OpenAI locks down Shutterstock's library for 6 years

While the rest of the AI world spends millions fighting copyright lawsuits in court, OpenAI is taking out its checkbook and buying peace of mind for the rest of the decade.

Shutterstock just signed an expansive new six-year agreement with OpenAI, cementing itself as the primary training data engine behind future models.

Under the new deal, OpenAI secures broad licensing rights to Shutterstock's vast library of images, videos, and music, alongside all the valuable metadata that ties them together.

More than just static images

The two companies have been tight since 2021, and Shutterstock already baked DALL·E directly into its stock platform earlier this year. But this renewal goes way beyond photos.

OpenAI gets priority access to video and audio tracks, which is exactly the fuel required to train next-generation multimodal models.

In return, Shutterstock gets priority access to OpenAI's latest models and plans to roll out synthetic editing tools across its entire catalog. Users won't just generate images from scratch; they will be able to transform and remix any existing asset in the library.

Mobile users are getting a slice of the pie too. Shutterstock recently acquired GIF juggernaut GIPHY, and both teams are collaborating to bring generative AI features directly into GIPHY's mobile footprint.

To keep artists from revolting, Shutterstock pointed to its Contributor Fund, which it claims has already paid out compensation to hundreds of thousands of creators whose work helped train these engines, with ongoing royalties attached to generated asset licenses.

Why it matters

We think locking in six years of fully cleared media is a massive strategic moat for OpenAI.

As regulators circle and artists file lawsuits over scraped web data, having legitimate, commercially indemnified training data across video, audio, and images makes OpenAI's enterprise pitch almost impossible to beat. It basically forces competitors to either secure their own multi-year content deals or risk getting sued into oblivion.

For designers and marketers, synthetic editing inside stock libraries is where generative tech actually becomes useful. Prompting from a blank canvas is fun, but tweaking an existing high-res asset to fit a brand guideline in two clicks is real utility.

Still, the lingering question is the payout economics. Shutterstock is doing the right thing by compensating contributors, but whether those micro-royalties can replace traditional licensing income for working photographers remains to be seen.