The Death of Stock Photography (and What Replaced It)

Getty Images’ annual revenue peaked in 2021 at approximately $935 million. By 2026, the company’s revenue had declined to an estimated $540 million — a 42% drop in five years, driven primarily by the collapse of the commercial stock photography market. Adobe Stock and Shutterstock reported comparable declines. Corbis, which had been acquired by Visual China Group, essentially ceased its independent commercial operations.

This is a clean economic story: AI image generation crossed a quality threshold, prices fell to near zero, and a market that had supported hundreds of thousands of photographers worldwide contracted sharply. The clean version is also incomplete, because the market that replaced stock photography is not nothing — it is a different market with different winners and different rules.

The Quality Threshold

The specific claim that needs examination: AI image generation crossed the quality threshold for commercial use.

The threshold is not uniform. For a photograph that needs to be realistic and specific — a particular building in a particular city, an identifiable person, a documented event — AI generation cannot substitute because it can only generate plausible fictional instances, not actual photographs of real things. The photojournalism market has not been disrupted by AI generation; photojournalists’ work is valuable precisely because it is documentary, and the value of documentary photography has if anything increased as AI-generated imagery has proliferated.

The threshold was crossed for a different and much larger market: generic illustrative imagery for digital content. The blog post that needs a “person at computer” header image. The marketing brochure that needs “diverse team in meeting.” The social media post that needs “beautiful landscape at sunset.” These applications required, until 2022, either buying a license from a stock library or paying a photographer to shoot to specification. They now require a text prompt and thirty seconds.

Midjourney version 6 (late 2023), the DALL-E 3 generation, and Stable Diffusion XL were the approximate threshold markers — the point at which the quality of AI-generated imagery for commercial illustrative purposes became sufficient to substitute for stock photography in most professional contexts. The transition was faster than anyone expected: Adobe Stock, which had introduced an AI generation service in 2023, reported that AI-generated images constituted over 50% of the images used by its subscribers by mid-2025.

What Happened to the Photographers

The impact on photographers depends heavily on which segment of photography is under discussion.

Stock photography contributors — the hundreds of thousands of photographers worldwide who supplied images to Shutterstock, Getty, and Adobe — have experienced significant income decline. The typical stock photographer on these platforms had built a passive income stream over years by uploading a large catalog of images that received micro-royalties on individual downloads. This stream has largely dried up. Shutterstock’s royalty payments to contributors declined by over 60% between 2022 and 2026 in aggregate, and per-image royalties declined even faster as the library deprioritized human-created content in favor of AI generation.

Editorial photographers — those who supply images of news events, public figures, and documentary subjects — have been essentially unaffected by AI generation for the reasons described above. Getty’s editorial photography business has held up considerably better than its commercial business. Associated Press, Reuters, and AFP have seen stable or slightly growing demand for their photojournalism.

Commercial photographers — those working on direct assignment for advertising, brand, and corporate clients — have experienced a bifurcated market. The low end of commercial photography (small businesses, local marketing, simple product shots) has seen significant price pressure because clients can generate adequate substitutes with AI. The high end (brand campaigns, luxury advertising, fashion editorial) has held up because clients explicitly want the credibility and specificity that human photography provides, and are increasingly using “shot by photographer X” as a brand attribute.

The New Visual Economy

What replaced stock photography is not a single thing but a cluster of smaller markets.

AI generation services are the most obvious replacement, but the economics are different from stock photography in a structural way: there are no individual creators being compensated. Midjourney, Stability AI, and Adobe Firefly collect subscription fees. The photographers whose work trained these models receive nothing (absent the licensing deals being litigated in copyright cases). The revenue that was distributed across hundreds of thousands of photographers worldwide is now concentrated in a small number of companies.

Authentic photography is a growing premium category. “AI-free” is now an explicit marketing claim in some advertising contexts — the equivalent of “organic” in food. Several advertising agencies have begun explicitly crediting photographers in campaigns and featuring the photography process as part of the brand story. This is a real premium market, but it is small relative to the volume of the old stock photography market.

The “personal photograph” market — images of real people’s real lives — has become more valuable as AI imagery has become ubiquitous. Authentic personal photography for social media, for family archiving, for personal documentation has seen increased investment from some users precisely as a differentiation from the flood of generated content. Portrait photographers and event photographers (weddings, corporate events) have experienced stronger demand than before, partly because human presence and documentation carry a value that generated imagery cannot replicate.

The Aggregation Problem

The stock photography market, for all its imperfections, had one structural virtue: it distributed income across a global community of photographers. The images in a stock library represented the work of photographers on every continent, at every career stage, shooting in every context. The economics were often exploitative — the platforms took large percentages, the royalties were small — but the distribution was wide.

The replacement economy is more concentrated. The companies profiting from AI image generation are fewer, richer, and their revenue flows differently. The photographers who trained the models are uncompensated (under current legal frameworks). The visual culture of AI-generated imagery is pulling toward a kind of aesthetic homogeneity that reflects the training data and the aesthetic preferences baked into the guidance systems — more saturated, more perfect, more generic than the idiosyncratic human photography it replaced.

Whether the overall visual culture is worse is a question that does not have a single answer. There is more imagery, more cheaply accessible. It is less diverse in origin and less documentary in nature. The photographers who built the stock library over thirty years are not replaced — they have skills and perspectives that still have value in specific contexts. The economic structure that compensated them for those skills, broadly and regularly, no longer exists.

The stock photography market was not destroyed by a failure of quality or relevance. It was destroyed by a sufficient drop in the cost of an adequate substitute. This is not unusual in economic history — the same thing happened to recorded music (the CD market) when digital distribution reduced distribution costs, and before that to live music when recording reduced performance costs. What is unusual is the speed: the transition from healthy market to structural disruption took approximately three years. That pace is the story.