← All perspectives

Company · Marketing & AI · Equity · 2026

The Hour Trap: Why AI Breaks the Agency Holding Company — and How We’re Buying What Comes Next

The advertising holding company is breaking in public. AI is deflating the billable hour it was built on. That combination has put proven, cash-flowing agencies on sale at the exact moment a new operating model can transform their economics.

The world spent $1.14 trillion on advertising in 2025 — the largest marketing market in history. The companies built to serve it are being dismantled in public.

Start with the scoreboard. WPP — the largest agency group on earth for fifteen years — lost roughly 60% of its share price in 2025 alone, cut guidance three times, and watched its market value fall from £24 billion in 2017 to under $4 billion today. Its new CEO’s own words: WPP is “no longer a holdco.” Omnicom and IPG merged in a $13.3 billion all-stock deal that closed in November 2025 — and promptly doubled the cost-cut target to $1.5 billion, with roughly 4,000 further jobs to go. Dentsu posted a quarterly loss, announced 3,400 international job cuts, and tried to sell its entire international business; no buyer would take it. S4 Capital, the “digital-first” challenger of the last cycle, sits 98% below its 2021 peak.

This is not an advertising recession. Ad spend grew almost 9% in 2025. The clients are spending; the model is what’s breaking.

The Hour Trap

The holding companies have a structural problem, not a cyclical one, and one datapoint proves it: Publicis — the one major group that spent a decade rebuilding itself around data and AI — grew +5.6% organically in 2025, its sixth straight year of outperformance, at an 18.2% operating margin, while WPP shrank 5.4%. Same market. Same clients. Opposite outcomes. The difference is not talent; it is architecture.

The traditional agency sells time. Headcount and hours are the inventory; markup on both is the margin. That model survived every prior technology wave because each wave made marketing more complex, and complexity billed more hours. AI is the first wave that deletes the hours themselves. Inference costs have fallen roughly 1,000x since 2022; work that consumed a creative team’s week now takes a supervised afternoon. Meta has said advertisers will be able to fully automate ad creation on its platforms by end-2026. 92% of US digital display is already bought programmatically.

Clients have noticed. 39% of CMOs say they are cutting agency budgets, and roughly one in five report that generative AI has already reduced their reliance on external agencies. 60% of senior US marketing leaders say AI has them spending less on agencies. Forrester projects US agencies will shed 32,000 jobs to automation by 2030. Meanwhile 82% of major advertisers now run an in-house agency, up from 58% a decade ago.

We call the result the Hour Trap: AI does not compress margins for agencies that price on value. It compresses margins for agencies that price on time. The holdcos — carrying high fixed costs, siloed fiefdoms, and public-market scrutiny — cannot escape the trap, because every hour AI saves is revenue they must give back. An agency that prices outcomes keeps that hour as margin.

The incumbents’ distress is rational. So is the opportunity it creates.

The Sale Rack Nobody Is Shopping

Fear reprices assets. Independent agencies — the profitable, founder-owned firms that hold real client relationships — currently trade at roughly 3–7x EBITDA depending on size and quality, per 2025–26 M&A survey data, with even strong mid-sized firms changing hands in the mid-single digits. Consolidated platforms command 8–12x, and strategic buyers pay well beyond that: KKR valued FGS Global, a single premium asset carved out of WPP, at $1.7 billion; Accenture Song built a $20 billion-revenue marketing business — overtaking WPP itself in 2025 — largely by acquisition.

The discount exists because sellers and buyers are both afraid of the same headline: AI kills agencies. The evidence says something more specific — AI kills time-priced agencies. The buyer who can re-price acquired revenue around outcomes, and re-platform delivery around AI, is buying dollar bills marked down for a risk they’ve engineered out.

Supply favors that buyer too. A generation of agency founders who built firms in the internet’s first wave is reaching succession age — part of the broadest owner-transition wave in US small business history, peaking in 2027–28 — with no holdco bid waiting, because the holdcos are busy shrinking.

What Our Investment Unlocks

One Marketing is 1V1sion’s answer: a platform acquiring proven, outcome-focused independent agencies and integrating them on a shared AI-native operating layer — built with the conviction that the winning model is human judgment orchestrating machine execution, not the reverse.

The founding platform combines agencies spanning media, creative, experiential, and full-service — roughly 180 professionals across the US, Canada, and the UK, serving 600+ client relationships across education, telecom, automotive, entertainment, luxury, and consumer brands. These are not turnarounds. They are award-winning operators with retainer-dominant, fixed-price revenue — the flagship media agency lifted its net revenue retention from 82% to 107% over the past three years while cutting revenue churn from 24% to 7%.

The wedge is deliberate: direct-to-consumer and mid-market brands — a segment of roughly 110,000–120,000 US DTC companies spending 10–20% of revenue on marketing, within a DTC economy of some $600 billion in spend. These clients need integrated, outcome-driven marketing most and get it least: the holdcos gravitate to global accounts, and boutique shops can’t offer the full stack. An AI-leveraged platform can serve this segment profitably at a price point neither incumbent model can match.

The integration thesis has three engines:

1. Re-pricing captures what AI saves. Because the platform’s contracts are retainer- and outcome-based rather than hourly, every efficiency the AI layer produces lands as margin, not as a client rebate. Industry benchmarks put AI-native delivery margins at 40–60% against the holdcos’ 13–17% — the gap is the prize, and the window matters: early movers pocket the spread for years before pricing resets.

2. Cross-sell is mapped, not hoped. The platform’s operating leadership — drawn from the CMO seats of major streaming platforms, the executive team that sold a $5 billion media business, and the growth leadership of top digital agencies — carries a mapped network of hundreds of senior brand relationships representing multibillion-dollar marketing budgets. Land-and-expand across four complementary service lines converts one relationship into four revenue streams.

3. A data flywheel the startups can’t build. Every campaign across the owned agencies feeds a shared intelligence layer — live spend, creative performance, audience data — that no software-only startup possesses and no conflicted holdco can neutrally offer. The platform deploys its technology on its own clients first, proves the ROI with real budgets, then licenses outward. Owning the execution layer is the moat.

And the forward bet, stated plainly: as buying decisions migrate from humans to AI agents, the scarce asset in marketing stops being attention and becomes trust — being the brand that both people and machines confidently recommend. The platform is being architected for that world — machine-readable, verifiable, outcome-priced — not retrofitted to it.

The Upside, Priced by Precedent

The arbitrage is visible in public marks. Buy at mid-single-digit multiples; operate into the 8–12x platforms command; sell into a strategic market that has paid premium multiples for scaled, tech-enabled marketing assets — while Accenture, the consultancies, and the surviving holdcos all need exactly this capability and have shown they buy it rather than build it. S4 Capital proved the market will award billions to a credible new-model agency narrative — reaching a £4 billion market cap within three years of founding — and its subsequent collapse is equally instructive: it overpaid with equity, over-levered, and listed too early. The playbook here is the inverse: disciplined entry prices, conservative leverage, private ownership, and cash flow from day one.

What Protects the Downside

We underwrite One Marketing as a private equity risk profile with venture-scale upside — in that order.

  • Cash flow at entry. The platform acquires profitable agencies with established client bases — not pre-revenue technology bets. The base case works if the AI thesis merely proceeds slowly.
  • Retainer revenue with visibility. Fixed-price, retainer-dominant contracts with multi-year client tenure — not project work that evaporates in a downturn.
  • Diversification by design. Hundreds of client relationships across six-plus sectors; no single client dominates at platform level.
  • Variable cost structure. Agency cost bases flex with revenue; capex is minimal; integration synergies are upside in the model, not a requirement of the base case.
  • Aligned sellers. Founding agency leaders roll equity and sign multi-year employment agreements — sellers keep skin in the outcome.

The honest risks: integration execution across acquired cultures, the pace at which AI-driven pricing pressure arrives, and in-housing — though more than 90% of what brands have in-housed is narrow digital execution, leaving integrated strategy, creative, and orchestration squarely with agencies. What this investment does not depend on: the AI platform achieving venture outcomes, holdco multiples recovering, or the ad market accelerating.

The open question is timing — how fast the industry’s revenue re-prices from hours to outcomes. If it’s slow, we own growing, cash-generative agencies bought at a discount. If it’s fast, the operating layer converts that speed into margin the incumbents structurally cannot match.

Either way, the hour is dying. We’d rather own what replaces it.

The full underwriting file — market model, agency-level diligence, and integration plan — is available to qualified investors. If marketing services, AI-enabled platforms, or founder-succession consolidation is on your radar, we’re glad to walk through the work.