The trouble with AI advertising statistics is that three different things share the name: AI that buys and optimises media, AI that makes the creative, and AI that decides who sees it. Vendors blend them into one slide, and a platform's own lift claim gets repeated as if it were independent research. This page keeps them apart, labels every figure as a platform claim, a survey of opinion or a measured result, and links each one to its publisher. Cite freely.
AI advertising statistics at a glance
| Category | Statistic | Source |
|---|---|---|
| Buying | Meta's Advantage+ reached a $75B annual revenue run rate in Q2 2026, Meta's own disclosure | Meta Q2 2026 earnings call, 29 July 2026 |
| Buying | That is about 32% of Meta's Q2 2026 Family of Apps ad revenue, annualised | Computed from Meta Q2 2026 |
| Buying | Two-thirds of US ad buyers are focused on agentic AI for buying and campaign execution | IAB 2026 Outlook Study, January 2026 |
| Creative | 9 million small businesses use at least one of Meta's AI ad creative tools, Meta's own count | Meta Q2 2026 earnings call |
| Creative | AI-generated ads beat a professional designer's on click-through rate, 0.98% against 0.65% | Daviet and Nishimura, Marketing Science, August 2026 |
| Creative | Over 40% of ads where the AI use is not obvious land in the top tier for branded cut-through | Kantar, November 2025 |
| Trust | 49% of US adults view businesses using AI to make ads negatively; 19% view it positively | Gallup and Bentley University, August 2026 |
| Trust | 62% of US adults say AI-created people or voices in ads are unacceptable | Gallup and Bentley University, August 2026 |
| Adoption | 96% of CMOs claim end-to-end AI transformation; 8% run campaigns with autonomous agents | BCG, June 2026 |
| Adoption | CMOs put 15.3% of marketing budget into AI, and 30% say they are ready to scale it | Gartner 2026 CMO Spend Survey, May 2026 |
| Measurement | 60% to 75% of advanced measurement users say current approaches fall short | IAB State of Data 2026, February 2026 |
| Measurement | Better AI measurement could release $26.3B in media investment and $6.2B in productivity | IAB, February 2026 |
| Waste | $26.8B of global media value is lost every year to programmatic inefficiency | ANA, August 2025 |
| Market | US AI advertising spend will more than double to $68.25B by 2030 | eMarketer, June 2026 |
How much advertising money now runs through AI buying systems?
Meta's Advantage+ products alone carry a $75 billion annual revenue run rate as of Q2 2026. That is roughly 32% of Meta's annualised advertising revenue, and close to 6% of the $1.30 trillion WARC expects the world to spend on advertising this year.
| Statistic | Source |
|---|---|
| Meta's Advantage+ reached an annual revenue run rate of $75 billion in Q2 2026, on Meta's own disclosure to investors. | Meta Q2 2026 earnings call, 29 July 2026 |
| Meta Family of Apps ad revenue was $59.4 billion in Q2 2026, up 27% year over year and 26% in constant currency. | Meta Q2 2026 earnings call |
| Alphabet's Search and other revenue grew 17% and YouTube ads grew 13% in Q2 2026, on Alphabet's own reporting. | Sundar Pichai, Alphabet Q2 2026 earnings remarks, 22 July 2026 |
| Global ad spend is forecast to grow 9.1% in 2026 to $1.30 trillion, about $150 for every person on earth. | WARC, The Future of Media 2026, January 2026 |
| Almost 80% of ad spend now flows into retail media, paid search and social platforms. The rest of the media industry shares the other 20%. | WARC, January 2026 |
| US ad spend is forecast to grow 9.5% in 2026, in a survey of more than 200 brand and agency buyers. | IAB 2026 Outlook Study, January 2026 |
| Two-thirds of buyers say they are focused on agentic AI for ad buying and campaign execution in 2026. | IAB 2026 Outlook Study |
| US search advertising revenue was $114.2 billion in 2025 and total US digital ad revenue was $294.6 billion, up 13.9%. | IAB and PwC, Internet Advertising Revenue Report, April 2026 |
| US AI advertising spend will more than double to $68.25 billion by 2030, and more than 80% of AI advertising in 2026 sits next to AI content rather than inside a chatbot. | eMarketer, June 2026 |
The $75 billion is the figure I would put in front of a board, because Meta has already banked it. Automated buying arrived, and the share of the biggest ad platform running on it is closer to a third than to a pilot. WARC's 80% concentration number explains why: when four fifths of the money sits inside retail media, search and social, the auction owners decide how much of the buying decision you still get to make.
What performance lift do AI ad systems deliver?
The median AI advertising lift published in 2026 is 8.3%, the midpoint of five figures Meta and one peer-reviewed field experiment put on the record this year. Three of those five are Meta's own self-reported numbers, so read 8.3% as a median of mostly self-reported lifts rather than of independent measurements.
| Statistic | Source |
|---|---|
| Meta's GEM ads ranking model and sequence learning produced an 8.3% increase in ad clicks and a 15.7% uplift in conversions on Facebook. Meta's own figure, with no independent control disclosed. | Meta Q2 2026 earnings call |
| Early Meta pilots using large language models to read user preferences drove a 1% increase in app event conversions on Instagram. Again, Meta's own figure. | Meta Q2 2026 earnings call |
| In a live Instagram campaign for one outdoor activities advertiser, AI-generated ads recorded a 0.98% mean click-through rate against 0.65% for a professional human designer, a 50.8% relative gain. | Daviet and Nishimura, Marketing Science, August 2026 |
| Rerun 18 months later without retraining, the same AI portfolio averaged 3.38% click-through rate against 3.24% for the human-designed campaign, a 4.3% gain. | Daviet and Nishimura, Marketing Science, August 2026 |
| Only 31% of B2C CMOs and 20% of B2B CMOs report significant, measurable revenue impact from AI, in a survey of 300 CMOs. | BCG, June 2026 |
| Google announced AI Max for Shopping and Travel, the Asset Studio creative tools and the Ask Advisor agent at Google Marketing Live on 20 May 2026. The page carries product descriptions and no performance figures. | Google Marketing Live 2026, May 2026 |
Meta's numbers come from a company reporting to shareholders, so they are audited in spirit but not designed as experiments, and they measure Meta's own ranking upgrades rather than a marketer's decision to switch on automation. The Marketing Science field experiment is the opposite: one advertiser, one category, published with standard deviations.
Every agency deck I have seen this year quotes a platform lift as if it were independent evidence. It is not. Google and Meta publish the numbers that sell the product, and neither discloses the counterfactual. A 50.8% first run that decayed to 4.3% in 18 months says the advantage is mostly speed, and speed erodes. Build on the 8.3% median anyway.
How many marketers use AI in advertising, and how deep does it go?
Almost every marketing organisation says it has adopted AI and almost none has changed how a campaign runs. BCG found 96% of CMOs reporting significant end-to-end AI transformation in 2026 while 8% run campaigns where multiple agents operate autonomously.
| Statistic | Source |
|---|---|
| 96% of CMOs report significant end-to-end AI transformation, 42% use generative AI only for discrete tasks, just under a third have reached agent-led workflows and 8% run multi-agent autonomous campaigns. Sample: 300 CMOs plus 50 interviews. | BCG, June 2026 |
| CMOs allocate 15.3% of marketing budgets to AI in 2026, but only 30% report mature AI readiness while 70% call AI leadership a critical goal. Sample: 401 CMOs across North America, the UK and Europe, fielded January to March 2026. | Gartner 2026 CMO Spend Survey, May 2026 |
| 4 in 10 US companies now use generative engine optimisation, and no marketing technology capability scores above 5 on a 7-point performance scale. Sample: 308 marketing leaders, 97% VP level or above, fielded 7 to 29 January 2026. | The CMO Survey, March 2026 |
| More than 70% of marketers embrace generative AI for advertising creativity and 40% call it a top-three priority. Sample: 974 senior marketers, fielded 4 April to 15 June 2025. | Kantar Media Reactions 2025, September 2025 |
| 71% of brands with marketing budgets above $1 billion named AI for personalisation and optimisation as a top trend for the year. Sample: 1,400 marketers, fielded 25 February to 6 March 2025. | Nielsen Annual Marketing Report, May 2025 |
| 9 million small businesses use at least one of Meta's AI ad creative tools, on Meta's own count. | Meta Q2 2026 earnings call |
Adoption surveys ask whether AI is in use somewhere, which by 2026 is true of any team with a laptop. I read the gap between claimed transformation and real autonomy as a reporting problem: boards ask CMOs whether they are AI-led, and nobody answers no. Hold a team to BCG's 31% of B2C CMOs who can point to measurable revenue impact, because that one needs evidence. For the wider budget picture, see the brand vs performance marketing statistics page.
Does AI-generated creative beat human-made creative?
On click-through rate, yes, and by a wide margin in the one peer-reviewed field test published so far. AI-generated ads hit a 0.98% mean click-through rate against 0.65% for a professional human designer in a live Instagram campaign, published in Marketing Science in August 2026. On brand effects, the picture reverses when the AI is obvious.
| Statistic | Source |
|---|---|
| AI-generated ads recorded a 0.98% mean click-through rate (standard deviation 0.26%), against 0.65% for a professional human designer (0.31%) and 0.78% for an aesthetics-optimised AI benchmark (0.41%). The AI creatives beat the designer's batch in 99.59% of samples. | Daviet and Nishimura, Marketing Science, August 2026 |
| In the 18-month rerun with no model retraining, the AI portfolio averaged 3.38% click-through rate and the contemporary human-designed campaign 3.24%. No conversion data was reported. | Daviet and Nishimura, Marketing Science, August 2026 |
| More than 40% of ads where the AI use is not obvious landed in the top tier for branded cut-through in Kantar's LINK database. Ads with obvious AI use performed worse. | Kantar, Rethinking AI-generated advertising, November 2025 |
| 41% of consumers say AI-generated ads bother them, against 29% of marketers. This is 2024 fieldwork and the oldest number on this page. | Kantar Media Reactions 2024, via Kantar, November 2025 |
| 9 million small businesses use at least one of Meta's AI ad creative tools, and adoption compounds when advertisers use more than one. Meta's own count. | Meta Q2 2026 earnings call |
AI creative wins the cheap metric and has not proven the expensive one. The Marketing Science number is a click result from one advertiser in one category, and clicks are what AI optimises most easily. Kantar measures what the ad does to the brand, and its finding splits on visibility rather than authorship. Use AI for volume and iteration speed, keep a human on the idea, and never let the output look generated. On moving image, the AI-generated video statistics page follows the same logic.
What do consumers think about AI-made ads?
US consumers are hostile to AI in advertising and get more so with each step toward the finished ad. 49% view businesses using AI to create advertisements negatively against 19% positively, in a Gallup and Bentley University survey of 3,270 US adults fielded 4 to 11 May 2026 and published in August. These are attitudes, not performance data, and the distinction matters.
| Statistic | Source |
|---|---|
| 79% of US adults have seen advertisements in the past 30 days that looked AI-created. | Gallup and Bentley University, August 2026 |
| 49% view businesses using AI to create ads negatively, 19% positively, 32% neutrally. Sample: 3,270 US adults, fielded 4 to 11 May 2026, margin of error 2.4 percentage points. | Gallup and Bentley University, August 2026 |
| 75% say using AI for brainstorming and early drafts is acceptable when disclosed, 53% accept it for final text, images or video, and 62% say AI-created people or voices in ads are unacceptable. | Gallup and Bentley University, August 2026 |
| 66% of US adults aged 18 to 29 view AI-made advertising negatively, against 51% of 30 to 44 year olds, 38% of 45 to 59 year olds and 46% of those 60 and over. 73% of 18 to 29 year olds reject AI-created people and voices. | Gallup and Bentley University, August 2026 |
| More than half of consumers feel uneasy and distrustful of AI-generated ads and more than 60% worry generative AI could produce fake or misleading advertisements. Sample: 21,300 consumers aged 16 to 65 across 30 markets, fielded 8 May to 5 August 2025. | Kantar Media Reactions 2025, September 2025 |
| 40% of US adults say AI will have a negative impact on society over the next 20 years and 31% expect a negative effect on themselves. Roughly seven in ten expect AI to make their personal information less secure. Sample: 5,119 US adults, fielded 17 to 23 February 2026. | Pew Research Center, June 2026 |
Put the marketer and consumer numbers side by side and the gap is at least 51 points: more than 70% of marketers embrace generative AI for advertising creativity, 19% of US adults view AI-made ads positively. The age curve should worry brand teams most, because hostility peaks at 66% among 18 to 29 year olds and falls to 38% among 45 to 59 year olds. Gallup's 75% who accept AI for brainstorming is the way through. Consumers are not against the tool, they are against the finished ad being a machine's work.
Is AI fixing advertising measurement or complicating it?
Advertising measurement is the area where the buy side has both the most confidence in AI and the least working today. Between 60% and 75% of advanced measurement users say current approaches fall short on rigour, timeliness, trust and efficiency, in the IAB's State of Data 2026 survey of more than 400 senior decision-makers.
| Statistic | Source |
|---|---|
| Between 60% and 75% of users of advanced measurement say current solutions fall short on rigour, timeliness, trust and efficiency. Sample: 400+ senior planning and analytics decision-makers, run with BWG Global. | IAB State of Data 2026, February 2026 |
| The buy side believes AI could release $26.3 billion in additional media investment and $6.2 billion in productivity value within one to two years. | IAB, Project Eidos announcement, February 2026 |
| About 50% of organisations are already scaling AI inside advanced measurement, and more than 70% of those that are not expect to within one to two years. | IAB, February 2026 |
| No respondents believe all paid channels are well represented in their marketing mix models. Around 40% have AI-related clauses in brand-agency contracts, a share expected to double. | IAB, February 2026 |
| Only 32% of global marketers measure spending holistically across digital and traditional channels, falling to 23% in Europe. Sample: 1,400 marketers, fielded February to March 2025. | Nielsen Annual Marketing Report, May 2025 |
The zero is the number to sit with. Not one respondent in a 400-person sample thinks their marketing mix model covers every paid channel, and that is the model most large advertisers use to defend their budgets. The $32.5 billion the IAB's respondents say better measurement would release is the same order of magnitude as the $26.8 billion the ANA says programmatic already wastes. I would spend the first dollar on inputs rather than models. For how those models are built, see the marketing mix modeling statistics page.
Where does AI buying still waste money?
Automated buying has cleaned up the most visible waste and left a wide quality gap. Made-for-advertising exposure fell to a median 0.8% of programmatic spend in Q2 2025 from 15% two years earlier, while $26.8 billion of global media value is still lost each year to inefficiency, according to the ANA's Programmatic Transparency Benchmark.
| Statistic | Source |
|---|---|
| $26.8 billion of global media value is lost every year to programmatic inefficiency. Made-for-advertising exposure fell to a median 0.8% of spend, from 15% two years earlier, though top-quartile marketers still spent up to 28.7% on MFA domains. | ANA Q2 2025 Programmatic Transparency Benchmark, August 2025 |
| Advertisers with disciplined quality governance converted 56.7% of programmatic spend into benchmark-qualified impressions, against 37.5% for lower performers. Optimising to quality-adjusted metrics rather than CPM cut cost per conversion by nearly 40% in one case. | ANA Q4 2025 Programmatic Transparency Benchmark, February 2026 |
| More than 80% of AI advertising in 2026 appears next to AI content, such as search ads beside AI Overviews, rather than inside chatbots. | eMarketer, June 2026 |
| 4 in 10 US companies use generative engine optimisation, and no marketing technology capability scores above 5 on a 7-point scale. | The CMO Survey, March 2026 |
The 19-point spread between disciplined and undisciplined advertisers in the ANA's Q4 2025 benchmark is the most useful operating number in this set. Programmatic waste and AI optimisation are the same story told twice: the algorithm got better at hitting the target and the target was often the wrong one. The 56.7% against 37.5% split says the difference between a good and a bad 2026 is governance, not technology. On channel efficiency, the organic vs paid search statistics page has the click and cost benchmarks.
How we calculated the original numbers
Four figures on this page do not appear anywhere else. Here is the arithmetic, so you can check it or change the inputs.
- 32% of Meta's ad revenue runs through Advantage+. Meta reported $59.4 billion of Family of Apps ad revenue in Q2 2026, which annualises to $237.6 billion. Susan Li said Advantage+ had reached a $75 billion annual revenue run rate in the same quarter. 75 / 237.6 = 31.6%. Caveat: a run rate annualises current revenue, so the two sides of the ratio are built differently.
- A trust gap of at least 51 points. Kantar's Media Reactions 2025 found more than 70% of marketers embrace generative AI for advertising creativity. Gallup and Bentley University found 19% of US adults view businesses using AI to create ads positively. 70 minus 19 = 51 points, and Kantar's "more than" makes that a floor. Different questions of different populations, so read it as the size of the perception gap.
- 8.3% is the median published lift, mostly self-reported. Five lifts were published in 2026: Meta's 8.3% increase in ad clicks, Meta's 15.7% uplift in conversions, Meta's 1% increase in Instagram app event conversions, the Marketing Science 50.8% click-through gain over a human designer (0.98% vs 0.65%) and the same study's 4.3% gain in the rerun (3.38% vs 3.24%). Sorted: 1.0, 4.3, 8.3, 15.7, 50.8. Three of the five are Meta's own claims with no disclosed control.
- $32.5 billion, or 11% of US digital ad revenue. The IAB puts the prize from better AI measurement at $26.3 billion of media investment plus $6.2 billion of productivity value, which is $32.5 billion. US digital ad revenue was $294.6 billion in 2025 (IAB and PwC). 32.5 / 294.6 = 11.0%.
What the 2026 numbers say
Automated buying is settled and nobody voted on it. A $75 billion run rate at Meta, two-thirds of US buyers working on agentic AI, 80% of world ad spend inside retail media, search and social. The platforms took the decision about how much of the buy a machine makes, and the marketer's job is now the objective and the inventory rules.
The published lift is single digits, and the loudest numbers are the least independent. Five published lifts in 2026, median 8.3%, three of them Meta's own, and the largest decayed from 50.8% to 4.3% in 18 months. Build the business case on the median, not the maximum.
Creative is where the risk sits, not the buying. Consumers have no view on your bidding algorithm and a strong view on your ad: 49% negative against 19% positive, 62% rejecting AI-made people and voices, 66% hostility among 18 to 29 year olds. Kantar says the penalty attaches to visible AI rather than to AI itself, which makes it a production standard you control completely.
FAQ
How much of Meta's ad revenue runs through AI automation?
About 32%. Meta's Advantage+ reached a $75 billion annual revenue run rate in Q2 2026, against $59.4 billion of Family of Apps ad revenue in that quarter, which annualises to $237.6 billion. Both figures come from Meta's 29 July 2026 earnings call. A run rate annualises current revenue rather than reporting a trailing year, so treat 32% as a close approximation.
Do AI-generated ads perform better than human-made ads?
On clicks, yes. AI-generated ads recorded a 0.98% mean click-through rate against 0.65% for a professional human designer in a live Instagram campaign, published in Marketing Science in August 2026 by Remi Daviet and Yohei Nishimura. On brand effects the answer is conditional: Kantar found over 40% of ads where the AI use is not obvious reached the top tier for branded cut-through, while ads with obvious AI use scored worse. No large study has compared conversions.
What do consumers think about AI-generated advertising?
49% of US adults view businesses using AI to create advertisements negatively and 19% view it positively, in Gallup and Bentley University's survey of 3,270 US adults fielded in May 2026 and published in August 2026. Acceptance depends on how far the AI goes: 75% accept it for brainstorming when disclosed, 53% for final creative, and 62% say AI-created people or voices are unacceptable.
How much of marketing budget goes to AI in 2026?
15.3% of marketing budgets, according to Gartner's 2026 CMO Spend Survey of 401 CMOs fielded January to March 2026. The same survey found 70% of CMOs call AI leadership a critical goal for 2026 while only 30% report mature AI readiness.
What performance lift should I expect from AI advertising tools?
Single digits. The median of the five AI advertising lifts published in 2026 is 8.3%, computed here from Meta's earnings disclosures and the Marketing Science field experiment. Three of the five inputs are Meta's own self-reported figures, so treat the median as a platform-weighted estimate. BCG found only 31% of B2C CMOs and 20% of B2B CMOs report significant, measurable revenue impact from AI.
Is AI improving advertising measurement?
Not yet. Between 60% and 75% of advanced measurement users say current approaches fall short on rigour, timeliness, trust and efficiency, and no respondents believe all paid channels are well represented in their marketing mix models, in the IAB's State of Data 2026 survey of more than 400 decision-makers. The same buy side thinks better AI measurement could release $26.3 billion in media investment within one to two years.
Sources
- Meta Platforms, Q2 2026 earnings call transcript (call held 29 July 2026)
- Alphabet, Sundar Pichai's Q2 2026 earnings remarks (July 2026)
- Google, Google Marketing Live 2026 announcements (May 2026)
- WARC, The Future of Media 2026 press release (January 2026)
- IAB, 2026 Outlook Study (January 2026)
- IAB, State of Data 2026: The AI-Powered Measurement Transformation (February 2026)
- IAB, Project Eidos announcement, carrying the State of Data 2026 figures (February 2026)
- IAB and PwC, Internet Advertising Revenue Report: Full Year 2025 (April 2026)
- eMarketer, US AI Advertising Forecast 2026 (June 2026)
- Gartner, 2026 CMO Spend Survey (May 2026)
- BCG, Moving the Agentic Marketing Transformation from Illusion to Reality (June 2026)
- Kantar, US Media Reactions 2025 (September 2025)
- Kantar, Rethinking AI-generated advertising (November 2025)
- Gallup and Bentley University, Americans Aren't Sold on Businesses Using AI in Advertising (August 2026)
- Pew Research Center, Americans and AI 2026 (June 2026)
- ANA, Q2 2025 Programmatic Transparency Benchmark (August 2025)
- ANA, Q4 2025 Programmatic Transparency Benchmark (February 2026)
- Nielsen, 2025 Annual Marketing Report (May 2025)
- Remi Daviet and Yohei Nishimura, "Leveraging Generative Artificial Intelligence to Create Visual Content in Digital Advertising", Marketing Science, doi:10.1287/mksc.2024.1130 (August 2026). The journal page blocks automated access, so the figures here are taken from the INFORMS release below, which reproduces them.
- INFORMS release on the Daviet and Nishimura study, via EurekAlert (4 August 2026)
- The CMO Survey, CMOs Face Headwinds Even as Marketing Value and AI Impact Grow (March 2026)
Methodology. Numbers were collected in September 2026 and checked against the publisher's own page, filing or press release, not against statistics roundups. Meta and Google performance claims are labelled as the platform's own figures wherever they appear, and no platform lift claim is presented as independent evidence. Gallup, Kantar and Pew data measures perception, not advertising performance. Kantar's 41% of consumers bothered by AI-generated ads comes from 2024 fieldwork and is the oldest number here. More pages like this one sit on the statistics hub. Last updated 22 September 2026.