State of
Generative
Media

From e-commerce teams to visual designers, anyone can now generate hundreds of production-ready images in minutes. A few years ago, this volume would have required thousands of photographers, studios, and production staff. The cost structure that has governed e-commerce and other digital verticals has shifted. Traditional barriers around content production are evaporating, due to generative media infrastructure.

The primary impact of breakthroughs in generative technology is the expansion of creative potential for users and builders alike. Entertainment applications initiated adoption of generative media, but in 2025 production applications (e.g. e-commerce, advertising, creative studios) drove scale[1] and by year’s end 88% of organizations deployed AI in at least one business function[2].

Jeffrey Katzenberg articulated the fundamental transformation:

It’s the democratization of storytelling at a level that has never happened in the existence of humankind.

This shift emerged from rapid advances in generative technology, as models reached levels of quality, controllability, and reliability once reserved for specialized production teams.

This report examines how generative technology and trends accelerated in 2025. These insights draw heavily from revelatory survey data collected across a diverse range of organizations and individual users[4]. We’ve included highlights from industry leaders who spoke at the Generative Media Conference in October, as well as the most impactful market research covering the changing landscapes in generative technology. Generative media is changing how we tell stories, build businesses and engage with users.

This signals the start of a new chapter in the digital age.

Enterprise generative AI adoption accelerated through 2025, with adoption rates varying by industry vertical and use case. Personal users bypassed technical requirements through emerging consumer applications, enabling immediate access without required expertise. Organizations faced distinct barriers: model orchestration complexity, integration decisions and cost management all constrained deployment pacing. Businesses used two pathways for access to generative technology, splitting evenly between applications (65%) and APIs (62%)[4], with many using both.

Production deployment maturity varied by modality. 31% of organizations are still in the prototyping phase of deploying generative models into their workflows[4]. Creative teams gravitated toward generative applications for rapid iteration without code, while engineering organizations prioritized API integration for programmatic control and workflow automation.

As frontier model access becomes increasingly commoditized, adoption is expanding beyond early entertainment-led experimentation. Organizations across advertising, e-commerce, and creative production are moving toward reliable production infrastructure, where consistent performance, scalability, and cost efficiency matter most.

Enterprise ROI

Return on generative media investment materialized faster than expected for new enterprise software technology[4]. The details, however, reveal return on investment is still split: organizations achieving strong ROI concentrate on specific high-value use cases with clear metrics, while those pursuing broad experimentation report disappointing returns[9].

Organizations reporting measurable ROI focused on three categories: efficiency gains, cost reduction, and revenue expansion. 74% of companies report their initiatives meet or exceed ROI expectations[9]. For the creative marketing platform Pimento, results were achieved by eliminating cold-start delays rather than maximizing quality, since marketers needed to test dozens of variations quickly before increasing sophistication and fidelity. Deployment reduced generation times by 80%, doubling their feature shipping pace[21].

Game studios need speed more than hosting control, as competitive advantages came from offering the latest capabilities before competitors. The digital creative platform Layer built on this insight, enabling a lean team to release new models to studios within 24 hours[22].

Organizations achieving generative scale made structural changes beyond deploying new technology. 43% redesigned workflows and production pipelines, 33% invested in staff training and upskilling, and 30% allocated dedicated budget for media generation infrastructure[4].

Industry deployments

Advertising agencies

Marketing organizations showed 75% generative AI adoption, up from 61% in 2024. However, 80% reported using AI on less than half of their work. Legal concerns dominated hesitation: 94% cited intellectual property ownership and liability as implementation challenges[10]. Integration with existing creative workflows (e.g. through Adobe Creative Suite, DAM systems or campaign platforms) proved more challenging than anticipated.

Agencies achieving scale implemented generative media for content variation and A/B testing rather than primary asset creation. 72% of marketers identified GenAI as the most important trend for H2 2025[11], yet only 30% achieved full integration across the campaign lifecycle[12]. This gap showed pressing infrastructure needs: programmatic generation at campaign scale, brand consistency enforcement, and audit trails for legal compliance in certain industries.

E-commerce platforms

E-commerce platforms demonstrated high adoption rates, with product image generation becoming a core infrastructure capability. Matt Koenig articulated the critical constraint that distinguishes e-commerce from other verticals:

The creativity of models absolutely cannot interfere with product fidelity. Images and videos must have a faithful representation of every product.

Film and production studios

Film and television production exhibited cautiously optimistic adoption in major operational workflows. Major studios allocated less than 3% of production budgets to generative AI while shifting 7% of operational spending toward AI-enabled tools for contract management, permitting, and planning[13]. Independent studios followed different patterns: 65+ AI-centric film studios have launched since 2022[15] employing generative AI throughout production pipelines.

The 68% adoption rate among all media companies[14] reflects deployment in pre-visualization, automated editing, and post-production VFX augmentation, rather than primary content creation. Still, media companies’ AI spending is projected to grow at 37.2% CAGR between 2024-2029, from $2.6B to $12.5B[14], signaling sustained investment despite currently conservative production budgets.

The varying rates of generative adoption emphasize how established studios are optimizing operational costs, while new entrants compete on production economics restructured by new capabilities. Katzenberg highlighted the underlying institutional constraint at the Generative Media Conference:

The greatest innovations that occur, they don’t happen within the legacy enterprises. They’re just not able to let go of the past and innovate into the future.

Gaming companies

Gaming studios showed strong adoption, with 68% actively implementing AI in workflows[17]. Gaming’s generative media growth came from rapid iteration requirements incompatible with traditional asset timelines, and technical infrastructure focused on more predictable generations. 40% of studios experienced productivity gains exceeding 20%, with 25% achieving cost savings above 20%[18].

Use cases ranged from concept art acceleration to texture generation, NPC dialogue variation, animation in-betweening, and procedural level generation augmented with AI-generated details. Infrastructure selection emphasized generation speed (41%) and reliability (37%) over maximum quality[4]. This preference explained gaming’s significant adoption of generative media APIs: programmatic control enables integration into development pipelines.

Gaming’s infrastructure requirements differ from other verticals: real-time generation for dynamic content, batch processing for asset libraries, and game engine integration. There is high demand for infrastructure that enables rapid model deployment and scales with unpredictable player loads. Burkay Gur outlined a future for generative gaming:

Text-to-game will be the continuation of text-to-video; it’s essentially making the video output interactive. We are not too far away from that. It’s a great use case for world models.

This convergence of video and interactivity represents a fundamental shift from content creation to world simulation. Text-to-game capabilities would enable dynamic, AI-generated game environments responding to player actions in real-time, transforming gaming from pre-authored experiences to emergent narratives.

Education

The education sector represents one of the largest untapped opportunities for generative media, combining massive market size with historically limited technology adoption. Sonya Huang articulated this potential:

I’m most excited for the education use case. Education is a market that is so important and has never had that many compelling business cases behind it. The challenge is the bottleneck to create high quality content at scale that is most ideal for the learner.

Traditional educational content production faces the same constraints Huang identified; creating high-quality, personalized content at scale has been economically prohibitive. Gorkem Yurtseven had an equally strong prediction on generative AI in education:

The education market is almost untouched right now with video generation. And there is so much potential there, it’s just waiting for the quality and predictability to open up new use cases.

Current limitations, particularly consistency and controllability, constrain educational deployments. Educational content requires factual accuracy, cultural sensitivity, and curriculum coherence across multi-week lesson sequences. As these capabilities mature, education could become one of the largest generative media markets, driven by the need for personally optimized learning at massive scale.

The trajectory of generative media development through 2026+ is clear. Three major themes will dominate:

• 1) Multimodal advancements (e.g. world models)
• 2) Infrastructure optimization
• 3) Democratization of creative tools

Expertise will become orchestration rather than execution. Taste becomes scarce, while capability becomes abundant. As technical capabilities are more commoditized, the fundamental value proposition shifts. "It’s the storytelling that matters."[19]

The creative democratization that Katzenberg described will manifest in new forms. Solo entrepreneurs will generate visual content indistinguishable from large-scale production companies. Successful organizations building generative technology into their products will compete on orchestration, deployment reliability, and domain-specific optimization.

The durable competitive moats will belong to teams that understand how to best deploy generative media, now that generating professional media is easier than ever.

State of Generative Media Volume 1, published 2026