AI Video Statistics 2026: 100+ Verified Facts and Costs

AI video statistics show a category growing faster in money than in trust. AI video companies raised $3.08 billion in 2025, almost double the year before, and Google's Flow tool alone produced more than 500 million videos in its first six months. Yet only 3.5% of consumers in one 2026 survey fully trust a video made entirely by AI, and most people cannot reliably tell when they are watching one.

This page collects more than 100 AI video statistics on the market, the companies, adoption, advertising, cost per minute, quality, trust, deepfakes and platform rules. We checked every figure on the publisher's own page in November 2026, and we flag vendor surveys, small samples and estimates as we go. Where a widely repeated number could not be traced to its source, we say so rather than repeat it.

AI video statistics at a glance

• Market estimates for AI video generators in 2025 range from $716.8 million to $5.39 billion, a 7.5x spread for the same market.
• Funding for AI video companies reached $3.08 billion in 2025, up 94.6% on 2024.
• 62% of US video ad buyers use generative AI to build or enhance video ads in 2026, up from 51%.
• Surveys put the share of marketers using AI for video anywhere from 9% to 84%, depending on who was asked.
• Generated video costs roughly $4 to $30 per minute at public API and plan prices; avatar video costs about $1.50 per minute.
• On a new YouTube account, 21% of the first 500 Shorts were AI slop; on TikTok, 59% of the first 500 For You videos.
• 62% of Americans disapprove of online videos made with AI tools; 23% approve.
• Only 0.1% of people in one study correctly identified every real and fake image and video.
• OpenAI closed the Sora app on 26 April 2026, about seven months after launch.

AI video market statistics

1. Grand View Research values the AI video generator market at $788.5 million in 2025, rising to $3.44 billion by 2033 at a 20.3% annual growth rate (Grand View Research).
2. Fortune Business Insights puts it at $716.8 million in 2025 and $3.35 billion by 2034, an 18.8% annual growth rate, with North America holding 41% of the market (Fortune Business Insights).
3. Straits Research estimates $751.5 million in 2025 and $3.76 billion by 2034 (Straits Research).
4. Polaris Market Research estimates $793.5 million in 2025 and $4.02 billion by 2034 (Polaris).
5. Market Research Future is the outlier: $5.39 billion in 2025, growing to $62.89 billion by 2034 at 31.4% a year (Market Research Future).
6. Four of the five estimates sit between $717 million and $793 million for 2025. The firms do not agree on the biggest region either: Fortune Business Insights says North America, while Grand View, Straits and Polaris say Asia Pacific.

Market reports define "AI video" differently, so treat any single figure as one firm's estimate, not a measurement.

AI video company statistics: funding, revenue and users

7. Funding for AI-related video companies reached $3.08 billion in 2025, up 94.6% from $1.58 billion in 2024, according to Crunchbase data.
8. Higgsfield raised $400 million at a $5.4 billion valuation in August 2026 and said it had reached $700 million in annualized revenue and more than 30 million users (Higgsfield). In January 2026 it had reported a $200 million run rate.
9. Runway raised $315 million at a $5.3 billion valuation in February 2026, up from $3.3 billion in April 2025 (TechCrunch).
10. Synthesia raised $200 million at a $4 billion valuation in January 2026, nearly double its $2.1 billion valuation a year earlier, and says it is used by over 90% of the Fortune 100 (Synthesia).
11. Synthesia's CFO told CNBC the company had reached $150 million in annual recurring revenue and expected to pass $200 million during 2026 (CNBC).
12. HeyGen passed $200 million in annual recurring revenue in June 2026, doubling in eight months, with more than 30 million users and 118 million videos created (HeyGen).
13. Luma AI raised $900 million in a round led by Humain, owned by Saudi Arabia's Public Investment Fund, at a valuation CNBC confirmed as above $4 billion.
14. Kuaishou's Kling AI made more than RMB 850 million in revenue in Q2 2026, up more than 200% year on year, according to Kuaishou's unaudited Q2 2026 results.
15. Kling's monthly revenue passed $20 million in December 2025, a $240 million annualized run rate reached 19 months after launch. Kuaishou says Kling serves more than 60 million creators.
16. Mirage, maker of the Captions app, raised $75 million in March 2026; app tracker Appfigures estimated $28.4 million in in-app revenue over the previous year.
17. OpusClip took SoftBank Vision Fund 2 investment at a $215 million valuation and says it serves more than 20 million creators and businesses (OpusClip).
18. The Sora app had 627,000 iOS downloads in its first week, slightly more than ChatGPT's 606,000, according to Appfigures estimates.
19. By late January 2026, Sora had 9.6 million downloads and just $1.4 million in consumer spending, with installs falling 45% month on month.
20. OpenAI closed the Sora web and app experiences on 26 April 2026 and the Sora API on 24 September 2026 (OpenAI Help Center). Disney's planned $1 billion investment in OpenAI, tied to Sora character licensing, never closed.

The revenue figures above are self-reported and each company defines annual recurring revenue its own way, so they are not directly comparable.

How many AI videos are being made

21. Google's Flow tool produced more than 500 million videos in its first six months after launching in May 2025, and users had made more than 1.5 billion images and videos by February 2026 (Google).
22. More than 230 million videos had been generated with Veo 3 by Q3 2025, Sundar Pichai told investors.
23. Daily users creating videos in the Gemini app rose 40% after Google launched its Omni model in May 2026.
24. Kling had generated more than 600 million videos for over 30,000 enterprise users by January 2026.
25. TikTok had labelled more than 3 billion videos as AI-generated by July 2026, up from 1.3 billion in November 2025, using creator labels, detection and C2PA Content Credentials (TikTok).
26. On average, more than 1 million channels a day used YouTube's AI creation tools in December 2025, and over 6 million viewers a day watched at least 10 minutes of auto-dubbed content (YouTube).
27. Media generation in the Meta AI app rose more than tenfold after Meta launched its Vibes AI video feed in September 2025.
28. Google's SynthID has watermarked more than 180 billion images and videos across Google products.
29. The Captions app has been used to make more than 200 million videos, according to its CEO.

No platform publishes a running total of AI videos, so these figures are counts from individual tools, not a market-wide number.

What share of YouTube and TikTok is AI?

30. On a brand-new YouTube account, 104 of the first 500 Shorts (21%) were AI slop and 165 (33%) were "brainrot", in a hand-judged test by video tool maker Kapwing in October 2025.
31. Kapwing's TikTok test in May 2026 found 59% of the first 500 For You videos were AI slop, about three times its YouTube rate, and 57.4% in the Kids category.
32. Of the 100 fastest-growing YouTube channels in July 2025, 9 showed only AI-generated content, a Guardian analysis found. After the Guardian got in touch, YouTube removed three and cut two off from ad revenue.
33. Neither YouTube nor TikTok publishes the share of uploads or recommendations that are AI-generated. TikTok's cumulative label count is the closest official figure.

AI video adoption statistics for marketers and businesses

Adoption figures vary wildly because each survey asks a different group a different question:

Survey	Who was asked	Using AI for video
Semrush, 2026	224 SEO and content marketers	9% (video or audio production)
HubSpot State of AI in Marketing	1,754 marketers	33% (image or video generation)
Wistia State of Video 2026	Nearly 1,000 marketers	38% (active use in video production)
Synthesia L&D report 2026	421 L&D professionals	52% (video creation)
CMI B2B research for 2026	B2B marketers whose firms use AI	53% (images, video and visuals)
Wyzowl State of Video Marketing 2026	Video marketers	63% (have ever used AI for video)
HeyGen State of Enterprise Video 2026	Organizations, sample not disclosed	77% (use AI video in some way)
Animoto State of Video 2026	US marketers already using AI	84% (use AI to create videos)

34. 63% of video marketers have used AI to create or edit marketing videos, up from 51% a year earlier, the biggest rise in the report (Wyzowl).
35. 38% of marketers actively use AI in video production, down from 41%, while 62% use it or plan to within a year (Wistia).
36. AI avatars were the fastest-growing AI video use in Wistia's data, up 56% year on year; AI dubbing rose 29%, and captions remain the most common use.
37. 51% of marketers use AI to come up with ideas for videos and write their scripts, and 49% of companies have a dedicated AI budget, in a separate Wistia survey of 500+ marketers.
38. 33% of marketers use generative AI for image or video generation, second only to drafting and editing copy at 36% (HubSpot).
39. 53% of B2B marketers whose organizations use AI use it to create images, videos and other visuals, against 89% for written content (Content Marketing Institute).
40. Only 9% of SEO and content marketers use AI for video or audio production (Semrush).
41. 52% of learning and development teams use AI for video creation, 63% for voice generation and 38% for translation in Synthesia's AI in L&D Report 2026, a vendor survey its authors say likely over-represents early adopters.
42. 77% of organizations use AI video in some way and 50% use it regularly or extensively, in a vendor survey by HeyGen that does not disclose its sample.
43. 59% of organizations say they are likely to increase AI video investment in 2026, in the same HeyGen survey.
44. Companies with in-house video teams rose from 35.7% to 54.3% in two years, a shift Wistia attributes to AI.
45. B2B marketers expecting to increase agency spend on content creation fell from 41% to 26% between 2025 and 2026, though 93% still use agencies (Forrester).
46. Adoption of Vidyard's AI avatars grew 12x in eight months of 2024, across more than 940,000 business videos in its data (Vidyard).

For video adoption without AI in the mix, see our B2B video marketing statistics.

AI in video advertising statistics

47. 62% of US video ad buyers use generative AI to build or enhance digital video ads in 2026, up from 51% in 2025 (IAB, 360 buyers).
48. Among buyers using generative AI, the share of their video ads made or adjusted with it rose from 24% in 2025 to 33% in 2026, and they expect 43% in 2027.
49. In 2025, buyers projected generative AI would reach 40% of all video ads by 2026. Actual use is behind that projection, although the two figures use different bases.
50. 70% of large buyers, those spending $50 million or more a year, use generative AI for video ads, against 54% of smaller buyers.
51. 96% of small and mid-size buyers are not satisfied with how much they use generative AI for creative; 46% say proof of performance would help most.
52. 83% of ad executives say their company has used AI in the creative process, up from 60% in 2024, most often for social ads (85%) and display (73%), in IAB's AI Ad Gap study.
53. Cost efficiency is now the top benefit advertisers see in AI, cited by 64%, up from fifth place in 2024.
54. Nearly 10 million advertisers used at least one of Meta's generative AI ad creative tools in Q3 2026, and adoption of its video generation features grew 25% in the quarter (Meta Q3 2026 call).
55. Meta said the combined revenue run rate of its video generation tools hit $10 billion in Q4 2025 (Meta). Meta did not define the figure; it most likely measures ad revenue from campaigns using the tools, not sales of the tools themselves.
56. Ads made with generative AI average the 54th percentile for overall impact in Kantar's LINK database, against the 65th percentile for ads made without it (Kantar).
57. 89% of advertisers using generative AI in ads disclose it at least sometimes, but fewer than half always do.

What AI video costs per minute

Public prices for generated video, worked out per minute from each vendor's own pricing page in November 2026:

Tool and setting	Price per second	Price per minute
Gemini Omni 1.2 Flash, 720p with audio	about $0.07	about $4.17
Veo 3.1 Lite, 1080p with audio	$0.08	$4.80
Veo 3.1 Fast, 1080p with audio	$0.12	$7.20
Runway Gen-4.5, via API	$0.12	$7.20
Kling 4.0, 1080p, via API	$0.235	$14.10
Veo 3.1, 1080p with audio	$0.40	$24.00
Adobe Firefly, 1080p, Standard plan	about $0.50	about $30

58. Veo 3.1 costs $0.40 per second of 1080p video with audio on Google Cloud, so a single 8-second clip costs $3.20 (Google Cloud pricing). Google's table bills per "count"; Runway's API and Artificial Analysis both price the same model per second at the same rate.
59. Google's Gemini Omni 1.2 Flash costs about $0.07 per second of 720p video with native audio, and $0.209 at 4K, which is upsampled (Gemini API pricing).
60. Runway's Gen-4.5 uses 12 credits per second; the $12-a-month Standard plan (billed annually) buys about 52 seconds of it (Runway).
61. Kling 4.0 costs $0.157 per second at 720p and $0.235 at 1080p through Kling's own API, sold only in prepaid packages (Kling).
62. Adobe Firefly video uses 100 generative credits per second at 1080p, so the 2,000 monthly credits on the $9.99 Standard plan cover 20 seconds of 1080p video (Adobe).
63. Avatar video is far cheaper per minute: about $1.50 per minute for HeyGen's Avatar IV on the $29 Creator plan, and about $1.58 per minute on Synthesia's $19 Starter plan, which includes roughly 12 minutes a month.
64. OpenAI last charged $0.10 per second for Sora 2 at 720p before shutting the API down (OpenAI).
65. On Pika's $10 Starter plan, a 5-second 720p clip from ByteDance's Seedance 2.5 costs about 122 credits, around $2.03, according to Pika's own pricing page.

These prices buy raw generated clips. Scripting, editing, sound, captions and review are separate costs.

AI video quality and capability statistics

66. Wan 3.0 leads Artificial Analysis's blind-vote text-to-video leaderboard with an Elo score of 1175, statistically tied with Utopai X (1168) and MiniMax H3 (1161), when we checked on 9 November 2026.
67. Google's Veo 3.1 ranked 23rd of 32 on the same leaderboard, with an Elo of 965, while costing $24 per minute at 1080p.
68. ByteDance's Dreamina Seedance 2.5 leads image-to-video with an Elo of 1161, but it is also among the most expensive at $34.12 per minute.
69. Veo 3.1 generates clips of 4, 6 or 8 seconds, at up to 4K and 24 frames per second, with sound.
70. Runway Gen-4.5 generates clips of 2 to 10 seconds at 720p, and Kling 4.0 up to 30 seconds at 720p or 1080p.
71. Adobe Firefly video defaults to 5-second clips at 24 frames per second.
72. HeyGen supports more than 175 languages and dialects and videos of up to 30 minutes on its Creator plan; Synthesia offers more than 140 languages and voices.

So a 60-second video made purely from generated clips means stitching together between 2 and 12 separate generations, even at each model's maximum clip length, and every join is a consistency risk.

AI video productivity and time savings statistics

73. 93% of creators say creative AI helps them produce content faster, but 57% say its output needs moderate or extensive editing before it is ready (Adobe Creators' Toolkit Report, 16,000+ creators).
74. Creative professionals say AI saves them 17 hours a week on average, in an Adobe survey of 400+ creatives and 400+ marketers (self-reported, all creative work).
75. 78% of B2B marketers say AI has improved the speed of content creation; only 17% say it has improved quality, in Content Marketing Institute's research for 2027.
76. In a peer-reviewed study of AI-generated edit alternatives, editors answered comparison questions in 38 seconds on average against 74 seconds without, roughly half the time (VideoDiff, CHI 2025, 12 participants).
77. LinkedIn's video team saves about 1 hour per project with transcript-based rough cuts, in a vendor case study published by Descript.
78. 37% of organizations publish video without formal review at least sometimes, even though 84% call review before publishing essential or important (HeyGen survey).

We found no controlled study measuring the time saved on a complete AI-assisted video edit, only studies of single tasks and vendor case studies. For where the line sits in practice, see our breakdown of AI video editing vs a human editor.

Audience trust and attitudes toward AI video

79. 76% of US adults say it is extremely or very important to be able to tell whether pictures, videos and text were made by AI or by people, and 53% are not confident they could (Pew Research Center, 5,023 adults).
80. Americans disapprove of online videos made with AI tools by 62% to 23%, and 73% are very concerned about misleading deepfakes (YouGov, 1,091 adults).
81. Only 3.5% of consumers fully trust videos made entirely by AI, against 77.9% who trust videos featuring real people, and 36% say AI-generated videos lower their trust in a brand (Animoto).
82. 82% of ad executives think Gen Z and Millennial consumers feel positive about AI-generated ads, but only 45% of those consumers do; the gap widened from 32 to 37 points in a year, in IAB's AI Ad Gap study.
83. 39% of Gen Z feel negative about AI ads, almost double the 20% of Millennials.
84. 71% of young consumers believe they have seen an AI-made ad, up from 54% the year before.
85. 50% of US consumers would rather do business with brands that avoid generative AI in consumer-facing content (Gartner, 1,539 consumers).
86. Only 26% of consumers prefer AI-generated creator content, down from 60% in 2023 (Billion Dollar Boy, a creator marketing agency).
87. 93% of US consumers now question whether online content was made by a person or by AI, in a Kantar study with billboard company OUTFRONT Media.
88. More than 60% of consumers worry that generative AI could produce fake or misleading ads, in Kantar's survey of 21,300 people across 30 markets.
89. 52% of social media users are concerned about brands posting AI-generated content without disclosing it (Sprout Social).
90. Only 12% of people are comfortable with news made entirely by AI, rising to 19% when a human oversees it (Reuters Institute, nine countries).
91. Concern about fake news online rose to 62% globally in 2026, and 77% of people watch online news video every week.
92. In NielsenIQ's brain-scan research, consumers found AI-generated ads more "annoying", "boring" and "confusing" than traditional ads, and even high-quality AI ads triggered weaker memory activation (NIQ, 2024).
93. 73% of young consumers say knowing an ad was made with AI would raise or not change their likelihood to buy.

Trust is where AI avatars struggle most in B2B; our comparison of AI avatar video vs real video covers when each works.

Deepfake statistics

94. Only 0.1% of 2,000 UK and US consumers correctly identified every real and deepfake image and video, even when told to look for fakes, and people were 36% less likely to spot a fake video than a fake image (iProov, a detection vendor).
95. 1,276 people judging AI-generated and real images, video and audio scored close to chance, in research by Frank and colleagues published in Communications of the ACM.
96. The FBI logged 22,364 AI-related complaints in 2025, with losses of more than $893 million (FBI IC3 2025 report).
97. Investment scams with a reported AI link cost more than $632 million in 2025; the FBI says scam "investment clubs" use AI-generated videos and voices of celebrities and CEOs.
98. Deloitte projects generative AI could push US fraud losses from $12.3 billion in 2023 to $40 billion by 2027 (Deloitte).
99. Deepfakes made up 11% of identity verification firm Sumsub's top fraud schemes in 2025.

Platform rules for AI video

100. YouTube requires creators to disclose realistic AI-generated or altered content through the "AI use" setting, and says disclosure alone does not limit reach or earnings (YouTube Help).
101. On 15 July 2025, YouTube renamed its "repetitious content" monetization policy to "inauthentic content". Mass-produced AI videos from generic templates do not earn money; AI used for scripts, backgrounds or your own characters is allowed.
102. TikTok requires labels on realistic AI-generated content and is testing a control that lets people see more or less AI content in their feed.
103. Meta adds "AI info" labels when it detects AI signals and may require labels on photorealistic video and realistic audio (Meta Help Center).
104. LinkedIn shows a Content Credentials icon on images and videos signed with C2PA data, but says it cannot yet identify all AI-generated content (LinkedIn Help).
105. YouTube expanded likeness detection, which flags AI videos using a creator's face, to all creators aged 18 and over in May 2026.

Numbers we could not verify

Several AI video figures circulate widely but could not be traced to the named publisher:

• An "AI video market" of $42.29 billion by 2030, credited to Grand View Research, does not appear on its site; the matching page says the report is still being developed.
• "30% of video ads built with generative AI in 2025, rising to 39% in 2026" comes from press coverage; IAB's own reports give 24% and 33% among buyers already using generative AI.
• Market figures attributed to MarketsandMarkets cover image and video generators together, and its own page was not accessible.

What the AI video statistics mean for B2B teams

The numbers point one way. Generating footage is now cheap, from about $4 a minute, and nearly every team is experimenting. But the same studies show the gap is in finishing: 57% of creators say AI output needs real editing, only 17% of B2B marketers say AI has improved quality, and AI-made ads score lower on impact than ads made without it. Audiences notice, and many of them mind.

The practical split for B2B video is to let AI handle drafts, transcripts, translations and rough cuts, and to keep people on the parts viewers judge you by: structure, brand fit, accuracy and the final review. For a buyer's view of the tools themselves, see our guide to the best AI video editing tools for business.

How we verified these AI video statistics

Every figure on this page was read on the publisher's own page, report, filing or earnings release between 7 and 9 November 2026. We used company announcements for company numbers, and major outlets such as TechCrunch and CNBC where the company did not publish the figure itself. Prices come only from vendors' own pricing pages, and per-minute costs are our own arithmetic from those prices. We label vendor surveys, small samples and estimates, and we left out anything we could not confirm at the source.

Where Pixel8 fits

Pixel8 Production is a video editing agency for B2B brands, SaaS companies and agencies. We do not film. We edit the footage you already have into finished videos, with 48-hour turnaround. Pricing is a $900 one-time trial, then $2,000 to $3,200 per month; see pricing.