Content & Media
How many AI images are generated every day?
Thirty billion AI images in three years. That is the rough total since Stable Diffusion launched in August 2022 and put text-to-image generation on anyone's consumer GPU. By 2025/2026, the daily volume sits at about 80 million images, around 926 per second, according to Everypixel Journal. Because Stable Diffusion runs locally without any central log, all estimates carry significant uncertainty.
Roughly 926 images every second.
Source: Everypixel Journal AI Image Statistics 2024; platform estimates. AnythingCounter overview
What this means for you
In 2022, virtually every image you encountered online was taken or designed by a human. By 2024, AI-generated images had become part of standard workflows across stock photography, marketing, and journalism. The ratio to human-produced images keeps shifting.
For most people, this raises a practical question: can you tell the difference? In 2023 studies, participants correctly identified AI images only 50-70% of the time under relaxed conditions, and detection rates fell further when images were styled to look authentic. The "face test" (checking for distorted hands, misaligned symmetry, unnatural backgrounds) still works for low-quality outputs, but is increasingly unreliable for state-of-the-art models.
For creators and publishers: the legal landscape around AI-generated imagery (copyright ownership, disclosure requirements, platform rules) is still being defined. Using AI images in commercial contexts without understanding the terms of the model's licence carries real risk.
AI-generated images vs. selfies taken, today
In 2022, AI-generated images barely existed. By 2024, machines produced one AI image for roughly every three human selfies taken. The gap keeps narrowing.
The scale: key statistics on AI image generation
Approximately 80 million AI images are generated every day globally as of 2025/2026, ~926 per second (Everypixel Journal 2026)
EverypixelOver 30 billion AI images have been created in total since 2022 – roughly the same number as photos taken by humans in the first 150 years of photography (Everypixel Journal 2025)
Everypixel34 million AI images were created every day globally in 2024, 393 per second
EverypixelThe March 2025 GPT Image 1 viral launch temporarily pushed daily AI image generation to ~100 million images per day
EverypixelStable Diffusion accounts for ~80% of all AI image generation (12.59 billion of 15B total through 2024)
EverypixelAdobe Firefly reached 1 billion AI images generated in just 3 months after its 2023 launch
EverypixelThe AI image revolution: creation, deception, and the future of visual media
The text-to-image revolution
AI image generation was transformed by diffusion models in 2021-2022. Unlike previous GAN-based systems, diffusion models (DALL-E 2, Stable Diffusion, Midjourney) can generate photorealistic, artistic, or stylised images from plain-language text prompts in seconds. The August 2022 open-source release of Stable Diffusion changed the scale: for the first time, anyone with a consumer GPU could generate images locally, without a subscription or API. DALL-E alone was generating over 2 million images per day by September 2022; two years later the combined daily volume across all platforms had climbed to 34 million.
Scale and creative disruption
Since 2022, over 30 billion AI images have been generated, roughly the same number as photos taken by humans in the first 150 years of photography. At 80 million per day, AI image generation has surpassed professional stock photography output by orders of magnitude, which has put real pressure on Getty Images, Shutterstock, and the concept art sector. Illustrators have lost contracts to it. Others use it to produce work they could never have afforded to commission. Both things are true at once.
Growth of AI image generation: 2022-2025
AI image generation went from a research curiosity to a mass-market tool in under three years, with daily generation volumes growing from near zero in 2021 to 34 million images per day by 2024, and approximately 80 million per day by 2025/2026.
| Year | Rate | Est. per day | Context |
|---|---|---|---|
| 2022 | 50/sec | 4M | Stable Diffusion open-sourced; adoption explosive |
| 2023 | 145/sec | 13M | Midjourney, SDXL, Firefly all launch |
| 2024 | 394/sec | 34M | Integrated into phones, apps, workflows |
| 2025/2026 | 926/sec | 80M | GPT Image 1 launch; smartphones, apps and workflows fully integrated |
| 2028 (forecast) | 2K/sec | 150M | AI image generation embedded in every creative app and smartphone |
Key moments in AI image generation history
- 2021DALL-E (OpenAI) and VQGAN+CLIP demonstrate text-to-image publicly; AI image generation enters awareness
- 2022Stable Diffusion open-sourced (August); DALL-E 2 public release; Midjourney public beta; mass adoption begins
- 2022DALL-E 2 generates 2+ million images/day by September 2022
- 2023Adobe Firefly: 1 billion images in 3 months; AI images in mainstream advertising
- 202434 million AI images/day; 15 billion total since 2022; image generation integrated into phones, apps, and workflows
- 2025GPT Image 1 (OpenAI, March 2025) launch drives temporary peak of ~100M images/day; Everypixel Journal 2026 reports established baseline of ~80M images/day; 30+ billion total since 2022
The volume in perspective
At 80 million AI images per day, humanity now creates more AI-generated images in a single day than professional photographers produced in the entire decade of the 1980s
30 billion AI images created since 2022, enough to give every person on Earth roughly four AI-generated images each
AI image generation reached 30 billion total images in ~3 years; traditional photography took 150 years to reach the same volume
Research on AI image generation
| Year | Finding | Value | Source |
|---|---|---|---|
| 2022 | Stable Diffusion launched Aug 2022; DALL-E 2 generates 2M+ images/day by Sept 2022; Midjourney reaches 964M images in first 13 months | 2M DALL-E images/day (Sept 2022) | Everypixel |
| 2023 | Adobe Firefly reaches 1 billion AI images in 3 months post-launch; Midjourney market share 26.8% | 1.0B Adobe Firefly total images (first 3 months) | Everypixel |
| 2024 | Total ~34 million AI images/day; 15+ billion created since 2022; Stable Diffusion 80% market share | 34M AI images/day | Everypixel |
| 2025 | Everypixel Journal 2026: ~80 million AI images/day across all platforms; 30+ billion total since 2022; GPT Image 1 launch (March 2025) drove temporary peak of ~100M/day; 150M+ monthly users of AI image generators | 80M AI images/day (2025/2026 baseline) | Everypixel |
Frequently asked questions
- How many AI images are created per day?
- Approximately 80 million AI images are created per day globally as of 2025/2026, according to Everypixel Journal analysis. This equals about 926 images per second. In 2024, the figure was 34 million per day (393/sec); the increase reflects wider platform integrations and the March 2025 GPT Image 1 launch.
- Which platform generates the most AI images?
- Stable Diffusion-based models (running locally, on cloud services, and embedded apps) account for approximately 80% of all AI image creation through 2024, roughly 12.59 billion of the ~15 billion total at that time. Since then, OpenAI's GPT Image 1 and integrated smartphone tools have grown significantly.
- How does AI image volume compare to photography?
- By 2025/2026, approximately 80 million AI images are created daily, while conventional photography produces about 5.3 billion photos per day. AI images represent roughly 1.5% of daily image creation, growing rapidly from 0.6% in 2024. Total AI images since 2022 have passed 30 billion, matching roughly 150 years of traditional photography.
Cite this rate
AnythingCounter derived rate, last checked September 2026: about 926 images per second (about 80 million per day).
Method: Based on Everypixel 2026, converted to a per-second rate (daily total ÷ 86,400).
Suggested citation: AnythingCounter, "AI images spawned," September 2026, https://anythingcounter.com/ai-images-generated-per-day
Primary source: Everypixel.
How this number is calculated, and how much to trust it
Everypixel Journal's AI Image Statistics 2026 reports approximately 80 million AI images generated per day across all platforms. Dividing 80,000,000 by 86,400 seconds yields the live rate of approximately 926 images/sec. The prior 2024 figure was 34 million/day (393/sec); the 2025/2026 increase reflects broader platform integrations (GPT Image 1, smartphone AI tools, expanded Midjourney access). Stable Diffusion's open-source distribution makes precise counting impossible; the Everypixel estimate is the most-cited public figure.
The Everypixel estimate draws on platform APIs, Stable Diffusion community metrics, and direct data from commercial providers. The 2024 baseline (34M/day) from the same source was corroborated by Information Is Beautiful and MIT Technology Review. The March 2025 GPT Image 1 launch temporarily drove daily volumes to ~100 million; the 80M figure represents the 2025/2026 established baseline. Given the open-source nature of Stable Diffusion, all estimates carry significant uncertainty.
Sources: see below. More detail on our methodology page.
Worth a look... Deepfakes created - Fake news shared - Selfies taken.