State of Generative Media Survey Report 2025
Supported by Fal
ImageVideo
A comprehensive look at the generative media landscape in Q3 2025, based on our survey of personal and organizational users.
Key Takeaways
Image Generation sees strong personal adoption (89%), while video lags behind but is catching up.
Quality is the #1 factor for model selection for both personal and organizational use. Cost is critical for users selecting Video Generation APIs.
Google's models (Gemini for image, Veo for video) are the most popular choices among respondents.
Organizations are seeing fast ROI, with 65% reporting returns within 12 months.
About This Survey
We conducted a survey to understand the current state of Generative Media in Q3 2025. The survey collected responses from both personal users and organizations to understand adoption patterns, use cases, and preferences across image and Video Generation models. The survey was conducted before the release of Sora 2 and the OpenAI Video API.
Adoption Maturity
BothImage Generation is more commonly adopted than video generation, personal adoption outpaces organizations
Do you use [image/video] generation?
Personal (n=156)
Organization (n=138)
BothPersonal: Image generation usage is more mature than video, with more integration into creative workflows
Maturity - Which of the following best describes your current use of [image/video] generation models?
Image (n=126)
Video (n=78)
BothOrganizations: Organizations show similar deployment patterns for image and video, with 44% (Image) - 39% (Video) already in production
Maturity - Which of the following best describes your organization's current use of [image/video] generation models?
Image (n=78)
Video (n=41)
Use of Models
BothPersonal: One-off personal projects dominate both image and video model use amongst respondents, entertainment and storytelling follow
Current use – What do you primarily generate [image / video] for? (Select all that apply)
Image (n=130)
Video (n=88)
BothOrganizations: Marketing leads video model adoption at 55%, while image use is split between marketing, entertainment, and creative tools
What are the main end use cases of the [images/videos] generated by your organization? (Select all that apply)
Image (n=79)
Video (n=42)
Choice of Models
ImageImage: Google Gemini leads image model adoption at 74%, followed closely by OpenAI, then by Black Forest Labs (FLUX)
Models Used – Which Image Generation model families [are you / is your organization] currently using? (Select all that apply) (n=209)
ImageImage: Google Gemini leads both personal and organization adoption at 77% and 68% respectively. Organization and personal adoption between offerings is closely aligned
Models Used – Which Image Generation model families [are you / is your organization] currently using? (Select all that apply)
Personal (n=132)
Organizations (n=77)
VideoVideo: Google leads video model adoption at 69%, followed closely by Kling and then Hailuo, Runway and Alibaba
Models Used – Which Video Generation model families [are you / is your organization] currently using? (Select all that apply) (n=129)
VideoVideo: Google's Veo leads both personal and organization adoption at 67% and 72% respectively. Organization and personal adoption between offerings is closely aligned
Models Used – Which Video Generation model families [are you / is your organization] currently using? (Select all that apply)
Personal (n=86)
Organizations (n=43)
BothPersonal: Quality is the most important consideration for personal use respondents at 76%, with cost (46%) a larger consideration than speed (37%) in model selection
Model Criteria – What factors most influence your choice of model for [image/video] generation? (Select all that apply)
Image (n=131)
Video (n=82)
BothOrganizations: While output quality is still the greatest driver of model choice (82% for image), cost still remains a large consideration (55%), speed of generation is a much smaller consideration for image than video for organizations
Model Criteria – What factors most influence your organization's choice of model for [image/video] generation? (Select all that apply)
Image (n=78)
Video (n=43)
Access Channels
BothPersonal: Personal users overwhelmingly access both image and video models directly through applications; APIs are second and relatively few self-deploy
Access Channels – How do you primarily access [image/video] generation models? (Select all that apply)
Image (n=139)
Video (n=87)
BothOrganizations: Organizations split nearly evenly between applications (65%) and APIs (62%), making both important ways to access the enterprise market
Access Channels – How does your organization primarily access [image/video] generation models? (Select all that apply)
Image (n=78)
Video (n=44)
BothPersonal: Personal users prioritize model availability for Image Generation (73%) while cost/price is most important for Video Generation (61%)
Decision Criteria – Which of the following are most important when you select your access channel(s) for [image/video] generation? (Select all that apply)
Image (n=131)
Video (n=87)
BothOrganizations: Cost is the primary driver of access channel decision making for 58% of respondents, outweighing model availability (49%), and generation speed (41%)
Decision Criteria – Which of the following are most important when your organization selects your access channel(s) for media generation? (Select all that apply)
Organizations (n=132)
ImageImage Applications: ChatGPT (ChatGPT) and Google (Gemini) stand out as the clear leader amongst applications used for generation
Application Providers – Which application provider(s) [do you/ does your organization] use for Image Generation? (Select all that apply) (n=167)
ImageImage Applications: ChatGPT and Google products lead in personal adoption at ~70%; Adobe Suite and Midjourney see outsized usage in organizations
Application Providers – Which application provider(s) [do you/ does your organization] use for Image Generation? (Select all that apply)
Personal (n=116)
Organizations (n=51)
ImageImage Endpoints: Google AI Studio is the leader in API, followed closely by Fal.ai, OpenAI and Replicate
Endpoint Providers – Which endpoint provider(s) do you use for Image Generation? (Select all that apply) (n=96)
ImageImage Endpoints: Google leads in personal adoption whereas Fal.ai leads in adoption amongst organizations for image APIs
Endpoint Providers – Which endpoint provider(s) do you use for Image Generation? (Select all that apply)
Personal (n=50)
Organizations (n=46)
VideoVideo Endpoints: Google AI Studio is the leader in Video APIs, followed closely by Fal.ai and then by Replicate, Kling, and MiniMax
Endpoint Providers – Which endpoint provider(s) do you use for Video Generation? (Select all that apply) (n=57)
Organizational Return
Both65% of organizations report ROI within 12 months, 34% already profitable, 31% expecting returns soon
ROI timeframe – When do you expect to see return on investment (ROI) from your organization's media generation initiatives?
Organizations (n=128)
BothOrganizations are making changes to their processes, teams and budgets to take advantage of media generation models
Organizational changes – What organizational changes has your company made to support adoption of media generation? (Select all that apply)
Organizations (n=138)