SASIGNAL ATLASCross-industry intelligence / Research desk
SIGNAL ATLAS / RESEARCH DESK

Alphabet, Meta and Amazon taking a widening majority of US advertising revenue, and what the rest of the value chain does about it

Scenario set · Advertising, marketing & brand · to 2030

Six alternative branches. These are subjective probabilities from the seed, not observed frequencies or investment recommendations.

Why this trend

Load-bearing assumption

That platform share gains are driven by advertiser self-selection into environments with logged-in, measurable audiences, and therefore continue regardless of antitrust remedies — rather than by the auction and inventory privileges the 2026 remedies are designed to remove.

Preconditions

Not recorded.

Concentration grinds on, the middle thins

Madison and Wall's big-three share series climbs from about 60% of North American advertising revenue toward two-thirds. Behavioural remedies in the Google ad-tech case are implemented, audited and satisfied without moving share. Holding companies stabilise around four global groups, with Publicis and Omnicom taking share from WPP and Dentsu, and sector headcount declines slowly while revenue per head rises on the denominator. Advertising inside AI assistants reaches a real but unremarkable few billion dollars and is bought as a test line, not a channel.

Mechanism

Advertisers self-select into environments where the audience is logged in and the outcome is measurable, because the corpus records that only 15.4% say effectiveness outputs primarily drive budgets - which means budget moves follow ease of buying rather than measured incrementality. The three platforms sell that ease. The independent middle cannot match it because its addressable inventory shrinks as the platforms absorb both volume and price: Meta's Q2 2026 price per ad was up 12% with impressions up 14% at the same time.

Preconditions, early indicators and assumptions
Band Width Pp

20

Calibration Basis

base_rate

Thesis

Platform share keeps climbing roughly three points a year while behavioural remedies comply on paper and change nothing about share.

Preconditions
  • Behavioural remedies are entered as a final judgment and complied with
  • No appellate reinstatement of structural relief before 2028
  • US ad market growth normalises to mid-single digits after the 2026 sporting calendar
Early Indicators
  • Indicator

    Madison and Wall big-three North American share series printing between 62% and 64% for calendar 2027

    Source

    S-16-18 Madison and Wall

    Would Be Visible By

    2028-03

  • Indicator

    The Trade Desk full-year 2026 revenue growth below 8% against an IAB US market growth figure above 10%

    Source

    S-16-20 The Trade Desk IR; S-16-09 IAB Outlook

    Would Be Visible By

    2027-03

  • Indicator

    Omnicom FY2026 10-K employee count at or below 108,000 with WPP full-year like-for-like net revenue still declining

    Source

    S-16-08 SEC EDGAR; S-16-02 WPP IR

    Would Be Visible By

    2027-03

Affected Industries
  • 16
  • 17
  • 14
  • 24
  • 12
What Businesses Should Do

Budget for a jurisdictionally fragmented, platform-concentrated buy: negotiate audit rights against the MRC auction transparency standard now while it is a differentiator rather than a requirement, and stop paying for identity infrastructure whose neutrality has already been sold.

Assumptions
  • Text

    Behavioural remedies do not materially reallocate open-web display supply to rival ad servers before 2028

    Confidence

    medium

    Load Bearing

    true

    Basis

    T-16-01 and the French Autorite de la concurrence 2021 behavioural precedent as characterised in the corpus

    If Wrong

    Probability mass moves from base to upside and the independent sell-side re-rates

Precedence Note

Distinguished from downside by the direction of total market growth, not by share: base has share rising in a growing market, downside has share rising in a shrinking one.

Would Change Our Mind

Madison and Wall printing a North American big-three share below 58% for any full year.

Interoperability actually works

The final judgment turns out to require exposure of real-time AdX bids to rival ad servers, deprecation of Unified Pricing Rules and per-bidder floors in Google Ad Manager, and Google implements rather than litigates. Independent SSP and DSP revenue re-accelerates on open-web display specifically, publisher yields improve, and the MRC auction transparency standard becomes the de facto audit benchmark. Advertisers gain enforceable transparency for the first time since the 2016 ANA/K2 report.

Mechanism

The named constraint that releases is auction privilege, and the actor that releases it is the E.D. Va. court through the final judgment. Once first look and last look are barred and per-bidder floors are available, the same impression can clear at a different price through a different pipe, which is the mechanical condition for independent demand to return to open-web display.

Preconditions, early indicators and assumptions
Band Width Pp

20

Calibration Basis

analogue

Thesis

The final judgment's real-time-bid exposure and Prebid parity requirements genuinely open AdX supply, and the independent sell-side re-accelerates.

Preconditions
  • A final judgment with operative interoperability terms, not merely a memorandum opinion
  • Google implementing rather than seeking a stay pending appeal
  • Magnite, PubMatic or The Trade Desk naming remedy compliance as a revenue driver
Early Indicators
  • Indicator

    Two consecutive quarters of accelerating open-web display revenue at Magnite or PubMatic with remedy compliance named as the driver in the earnings materials

    Source

    S-16-08 SEC EDGAR company filings

    Would Be Visible By

    2027-11

  • Indicator

    A second major platform voluntarily adopting the MRC Digital Advertising Auction Transparency Standards

    Source

    S-16-10 Media Rating Council

    Would Be Visible By

    2028-06

Affected Industries
  • 16
  • 14
  • 24
What Businesses Should Do

Sell-side publishers should instrument per-bidder floor testing before the remedy lands, because the first six months of a newly opened auction is where yield is won; buy-side should renegotiate platform fee terms while compliance is fresh.

Assumptions
  • Text

    Behavioural remedies change conduct rather than being satisfied on paper

    Confidence

    low

    Load Bearing

    true

    Basis

    The corpus records the French 2021 behavioural settlement as, on practitioner accounts, not eliminating the underlying incentives

    If Wrong

    Upside collapses entirely into base; no mass moves to downside

Precedence Note

Requires disclosed revenue acceleration at a named independent with open-web display, not CTV, as the stated driver. Announcements and compliance filings alone leave the outcome in base.

Would Change Our Mind

Two consecutive quarters of remedy compliance reporting with no revenue acceleration at any independent SSP.

The 2027 comparison bites

IAB's 12.3% US growth figure for 2026 was attributed by the forecaster itself to the sporting calendar, not to AI. 2027 laps both events into the macro brief's sticky-inflation, hawkish-lean environment. Advertising, the most cuttable line in a CFO's budget, contracts; holding companies that cut to 105,000 and 97,000 heads discover they cut capacity rather than overhead; the independent middle consolidates or exits. Platform share still rises, because it rises faster in a downturn.

Mechanism

The transmission channel is the marketing budget approval cycle, not consumer demand: in a restrictive-rate environment with core PCE near 3.3%, CFOs cut the discretionary line whose effectiveness only 15.4% of advertisers say primarily drives budget decisions. Agencies that have already taken 8% of headcount out cannot cut again without losing delivery capacity, so margin compresses rather than holds.

Preconditions, early indicators and assumptions
Band Width Pp

20

Calibration Basis

base_rate

Thesis

2027 laps the Winter Olympics and the FIFA World Cup into a sticky-inflation economy, and advertising contracts.

Preconditions
  • No repeat of 2026's sporting calendar in 2027
  • Federal funds target at or above 3.25% through mid-2027
  • At least one holding company guiding to negative organic growth for 2027
Early Indicators
  • Indicator

    IAB's outlook for 2027 forecasting US advertising growth below 5%

    Source

    S-16-09 IAB Insights

    Would Be Visible By

    2027-03

  • Indicator

    Any of Omnicom, WPP or Publicis guiding to negative organic growth for full-year 2027

    Source

    S-16-01, S-16-02, S-16-03 holding-company IR

    Would Be Visible By

    2027-03

Affected Industries
  • 16
  • 17
  • 14
  • 24
  • 12
  • 11
What Businesses Should Do

Model a 2027 plan on flat-to-down gross budget with the platform share of it rising, and renegotiate agency scope on deliverable volume rather than on headcount, because headcount has already been taken out.

Assumptions
  • Text

    The 2026 growth upgrade is cyclical rather than structural

    Confidence

    medium

    Load Bearing

    true

    Basis

    The forecaster (IAB, via ANA) attributed its own revision to the Winter Olympics and the FIFA World Cup, per T-16-02 and dossier section 8

    If Wrong

    Mass returns to base; the sector simply grows more slowly rather than contracting

Precedence Note

Downside is the base mechanism at a lower growth parameter. It becomes failure only if measurement legitimacy, not the budget, is what breaks.

Would Change Our Mind

IAB forecasting 2027 US growth above 8% in its January 2027 outlook.

The assistant becomes the shelf

Click-to-chat formats keep the transaction inside the assistant. Publishers lose the referral click entirely and demand capture re-prices around conversational placement. Search advertising does not collapse but stops growing, and a fourth large ad-funded pole forms alongside Alphabet, Meta and Amazon. The corpus's query-shape data is the early trace: one-to-two-word queries fell from 42% to 24% of Google Ads impression share since January 2025 while three-to-four-word queries rose from 33% to 48%.

Mechanism

The substitute is conversational commercial discovery, and the two independent parties pursuing it are OpenAI, which reached a disclosed $1bn annualised advertising run rate within 200 days, and Amazon, which began selling ChatGPT inventory through Amazon DSP on 2026-09-10. One company is a project; two is a sector, and here the second is the largest retail-media seller carrying the first's inventory.

Preconditions, early indicators and assumptions
Band Width Pp

25

Calibration Basis

judgement_only

Thesis

AI assistants become the primary commercial discovery surface for a material share of categories and a fourth platform pole forms.

Preconditions
  • OpenAI or a peer continues to sell advertising rather than withdrawing it
  • At least two large demand-side intermediaries carrying assistant inventory
  • Advertisers reporting assistant-attributed conversions in their own reporting
Early Indicators
  • Indicator

    A disclosed OpenAI advertising run rate above $5bn, attributed to a named executive

    Source

    S-16-12 Digiday

    Would Be Visible By

    2028-06

  • Indicator

    A top-20 US advertiser disclosing a double-digit percentage of performance budget allocated to assistant surfaces

    Source

    S-16-16 ANA; S-16-17 WARC

    Would Be Visible By

    2028-12

Affected Industries
  • 16
  • 01
  • 17
  • 12
  • 24
What Businesses Should Do

Instrument assistant-referred traffic and conversions now as a separate channel in analytics, and treat assistant CPMs as a test budget rather than a plan line until a second seller exists at scale.

Assumptions
  • Text

    Assistants substitute for, rather than complement, commercial search

    Confidence

    low

    Load Bearing

    true

    Basis

    The corpus lists this explicitly under things nobody can know: the behaviour is two years old, the assistants change monthly, and the measurement instruments are themselves unstable

    If Wrong

    This branch collapses into base and assistant advertising becomes a modest incremental channel

Precedence Note

Requires disclosed advertiser budget allocation, not assistant usage surveys. The 61% consumer-usage figure the corpus records measures adoption, not substitution, and does not move an outcome into this branch.

Would Change Our Mind

OpenAI's disclosed run rate failing to roughly double between mid-2026 and mid-2027.

Auction transparency becomes enforceable

The FTC and 22 state attorneys general sued Amazon on 2026-08-26 over a soft reserve price feature described internally as an invented auction participant, alleging that by 2024 about 80% of winning bidders paid nearly their full bid despite a publicly described second-price auction. If that theory holds, the MRC's 2026-01-29 auction transparency standard becomes the compliance benchmark, advertiser audit rights become contractually routine, and the opacity margin that has funded parts of the value chain since 2010 compresses.

Mechanism

The instrument is FTC and State enforcement in the 2026 Amazon advertising action, plus the MRC standard issued seven months earlier as the ready-made benchmark a court or a consent order can adopt. The direction is restricting for sellers and liberalising for buyers: once one platform's auction is disclosed, the others' non-disclosure becomes a procurement question rather than a technical one.

Preconditions, early indicators and assumptions
Band Width Pp

20

Calibration Basis

base_rate

Thesis

The FTC prevails or settles favourably against Amazon on the auction-mechanics theory and mandatory auction disclosure spreads.

Preconditions
  • The auction-mechanics theory survives a motion to dismiss
  • The MRC standard is not superseded by a competing industry standard
  • At least one platform adopts disclosure voluntarily to pre-empt
Early Indicators
  • Indicator

    A settlement, consent order or adverse ruling in the FTC Amazon advertising action containing auction-disclosure obligations

    Source

    S-16-23 FTC; S-16-22 CourtListener

    Would Be Visible By

    2029-06

  • Indicator

    A second platform voluntarily adopting the MRC Digital Advertising Auction Transparency Standards

    Source

    S-16-10 Media Rating Council

    Would Be Visible By

    2028-06

Affected Industries
  • 16
  • 12
  • 17
  • 14
What Businesses Should Do

Write MRC auction-standard conformance into media contracts at the next renewal rather than waiting for an order, because the standard already exists and the cost of asking is currently zero.

Assumptions
  • Text

    A US federal court treats auction mechanics as a deceptive-practice question rather than a business-judgement question

    Confidence

    low

    Load Bearing

    true

    Basis

    Novel theory; the corpus records no precedent and notes the FTC's $20bn estimate against Amazon's claimed $8bn of advertiser savings is contested

    If Wrong

    Mass returns to base; the MRC standard remains voluntary and largely unadopted

Precedence Note

Regulatory outranks base when an order with disclosure obligations issues, even if market shares look base-like. A motion-to-dismiss denial alone is not enough.

Would Change Our Mind

Dismissal of the auction-mechanics counts with prejudice.

Measurement legitimacy collapses

Cloudflare already reports automated agents and bots exceeding half of all web requests, and the corpus records that AI agents defeat conventional bot filters by adding cart items and signing up for newsletters. If a credible verifier publishes an agentic invalidity rate above 10% for open-web display, advertisers shift budget to environments with verifiable humans - retail media with logged-in users, live sports, out-of-home - and the open programmatic web is defunded. Publishers dependent on programmatic yield fail; the three platforms are structurally advantaged because their audiences are logged in.

Mechanism

The thesis that fails is that a digital impression is served to a person. The corpus's own risk register calls a single credible study quantifying agentic contamination of billable impressions a sector-level shock. Who is left holding the asset: independent open-web publishers and the SSPs that monetise them; recovery period for an advertising-funded publisher that loses programmatic yield is measured in years, if at all.

Preconditions, early indicators and assumptions
Band Width Pp

25

Calibration Basis

judgement_only

Thesis

Agentic traffic contamination is quantified, is large, and the open programmatic web loses its funding model for the second time in a decade.

Preconditions
  • A verifier or measurement body publishes an agentic invalidity rate for open-web display
  • Advertisers act on it rather than disputing it
  • No rapid standardisation of agent identification that restores confidence
Early Indicators
  • Indicator

    An MRC-accredited verifier publishing an agentic-traffic invalidity rate above 10% for open-web display

    Source

    S-16-10 Media Rating Council

    Would Be Visible By

    2028-12

  • Indicator

    A top-10 advertiser publicly withholding payment on agent-served impressions

    Source

    S-16-16 ANA; S-16-11 AdExchanger

    Would Be Visible By

    2028-12

Affected Industries
  • 16
  • 24
  • 14
  • 17
  • 01
What Businesses Should Do

Move measurable spend toward logged-in and verifiable-human environments before the number is published rather than after, and require agent-classification reporting from verification vendors now.

Assumptions
  • Text

    Existing invalid-traffic frameworks cannot detect agents that transact

    Confidence

    low

    Load Bearing

    true

    Basis

    T-16-10, single-source on Cloudflare and one agency's client data; the corpus scores this trend evidence-limited

    If Wrong

    The branch is near-remote and its mass returns to base and regulatory

Precedence Note

Failure outranks all other branches on occurrence. Distinguished from downside because the budget does not shrink - it relocates.

Would Change Our Mind

An MRC or verifier study finding agentic invalidity below 3% for open-web display.

Additional scenario notes

Precedence Rule

Assign an outcome to the branch whose distinguishing indicator fires first. Regulatory outranks base when a final judgment or an adverse FTC ruling with auction-disclosure obligations issues, even if share numbers look base-like. Failure outranks all branches on occurrence of a published agentic-invalidity rate above 10%.

Probabilities Sum

1

What Must Be True To Grow
  • Advertisers continue to buy ease of execution over measured incrementality - only 15.4% say effectiveness outputs primarily drive budget decisions
  • Platform-measured performance remains acceptable to buyers who cannot independently verify it
  • Total market growth continues at a rate that lets the three platforms absorb it without visibly taking absolute dollars from named competitors
What Could Stop It
  • Regulatory reversal - an appellate reinstatement of structural relief in the Google ad-tech case
  • Demand deflation - the 2027 lapping of the Winter Olympics and FIFA World Cup into a restrictive-rate economy
  • Measurement revision - a quantified agentic-traffic invalidity rate that invalidates the audience model underneath all digital advertising
  • Substitution - conversational assistants becoming the commercial discovery surface and forming a fourth pole
Uncertain Assumptions
  • That Madison and Wall's modelled share series is a fair measure of concentration - the corpus records it as a model, not a measurement, and notes the same modeller printing 56% for US 2025 and about 60% for North America 2026
  • That the sealed Google remedies memorandum opinion, public on or about 2026-09-16, contains no term that materially reallocates inventory - unresolved by construction at the authoring date
  • That Alphabet's advertising segment revenue is growing in line with the modelled share - the corpus could not retrieve it, which is its single largest gap
Review Required

true

Authored

2026-09-15

Subject Trend Evidence Quality

4

Subject Trend Composite Score

82

Subject Trend Continuity Prior

0.95