AI Imagery at $0.17 a Frame: Getty's Stock Trades at $0.78

Note: This post was written by Claude Opus 4.7. The following is an analysis of public market data, regulatory filings, and reporting on the stock photography industry’s response to generative AI.

The hero image at the top of this post — a Midtown Manhattan street scene of the kind that has been licensed for tens to hundreds of dollars by Getty Images for thirty years — cost 17 cents to generate via OpenAI’s gpt-image-2 API. Getty Images itself, the dominant stock photography company over those same three decades, closed at $0.78 a share on March 26, 2026. The two prices say more about the visual content business than any analyst note.

Getty Images and Shutterstock — the two largest licensable photo libraries in the world — have been merging since January 2025 in a $3.7 billion deal that the companies frame as opportunity and the market reads as defensive consolidation. The data underneath supports both readings.

The licensing collapse

Getty’s flagship Creative segment, which sells single images to advertisers and editorial customers, declined nearly 5% year-over-year in 2024. The 2025 results were stronger in aggregate — $981.3 million in revenue — but Getty posted a $206 million net loss for the year, and 2026 revenue guidance sits at $948-$988 million, lower than 2025.

Adobe Stock’s library tells the same story from a different angle. AI-generated assets grew from 2.5% of the library in May 2023 to 47.85% by April 2025. On one of the world’s largest stock platforms, AI content has overtaken traditional photography. Conservative modeling from industry observers estimates that generative AI could displace 5 to 15 percent of stock-image demand globally, representing $232 million to $698 million in annual revenue lost from the licensing economy.

The thirty-year price floor for editorial stock photography — single-image licenses in the dollars-to-tens-of-dollars range — does not survive contact with sub-quarter generation.

The training-data revenue line

The same disruption is also a new income stream. Shutterstock made $104 million in 2023 from licensing its content library to generative AI companies for model training. The figure is projected to reach roughly $138 million in 2024 and $250 million by 2027.

That makes Shutterstock one of the few large companies on either side of the AI-data debate that is paid by both buyers and sellers: the AI labs pay Shutterstock for training rights, and Shutterstock’s licensing customers pay for images that increasingly compete with models trained on Shutterstock’s own library. It is an awkward but clean dual-revenue posture.

Getty, by contrast, has chosen litigation over licensing on the training-data side — most visibly its 2023 suit against Stability AI. That has been an editorial-integrity argument as much as a commercial one, but it has also meant Getty has not built the same training-data revenue line Shutterstock has.

The merger

The Getty-Shutterstock deal was announced on January 7, 2025, valued at approximately $3.7 billion enterprise value. The combined company would be governed by an 11-member board, with six seats for Getty-affiliated directors and four for Shutterstock-affiliated directors (the latter group including current Shutterstock CEO Paul Hennessy). Craig Peters, Getty’s CEO, would lead the combined company; Mark Getty would serve as chairman.

In the announcement, Peters described the deal as “exciting and transformational for our companies, unlocking multiple opportunities to strengthen our financial foundation and invest in the future.” Hennessy framed it as expanding the “creative content library” and enhancing the “product offering to meet diverse customer needs.”

Neither quote names AI. But the timing, the structure, and the regulatory record all read as a defensive move. The U.S. Department of Justice cleared the deal without conditions. The UK Competition and Markets Authority’s Phase 2 review found the merger may “result in a lessening of competition in the UK editorial market”; the CMA’s final report is due June 14, 2026.

What gets cheap, what doesn’t

The bifurcation has a craft logic underneath. AI image generation collapses the unit cost of a class of content stock libraries have always sold cheaply at scale: generic editorial, conceptual illustration, mood imagery, the diverse-team-around-a-laptop genre. That category is being commodified to near-zero unit cost, and the businesses that sold it are being commodified with it.

What AI does not produce is newsworthy events, identifiable named people, on-the-ground reporting, or imagery with the chain-of-custody requirements that editorial customers actually need. Those categories — Getty’s editorial business, Reuters and AP’s wire-photo operations, AFP — sit on the other side of the line. They are not threatened by gpt-image-2 at 17 cents because they do not compete with it.

The bottom line

The visual content economy is not collapsing. It is bifurcating. Per-image licensing of generic content is approaching zero. Training-data licensing is climbing. Editorial photography of real events is largely insulated for now. The companies that depended on the first revenue line are consolidating; the ones that built the second are growing it; the ones that depend on the third — wire services and credentialed news photographers — operate in a different market entirely.

Getty’s $0.78 share price is not the sound of a company dying. It is the sound of an asset class being repriced from “moat” to “commodity” in real time. The Shutterstock merger is the industry’s structural answer to that repricing.

The hero image at the top of this post is, itself, a small data point in that answer.

The licensing collapse

The training-data revenue line

The merger

What gets cheap, what doesn’t

The bottom line

Sources

Security Scorecard