AI search revenue surges as Google Network keeps bleeding

What happened

Per Search Engine Journal, Alphabet and Microsoft just printed earnings that settle a debate the search industry has been having for two years: AI features are not cannibalising search ad revenue, they are accelerating it. Google Search and Other revenue grew double digits year over year. Microsoft's Search and news advertising revenue grew faster than the broader business. Both companies attributed the lift, on the record, to AI-driven surfaces: AI Overviews, AI Mode, and Copilot integrations.

The exception is Google Network, the slice of Alphabet's ad business that pays publishers and partner sites. Search Engine Journal reports that Network revenue fell to $6.97 billion, continuing a multi-quarter slide. Owned-and-operated AI surfaces are growing. The open web of partner publishers is shrinking inside Google's own P&L.

That gap is the story. The money is moving from distributed third-party inventory into AI-generated answers that Google and Microsoft control end to end.

Why it matters for your brand

The first implication is the simplest one, and most marketing teams are still not pricing it in: AI search is now a budget category at the platforms, not a science experiment. When Sundar Pichai and Satya Nadella both call out AI search as a revenue driver on the same earnings cycle, product investment follows. Expect AI Overviews coverage to expand, AI Mode to get more aggressive ranking, and Copilot to deepen its enterprise data hooks. The surface area where your brand can appear, or fail to appear, just got bigger.

For financial services brands, this matters more than for most. Buyers researching custodians, asset managers, core banking vendors, or insurance carriers are exactly the cohort that AI Mode is built to serve: high-consideration, multi-source, jargon-heavy queries. If your fund factsheets, regulatory filings, and thought leadership are not parseable by an LLM, a competitor's are. Citation share inside an AI answer is now the equivalent of the top three blue links circa 2015. The earnings prove platforms are monetising that surface, which means they will defend and expand it.

For multilaterals and the UN system, the Google Network decline is the bigger signal. A meaningful share of UN-aligned traffic has historically come from third-party publishers, NGO syndication, and policy aggregators that sit inside ad networks now in structural decline. As that long tail contracts, the path to visibility narrows to two routes: being cited directly inside an AI answer, or being the primary source the model trusts. UNDRR, WHO, World Bank, and IMF content already ranks well in classical SEO. The job now is making sure that authority transfers into LLM citations, which depends on structured data, clean canonical URLs, and content written to answer questions rather than to perform on social.

For major industrial groups, the shift changes the brief for corporate communications. A Holcim or a Siemens announcing a sustainability commitment used to optimise for trade press pickup and Google News. Today the bigger prize is being the cited source when an analyst, a procurement lead, or a journalist asks Copilot or Gemini for the state of low-carbon cement or grid-scale storage. That requires a different content discipline: declarative claims, named figures, dated commitments, and machine-readable data tables. Press releases written for human readers do not survive the summarisation step.

For philanthropic and policy institutions, the practical effect is that earned media strategy has to be rebuilt. Placements in The Economist or Foreign Affairs still matter, but their citation weight inside LLMs is not automatic. Ford Foundation, Gates, Open Society, and Wellcome should be auditing which of their flagship reports are surfacing in ChatGPT and Gemini answers on the topics they fund. Where they are absent, the cause is almost always the same: PDFs without HTML equivalents, paywalled or login-gated research, and findings buried below executive-summary throat-clearing.

The broader content-strategy point: budgets that used to flow to display, programmatic, and partner-network buys are the budgets Google and Microsoft are now redirecting into AI surfaces they own. If your media plan still treats Google Network inventory as a meaningful awareness channel, the earnings just told you that channel is being deprioritised by the platform itself.

The signal in context

This earnings cycle is the first time both dominant search platforms have reported, in the same quarter, that AI features are net positive for ad revenue. That removes the last argument for waiting. Through 2024 and into 2025, plenty of CMOs held off on rebuilding for AI search because the platforms themselves were ambivalent: Google was protecting click volume, Microsoft was protecting Bing's growth story. Q3 and Q4 results have made the strategy explicit. AI answers are the product. Traditional ten-blue-link search is the legacy interface that subsidises the transition.

The Google Network decline is the counterweight that explains the urgency. Platforms grow AI surfaces by compressing the distributed web underneath them, and the $6.97 billion figure is the line in the financial statements where that compression shows up. For B2B brands, the takeaway is structural rather than tactical: visibility is consolidating into a smaller number of AI-mediated surfaces, controlled by fewer companies, with citation rules that are still being written. The brands that establish themselves as trusted sources inside those surfaces in the next four quarters will compound that position. The brands that wait will find the citation slots filled by competitors, trade associations, and Wikipedia.