The U.S. ad market is expanding faster than expected, but the bigger story is where that growth is landing. IAB has revised its 2026 U.S. ad spending forecast up to 12.3% from 9.5%, yet new Madison & Wall estimates show the largest platforms are capturing a disproportionate share of each new dollar.
Google, Meta and Amazon accounted for 56% of U.S. advertising revenue in 2025, excluding political advertising, according to Madison & Wall. That is up from 53% in 2024. Google increased its share from 28% to 29%, Meta from 17% to 19%, and Amazon from 8% to 9%. The rest of the market slipped from 47% to 43%.
The automation flywheel
This shift is happening as more budgets move through platform-controlled automated buying tools. Madison & Wall estimates that AI-directed or automated spending now accounts for roughly 12% of the U.S. market, up from just 2% in 2023, with a projected 27% share by 2030.
Products such as Google’s Performance Max and AI Max, and Meta’s Advantage+, show how the mechanics have changed. Marketers increasingly provide objectives, creative assets and budgets, while platform algorithms decide targeting, inventory and bidding. That can improve efficiency, but it also reduces visibility and control.
For advertisers, the central challenge is proving these systems generate incremental value rather than simply making platform-reported performance look stronger. More spending also feeds more data back into the platforms, which improves their models and attracts still more budget — a scale loop that favors incumbents.
Why it matters for your media plan
Madison & Wall’s Luke Stillman told Digiday that the big three are likely to keep outperforming the average: “It’s going to be a little higher every year for the next five because they’re just going to outperform the average every year.”
He also argued that advertiser choice alone is unlikely to break the cycle, pointing to three potential disruptions: a major change in consumer behavior, a new hardware or computing platform that changes access to digital services, or regulation that alters the economics of the market.
For brands, that means AI should be treated as an optimization layer, not an accountability replacement. A practical test framework can help:
- Run incrementality tests: Use holdout audiences or geographic experiments to isolate true lift from platform-reported conversions.
- Triangulate metrics: Compare platform data with independent brand lift, search demand, CRM and sales data.
- Review budget controls: Set clear guardrails for audience, placement and cost per acquisition, even in automated products.
- Diversify data assets: Build first-party signals and conversion feeds to reduce dependence on any single platform’s optimization loop.
The broader signal
The same dynamics could shape newer AI surfaces such as ChatGPT and AI-generated search. If those channels scale, marketers should ask whether they create real competition or simply reproduce today’s concentration economics.
For now, performance and scale continue to outweigh privacy, brand-safety and antitrust concerns in many media plans. That makes measurement discipline — not platform promises — the strongest counterweight available to advertisers.
Source: Digiday
