How AI-Powered Advertising Totals $142 Billion By 2030
Today we are releasing a new long-form publication commissioned by Adobe, “How AI-Powered Advertising Totals $142 Billion By 2030.” A .pdf version of this document will be available at www.madisonandwall.com
—
Everyone agrees that AI is changing advertising. What almost no one can say is how much of the market it actually touches. Is AI reshaping a small layer of the business, or is it already moving meaningful dollars?
This analysis answers that question. We size where AI is already embedded across the advertising value chain, from creative production to content optimization to fully automated media buying. Because creative and media should be tightly linked, and AI tools increasingly make it possible to do so, the only way to understand AI’s impact is to measure it end-to-end. We show which parts of the market are truly AI-driven, which are only marginally affected, and which remain largely manual.
Key takeaways include:
1) $18 billion of U.S. creative spending is exposed to AI.
2) Content will be shaped for AI discovery through GEO tools.
3) AI-powered ad buying is scaling quickly, from 8% of U.S. ad revenue ($35 billion) in 2025 to 26% ($142 billion) by 2030.
4) True “agentic” advertising does not yet exist.
Creative Budgets & AI
AI is beginning to reshape creative work, which supports campaigns across every channel. Creative work in advertising spans a broad set of activities. Some are increasingly AI-addressable, while others are structurally less likely to be automated.
The adoption curve for AI-assisted creative is still early, and there is no hard data that measures the portion already produced with AI. But we can estimate the addressable universe of creative spending that can be optimized and streamlined by AI tools.
Based on IRS and Census data (see Appendix: Sizing of Creative Spending), we estimate there are $18 billion in U.S. creative production and service revenues that are realistically exposed to AI-driven substitution. Certain tasks will be automated almost immediately, others will take longer to unwind, but over time, nearly all these creative processes are eventually at risk.
Adobe Solution Spotlight: Accelerating Creativity & Productivity with Firefly’s Agentic AI
Marketers face increasing pressure to deliver more campaigns at an unprecedented speed and scale. Most organizations are struggling to keep pace with the demand that is only expected to grow as the need to personalize across regions, audiences, and channels increases.
Adobe Firefly family of models has taken a creator-friendly approach to AI — and integrated Firefly-powered features into Adobe products to help accelerate and expand the creative process. Agentic AI has potential to help every creator, at every skill level, work across every medium.
In addition, many creators are experimenting with different models at different stages of their creative process — especially the creative concepting or “ideation” phase. Adobe ensures that creators have control over which creative models they use when, and for which project.
Learn more about Adobe’s approach to creation and production with GenStudio
Content Optimization & AI
Before we reach ad execution, we must also consider how content is shaped for AI platforms. New tools focused on “Generative Engine Optimization” (GEO) have emerged because of changes in how people find information online. As more users turn to tools like ChatGPT or Gemini, information is no longer found by clicking links; it is delivered as a written answer.
We have argued that this shift will not dramatically change ad budgets, but it will change how content must be built to be seen. In the old model, weak optimization meant showing up lower on a results page. In the AI world, the risk is greater: not showing up at all. GEO exists to solve this visibility problem. Its goal is to help brands remain visible in a world where search results are no longer lists of links.
Most GEO tools focus on two main functions:
1) Visibility and competitive intelligence: “Do we show up in AI-generated answers, and if not, who does?” By tracking prompts and responses across LLMs, brands can see where their content appears relative to competitors.
2) Content and signal optimization: The real value is not just tracking, but helping brands change their content. Strong GEO tools will guide how to structure product listings, data feeds, metadata, and long-form content so AI systems can use them more easily.
The goal is to make sure strong content is not ignored because AI systems cannot read or interpret it properly. Over time, the best tools will combine tracking with automatic content changes – rewriting and restructuring material so it works better in AI answers. While this won’t represent a huge volume of direct spending, it will shape almost all content from the outset.
Adobe Solution Spotlight: Own Your Brand’s Presence in AI-Powered Search and Discovery
Brands have spent years investing in traditional search engine optimization (SEO), striving for coveted positions atop search results. But the emergence of AI-enabled search tools has changed the rules. AI agents are filtering, summarizing, and recommending brands based on how well their presence matches a new set of expectations.
In this changing ecosystem, trust takes priority, and brands need to be recognized as authoritative and relevant. Brands who want to lead through this evolution must act decisively by expanding their search goals to include generative engine optimization (GEO).
Adobe LLM Optimizer gives marketing teams a proactive way to engage with this evolution to AI-powered search and discovery — empowering them to enhance visibility, grow influence, and stay ahead in a discovery landscape defined by AI.
AI-Powered Media Execution
Next, we need to look at the execution layer: how much of campaign execution is “AI-powered?”
We believe that at the present time, almost every campaign involves some elements of AI processes. However, it’s clear to us that the entire advertising economy is not 100% “AI-powered” and most solutions assist around the edges. To quantify AI-driven spending, we define AI-powered advertising narrowly: campaign spending that flows through platforms where AI controls targeting, bidding, budget allocation, placement, and ongoing optimization with minimal human intervention. This definition focuses on the point of purchase, where decisions directly drive ad spend.
Nearly every ad seller now claims to be “AI-powered,” but only a subset provides end-to-end automated platforms and workflows (See Appendix: AI-Powered Execution Platform Inclusion Criteria). Most AI advances improve single steps in the process rather than compressing the whole loop.
Where solutions do qualify, they compress the legacy workflow process, i.e., objective setting, audience selection, creative, planning, activation, optimization, measurement, post-campaign reporting, into a single automated loop. The advertiser provides goals, assets, and constraints. The system handles targeting, bidding, placement, pacing, and continuous optimization.
AI-powered Market Dynamics
Overlaying these assumptions onto the Madison & Wall advertising revenue model leads to several conclusions (See Appendix: AI-Powered Adoption Assumptions).
Search and Social Remain the Core of AI-Powered Spending: These channels account for ~88% of AI-powered revenues from 2025–2030, reflecting both faster adoption and underlying channel growth. While AI-powered adoption in commerce and other digital formats reaches ~13% and ~18% of their respective totals by 2030, these categories collectively represent only ~13% of total AI-powered revenues.
Source: Madison & Wall
AI-Powered Solutions Drive Most Industry Growth: Three forces push spend towards AI-powered platforms, revealing advertiser priorities.
Ease over control: As channel fragmentation accelerates, even large advertisers are willing to trade control and transparency for price, performance, and operational simplicity. One-button platforms that optimize to specific KPIs, backed by scaled supply, enable further adoption.
SMB adoption: SMBs already drive 52% of search and 39% of social spend, attracted by simple, consistent, precise tools. These same characteristics support broad SMB adoption of AI-powered products. Enterprise adoption should advance as well, though more gradually.
Working-media orientation: Budgets continue shifting toward digital walled gardens, especially search and social, which maximize working spend when stacked up against linear formats and the open internet.
What About Agentic Advertising?
This analysis focuses on “AI-powered” advertising, which we define narrowly as campaign spending that flows through platforms where AI controls targeting, bidding, placement, pacing, and optimization with minimal human intervention. Platforms like Advantage+ and Performance Max already automate many intra-platform decisions. But this is not agentic advertising.
Current AI-powered tools automate execution within a platform; agentic systems would automate strategy and execution across platforms. Today, no such agentic advertising solutions exist in the market (See Appendix: What Would Agentic Advertising Require?).
We expect AI agents aimed at consumers to scale before AI agents that run marketing operations. Consumer use cases face fewer rules, less brand risk, and fewer internal approvals. Marketing teams operate under tighter controls and more complex processes. As a result, consumer AI agents are likely to become common one to two years before brands are comfortable letting AI systems autonomously manage multi-step, cross-platform marketing. That gap will give us an early signal for when truly agent-driven ad operations are likely to arrive.
Many offerings now labeled “agentic” are better understood as sophisticated AI applications rather than autonomous marketing systems. AdCP-enabled workflows such as those introduced by PubMatic illustrate this distinction, enabling buyers to use natural language to activate, optimize, and manage CTV campaigns with far less manual effort. These tools meaningfully streamline execution, but they remain constrained by platform boundaries. The agent executes a strategy set by a human, without independent control over business objectives, cross-channel budget allocation, or performance-versus-brand decisions. Put simply, they apply AI to make existing workflows more efficient; they do not yet represent marketing that runs itself.
Adobe Solution Spotlight: Redefining Customer Experience Orchestration with AI Agents
We are at the dawn of a new era where business processes and customer interactions are defined by generative AI and AI agents. The way customers discover, engage, and buy is shifting quickly.
Generative AI and intelligent agents are reshaping how customers discover and interact with brands. Increasingly, consumers aren’t searching for websites, they’re asking large language models (LLMs) for information. Brands are no longer just destinations — they’re participants in conversations happening through AI intermediaries.
And as teams embrace agentic AI to augment daily work and drive better results, interoperability amongst AI agents in different ecosystems is critical.
Where the Market Goes from Here
Over time, more creative will be made with AI. More content will be built for AI discovery. More ad spending will move into automated systems where AI controls targeting, budgets, and optimization. This will change how money flows to publishers, how agencies add value, and how brands run campaigns.
The direction is clear: more spending will flow through systems that automate execution from start to finish. The winners will be those who build their content, products, pricing, and workflows for an AI-first world.
APPENDIX
Additional Creative Details
These examples illustrate the creative sub-processes including in our “AI-exposed” creative universe; they are not all equally automatable near-term:
Creative strategy and concept development
Asset creation and production
Creative adaptation and optimization
Localization and cultural adaptation
Experiential and physical creative
Influencer and creator assets
Sizing of Creative Spending
Businesses deducted roughly $548 billion in advertising expenditures in 2024, according to our estimates, which are informed by historical IRS filings. We estimate publishers captured $401 billion, leaving $147 billion in non-working spend, including production, strategy, account management, and creative services.
Using historical U.S. Census data on agency service-line revenues, and reallocating blended “integrated services” across underlying categories, we estimate that creative services account for roughly 12% of this non-working layer.
Source: Madison & Wall, U.S. Census Bureau
Applying that share to the $147 billion implies approximately $18 billion in U.S. creative production and service revenues that are most directly exposed to AI-driven substitution. Put simply, the entire $18 billion represents creative work that AI can touch. Certain tasks will be automated almost immediately, others will take longer to unwind, but over time, nearly all these creative processes are eventually at risk.
Where Does “AI” Show Up in Advertising?
This is a taxonomy of AI features by channel; it is broader than our “AI-Powered” spending definition. Across channels, AI shows up in many forms:
Search: AI Overviews, smarter bid tools, automated campaign setup.
Social: automated targeting, measurement, and creative tools.
Commerce Media: automated bidding, improved incrementality/attribution, product recommendation engines, dynamic product ads, LTV optimization.
Other Digital (e.g., YouTube): AI tools for image/video generation, voice, editing, and automated bidding and pacing.
Television: creative versioning at scale, improved targeting and planning, improved modeling.
Other legacy formats (audio, outdoor, publishing): audience modeling, cross-channel forecasting, creative tools that speed production and lower cost.
Our “AI-Powered” definition includes instead the specific platforms below.
AI-Powered Execution Platform Inclusion Criteria
The following list is illustrative, and reflects products that materially automate the targeting / bidding / budget / placement within a platform:
Google Performance Max
Meta Advantage+
TikTok Smart+ and GMV Max
Pinterest Performance+
LinkedIn Accelerate Campaigns
Reddit Dynamic Product Ads
Snapchat Smart Campaign Solution
X Automated Campaigns
Microsoft Performance Max
Amazon Automated Targeting for Sponsored Products
Ad-tech solutions that automate campaigns across the open web in a similar way
Campaigns executed through these products are counted as “AI-powered” in our model because they automate the major steps of execution and optimization with minimal human involvement.
We exclude platforms that use AI in isolated parts, but still require advertisers to set up targeting parameters, placements, and structure their campaigns manually. Examples of these include:
Roku OneView
Disney DRAX
NBCU One Platform
Paramount EyeQ
WBD Neo.
We assume some end-to-end execution capability emerges in TV, print, audio, and OOH in the coming years. But meaningful penetration before 2030 is unlikely. The majority of AI-powered spend will remain concentrated in digital walled gardens throughout the forecast period.
AI-Powered Adoption Assumptions
These estimates combine platform disclosures where available with Madison & Wall analysis to reconcile to total revenues. Our directional view on where the market goes from here is shaped by several observations:
1) Meta and Google lead: As of 4Q25, we estimate that Meta Advantage+ campaigns represented ~25% of advertising revenue; Google Performance Max reached ~12% of revenue. These data points, along with past commentary, anchor our adoption curves.
2) Other digital walled gardens trail at a distance: Disclosures are limited, but we estimate adoption of automated products at ~1–10% of relevant platform revenues in 4Q25.
3) Search and social remain the dominant AI-powered channels through 2030: We estimate that search AI-powered adoption reaches ~37% of advertising revenues, and social reaches ~48% adoption by 2030.
4) Non-digital media exhibit limited adoption: Linear TV, outdoor, audio, publishing, and directories/direct mail show no meaningful AI-powered penetration through 2030. Although AI will influence workflow (e.g., targeting, data hygiene), these reflect incremental improvements to pre-AI processes rather than true end-to-end automation. We do assume that digital TV, audio, outdoor, and publishing gain some AI-powered adoption over time but stay below 10% of total channel ad revenues by 2030. We believe structural and sales-process constraints slow adoption irrespective of AI narratives.
Non-AI-Powered Advertising Breakdown
Non-AI-powered search and social decline through 2030 (-1% CAGR 2025-2030). Meanwhile, national TV and other non-digital formats maintain roughly stable share within the non-AI-powered portion of the market (31% share in 2025, falling to 27% by 2030, a similar -3% CAGR 2025-2030), reflecting the fact that most spend in these channels remains manually driven.
Source: Madison & Wall
Additional AI-Powered Market Spending Dynamics
This section expands upon the headline “AI-powered” spending key findings in the main text with additional channel-level rationale.
1. Social Media Leads AI-Powered Adoption
We estimate social reaches ~20% AI-powered in 2025, and ~48% by 2030. Meta drives this shift. This model aligns with Mark Zuckerberg’s stated goal: “Any business can basically tell us what objective they’re trying to achieve, like selling something or getting a new customer, and how much they’re willing to pay for each result, and then we just do the rest.” SMBs adopted these tools early; large marketers now accept the ‘black box’ in cases where it outperforms manual control. Other platforms are following the same path, especially as marketers are already on-board from their experience on Meta.
Source: Madison & Wall
2. Search Follows Closely
Google’s Performance Max shows strong traction. Search is the only channel where small and medium businesses still account for more than half of revenue, helping speed adoption of simplified, automated buying. Large advertisers increasingly recognize that, for certain KPIs, automated systems can outperform hands-on campaign management. Given the ease of use, AI-powered platforms don’t have to outperform manual campaigns, they just have to be roughly equal in order to cannibalize spend.
Source: Madison & Wall
3. Outside Search and Social, Adoption is Slower
YouTube and commerce media show meaningful progress, but beyond those channels, structural headwinds slow automation. U.S. TV buying still relies on upfront deals and relationship-based sales. Broadcasters are reluctant to give up direct control, especially for premium inventory and sports. Many of the factors that slowed addressable video buying and dynamic creative insertion technologies over the last two decades will apply to AI adoption as well. We estimate that in 2030, almost all of the 26% of AI-powered revenue will still come from digital channels.
Source: Madison & Wall
4. AI Isn’t Creating Growth, It’s Capturing It
It’s important to keep in mind that new processes don’t conjure money out of thin air. Despite all the hype when it comes to AI and advertising, marketers still set advertising budgets based on their underlying business (which typically grows at the rate of personal consumption growth), and then those budgets are allocated into various formats and strategies based on what is believed to be the most effective option for spending.
AI-powered tools are appealing to marketers. As a result, they are capturing a significant share of budgets and a large share of incremental dollars. And on the margin it’s possible that other marketing spending might migrate into paid advertising because of the effectiveness of AI-powered tools. But if the AI-powered tools didn’t exist, then a significant portion of those budgets would still be spent elsewhere, likely in legacy search, social, and commerce platforms. So while the “growth rate” of AI-powered spending is significant, it largely reflects a reallocation of dollars that would have gone elsewhere, rather than an expansion of the overall advertising pie.
Keeping this context in mind, we estimate that AI-powered budgets will grow at ~29% CAGR through 2030. While this growth comes from a small base and largely cannibalizes non-AI digital buying methods, the pattern is familiar. The largest platforms will continue to gain share. Walled gardens will compress the open internet. And buyers will continue to choose simplicity when it is backed by performance.
Source: Madison & Wall
What Would Agentic Advertising Require?
Agentic systems require a higher bar than “AI-powered”: multi-step, cross-platform autonomy. In our view, agentic advertising would involve an AI system that operates well beyond the boundaries of a single execution engine. Instead of optimizing a campaign after the marketer sets goals, an agentic system would autonomously:
Define goals and strategies across channels
Select publishers and inventory sources across platforms
Build creative variants and landing pages
Launch and pause campaigns across multiple walled gardens
Reallocate budgets among platforms like PMax, Advantage+, and TikTok Smart+
Optimize simultaneously for upper- and lower-funnel objectives
Escalate to humans only when business-level input is required
This mirrors how consumer-facing agentic tools may eventually handle shopping journeys: finding products, comparing retailers, identifying substitutes, executing the purchase, and managing delivery all without human micromanagement. Agentic marketing would apply the same logic to campaign execution.
This is a significantly higher bar than our definition of AI-powered.