AdSense Fell 4% While Google, Meta, and Amazon Split $150B in Q1 Ads

Google, Meta, and Amazon reported a combined $150+ billion in Q1 2026 ad revenue, with Google Search and Other at $60.4 billion (+19% YoY), Meta advertising at $55 billion (+33% YoY), and Amazon Ads at $17.2 billion (+24% YoY). In the same quarter, Google Network, the line that captures AdSense, AdMob, and Ad Manager, fell 4% YoY to $6.97 billion. The concentration gap, not the headline growth number, is the budget signal worth auditing this week.

The line that broke from the rest

The four big ad companies all posted growth. Three accelerated. One went the other way.

Google's first-party search and YouTube revenue grew double digits, with paid search up 31% on a query-volume tailwind. Meta accelerated to 33%, the fastest growth rate it has posted in years. Amazon Ads cleared $17.2 billion in a single quarter, with the trailing twelve months crossing $70 billion. Microsoft's search and advertising line grew 12%, though the company still does not break out a hard dollar figure.

Then there is Google Network, which captures AdSense, AdMob, and Ad Manager third-party inventory. That bucket dropped to $6.97 billion, down 4% year-over-year. On the ExchangeWire MadTech Podcast, COO Lindsay Rowntree called the decline "kind of a concern for the open web" and said it would likely keep falling.

That is the line worth staring at. The pie grew. The piece feeding non-owned-and-operated inventory shrank. If you are reading a Q1 summary that frames this as "Big Tech keeps winning," you are reading a headline that buried its own most useful detail.

Where the growth actually came from (the AI part the press releases gloss)

Each of the three winners attributed acceleration to AI ad systems they shipped in the last 18 months, not to a macro ad-market recovery.

Meta said 8 million advertisers are now using at least one of its AI ad creative tools, up from 4 million at the end of 2024. The company also reported video-generation users saw conversion rates roughly 3 percentage points higher in tests. Average price per ad rose 12%; impressions rose 19%. The 12% price line, not impression growth, is where Andromeda-class systems start to show in the P&L.

Google paid search grew 31%, separate from the headline 19% Search and Other number, which combines search and a few smaller lines. Query volume rose roughly 20%, which Google attributes to AI Overviews now serving 2+ billion monthly users. Forrester's own quarterly tracking framed this bluntly: GenAI is rebuilding search, and Google is still winning.

Amazon expanded Creative Agent, its agentic ad-creative tool, into Canada, France, Germany, India, Italy, Spain, and the UK. CEO Andy Jassy said on the call that AI tooling is expanding the advertiser base, particularly with sellers who never had agency support.

The pattern: AI tooling is functioning as an audience-expansion lever inside walled gardens, not as productivity that flows back to open-web inventory. The advertiser who never bought before now spends inside Meta or Amazon. They do not show up on a Criteo retargeting line.

The 90/10 ownership split hiding inside Google Network

One detail in the Alphabet disclosure worth surfacing. Google Network revenue is reported as a single line, but the company has previously indicated that owned-and-operated inventory accounts for the dominant share of the bucket. The piece that actually represents independent publishers, the long-tail AdSense placements on small sites, is a smaller subset still.

If you sell display through Google's mediation, or you run an ad-arbitrage play on AdSense pages, the 4% segment decline is directional only. The dollar decline on independent publisher inventory specifically is almost certainly larger. The visible -4% is a blended number that includes Google's own large surfaces. The number you actually care about, if you are an SMB publisher, is somewhere behind it.

That is the audit. Not "did we lose budget this quarter." Rather, "did we lose budget in the long-tail bucket that is shrinking under us, and are we still benchmarking against Q1 2025 splits we should have re-baselined three quarters ago?"

What this does to mid-tier ad networks

Per the PPC Land summary of Q1, the top four DSPs hold 85% of global programmatic share, and two of every three pounds in the UK ad market now flow to Google, Meta, or Amazon. The mid-tier (Criteo, StackAdapt, Outbrain, Taboola, every regional DSP) is splitting the remainder of a shrinking slice.

Two things break if you are a mid-tier network buyer right now:

First, your performance reporting probably still benchmarks against last year's Q1 share of total demand. The share of available budget you can intercept has shrunk faster than your fill rate, which means your CPM looks healthy while your absolute revenue from category leaders erodes underneath. The fill-rate metric was always a lagging proxy. In a year of concentrating spend, it lags worse.

Second, the AI ad-creative gravity is on the walled-garden side. An advertiser using Meta's video-generation tool is not exporting that creative to a Criteo retargeting buy. The creative pipeline that used to fund cross-network campaigns is increasingly built inside whichever wall the advertiser started with. On paper, cross-network reach still sounds like an upgrade. And sometimes it is. But the creative input cost is going up for buys outside the walls that own the tools.

And to be fair, this isn't entirely new. Walled gardens have been winning share for a decade. It just feels a lot less forgiving when the open-web line item finally turns negative on a year-over-year basis instead of merely decelerating.

The mid-tier dollar audit, in three queries

For paid-ads managers running diversified spend, the actionable check is straightforward:

Pull a 12-month trend on your non-Google, non-Meta, non-Amazon spend in absolute dollars, not as a percentage of total. Compare against the 12-month trend on revenue attributed to those same channels. If revenue is growing in line with spend, mid-tier channels are still earning their slot. If revenue is flat or declining while spend is steady, you are funding share donation to walled gardens at retail price.

For agencies pitching cross-network strategies as a "Google/Meta hedge," the macro case got harder this quarter. The case is still real in specific verticals (B2B intent buys, contextual placements in regulated categories, retail-adjacent CTV not yet swallowed by Amazon DSP), but "diversify your platform mix" as a general thesis ran into a 4% AdSense number that says the market is consolidating, not diversifying.

For publishers funded by AdSense and Ad Manager, the 4% drop is the second consecutive quarter of contraction. Rowntree's "every single publisher has to think about new ways" framing isn't a hot take, it is increasingly the only available read. Direct sales, paid memberships, and licensing deals (Reddit booked $39M in AI licensing this quarter alone) are not optional revenue lines anymore. They are the replacement line.

If you covered the related shift in OpenAI's pricing model, the same dynamic shows up there: ChatGPT Ads quietly opening to self-serve advertisers creates yet another walled-garden destination for budgets that used to land on open-web inventory.

One thing I would watch next quarter

Omnicom is running live media buys through AI agents that bypass DSPs and SSPs entirely. That is currently a footnote in most Q1 narratives, but it points at the next concentration step. Not just walled gardens taking share from the open web, but agent-mediated buying skipping over the programmatic infrastructure layer altogether. If that workflow scales through Q3, the next "mid-tier in trouble" line item is the SSP, not the DSP.

From what I have seen, the firms most exposed aren't the smallest ones. They are the ones whose entire pitch depends on programmatic plumbing remaining the dominant routing layer for digital ad spend. That plumbing is being routed around in real time by the largest agencies on the buy side. You can hedge it in your next planning cycle, or sit with a flat-revenue, flat-spend chart for two more quarters and pretend the concentration story is still about somebody else's P&L.

Notice Me Senpai Editorial