product.
Barry Schwartz walked through the specific AI Overviews product shifts:
Google AI Overviews & AI Mode gain preferred sources, plus new perspectives carousel and highly cited labels https://t.co/C3f82Vv7Of https://t.co/bYF0J1sxjt
Barry Schwartz@rustybrickMay 27, 2026
The return on citation share now sits in preferred sources, the perspectives carousel and highly-cited labels, the spots where third-party editorial trust turns into visibility inside the answer.
You're getting AI traffic. But do you know where it comes from?
RankControl credits every visit to the assistant that sent it: ChatGPT, Perplexity, Claude, Gemini, Copilot, or Grok. Full source attribution, next to your Google traffic.
API token throughput: 16B tokens a minute
Of everything disclosed, this is one of the two or three numbers I'd rank highest for AEO. Pichai put direct customer API throughput at 16 billion tokens per minute, up from 10 billion the prior quarter. That's a 60% jump.
Gemini Enterprise and third-party apps on Vertex consume those tokens, and so do AI Overviews and AI Mode. At that scale, one well-optimized page of yours can be pulled tens of thousands of times per minute at peak query load.
Small structural work is worth far more to you now. Schema and semantic-triple structure matter at a completely different scale than they did in 2024, and so do freshness signals. A single fix, whether it's a descriptive URL or moving the answer into the first 30% of the page, compounds across every query that pulls your page. Correct schema works the same way.
Search latency and response cost efficiency
Sundar's CEO note had two efficiency numbers in it. Search latency is down more than 35% over five years. AI response costs are down more than 30% since the Gemini 3 upgrade.
Quietly, that's Google telling you the marginal cost of an AI answer is falling fast. Cheaper answers mean AI Overviews keeps spreading into more query types (see Semrush's 71% AIO expansion on commercial-intent queries).
It also ends the "too expensive, so AI answers will retreat" argument. You've probably seen the comment-thread version: Google will pull AI Overviews coverage back once the CFO looks at unit economics. That's now empirically wrong. The CFO did look at unit economics, and Google's response was to raise capex guidance and confirm 2027 will grow significantly from there.
Product-level callouts: Gemma 4, Deep Research, Personal Intelligence
The CEO remarks listed a run of product callouts, and most are background if your job is getting cited:
| Product | Callout in the CEO remarks |
|---|---|
| Deep Research agent | Added MCP support and native visualizations |
| Personal Intelligence | Rolled across Gemini, AI Mode and Chrome |
| Lyria 3 | Generated 150M+ songs |
| Nano Banana 2 | Processes 1B images in half the time of v1 |
| Gemma 4 | 50M+ downloads in weeks; 500M+ open-model downloads total |
What matters for you is Deep Research and Personal Intelligence, along with MCP-augmented workflows, because that's where users do long multi-turn research. Content there has to survive being re-queried and extended, then combined with other sources over a longer session.
My read is that comprehensive comparison and analysis pages (2500+ words, a clear methodology, rich internal links) will beat short listicles built for single-answer AI Overviews retrieval in Deep Research.
That's the year's second big format-diversification signal, after Seer's 30% month-over-month drop in listicle citations. If your pipeline ships only listicles, you're over-indexed on a receding format and under-indexed on the one the fastest-growing Google surfaces prefer to retrieve. On the first-quarter signals, this is the mix I'd call reasonable for the second half:
| Format | Share of the mix | Why |
|---|---|---|
| Listicles | 30-40% | Still the largest single AI Overviews citation share, 21.9% per Wix Studio |
| Analysis and methodology pieces | 30-40% | The Deep Research surface |
| Comparisons and use-case pages | 20-30% | LinkedIn amplification and B2B tech AI Overviews coverage |
We'll show you exactly where your brand stands in AI search.
No commitment. $0 due today, cancel anytime. See how ChatGPT, Perplexity, Claude, Gemini, Grok, and Google AI Mode talk about your brand today.
Paid subscriptions cross 350M, Gemini BigQuery grew 30x
The subscription news is solid, but it isn't why this section is here. Paid subscriptions crossed 350 million in total, driven by YouTube and Google One, and YouTube Music and Premium had their largest quarterly subscriber increase since June 2018.
The number I'd underline is Gemini-powered BigQuery workflows, up 30x year over year and the fastest enterprise adoption signal on the call. Gemini is doing structured-data analytics at very high scale inside enterprise data warehouses. Its retrieval quality on structured business questions is compounding with it.
For you, that means schema and structured data on your public pages do two jobs. They're a citation signal for AI Overviews, and they're part of what Gemini is trained and evaluated on across the enterprise BigQuery footprint. If you make one technical move in the second half of 2026, make it structured data, the highest-ROI option I can see.
What this means for SEO strategy
Six operational changes fall out of those ten signals:
| Move | Why |
|---|---|
| Budget for AI Overviews citation share as a first-class KPI | $180-190B in capex says AI Overviews is permanent; 19% Search revenue growth says it monetizes |
| Prioritize the six ad-monetized verticals: retail, finance, health, healthcare-adjacent, education, B2B tech | Getting cited there defends against a paid competitor slot in the same real estate |
| Treat Gemini Enterprise and BigQuery adoption as a leading indicator for Gemini quality in your category | Verticals with heavy Gemini Enterprise adoption will see faster citation-quality gains in AI Overviews answers over the next two quarters |
| Ship semantic-triple structure and schema on every net-new page | HubSpot's 642% citation lift experiment maps to what the 16B tokens/minute pipeline can now retrieve at scale |
| Invest in Deep Research and Personal Intelligence-compatible content | Longer, methodology-forward pieces with clear internal navigation |
| Treat the Network revenue decline (-4% YoY) as the confirmed answer to "is publisher traffic coming back?" | It is not, and the compression will accelerate as AI Overviews ad units expand |
If you can only fund two right away, start with the first and the last. Once citation share is a KPI you report on, most of the rest follows, and the Network line stops you from waiting on publisher traffic that isn't coming back.
Our content engine covers the page-level part: it drafts with schema and an answer-first structure, with headings written as questions. For the strategic frame across all six major engines, see The 2026 State of AI Search: RankControl's Annual Report.
As I read it, the first quarter is when Google made the AI-first results page permanent, with no sign of hedging. The SEO strategies that survive the second half will treat the transcript as a confirmed permanent shift and adjust the pipeline to match. Rank still matters. Citation share and AI Overviews monetization matter more now, and Google's capex tells you which one it expects to matter for the rest of the decade.
Read as a to-do list, the transcript says you track citation share weekly and rebalance toward the six ad-monetized verticals, and every page you ship goes out structured.
The next checkpoint is the Q2 call, scheduled for July 22, which the market will read against these first-quarter baselines. If capex holds and AI Overviews users cross 2.5B while Search revenue keeps double-digit growth, the picture stays intact. Miss on any of them and I'd expect the AI search conversation to shift meaningfully in Q3.
Betting against Google's own guidance at this scale of capex is a specific position with a specific risk profile. I think most SaaS founders are better off assuming the trajectory holds and shipping the six moves above. The upside of being early on citation share dwarfs the downside of over-preparing for a permanent SERP shift Google has now confirmed twice.
Know exactly what AI says about your competitors.
RankControl's Recon Agent monitors competitor citations across ChatGPT, Perplexity, Claude, Gemini, Grok, and Google AI Mode. See where they show up and you don't.
