Is Stock Photography Still Worth It in 2026? The Numbers Say Mostly No

TL;DR: Is stock photography still worth it in 2026? For a human contributor deciding where to spend new hours, mostly no. Adobe Stock's library went from about 2.5% AI-generated in May 2023 to 47.85% by April 2025. Adobe's own management told investors on its Q1 fiscal 2026 call that the traditional stock business declined more steeply than expected. Shutterstock's Content segment fell 12% in the first quarter of 2026, and Getty's Creative segment fell 4.5% in the same quarter after growing 0.7% for all of 2025. Agency revenue is flat to down while the pile of assets nearly doubled, so the money per file fell for everyone. Keep the portfolio you have. Do not build a workflow around growing it.

We built SpreadPix, at first, to automate stock uploads. The plan on our whiteboard in 2025 listed Adobe Stock and Shutterstock as the top targets, on the logic that Adobe paid more per download and that neither had a contributor API worth the name. Then we did the market work, and the conclusion was not that stock is hard to automate. It was that stock is the wrong thing to automate. This post is the evidence for that call, laid out for anyone with a thousand images on an agency wondering whether to keep going.

Is stock photography still worth it when half the library is AI?#

The number that reorganised our thinking is the library share. According to the figures compiled in our August 2026 market research, Adobe Stock's generative AI content rose from about 2.5% of the library in May 2023 to 47.85% by April 2025. That is not a niche. It is the other half of every search result a buyer scrolls.

Adobe said as much to its investors. On its Q1 fiscal 2026 earnings call, management stated that the traditional stock business saw a steeper decline than expected and that the shift was playing out more quickly than planned, with the Adobe Stock line falling by roughly $450 million even as Firefly, Adobe's own generative tool, grew 75% quarter over quarter. Read those two sentences together: the same company that runs the marketplace is growing the product that replaces the marketplace's supply.

The rest of the industry rhymes. Shutterstock posted record revenue for 2025 overall while its Content segment fell 12% in Q1 2026, meaning the growth is coming from somewhere other than licensing photographs. Getty's Creative segment grew 0.7% in 2025 and then fell 4.5% in Q1 2026. None of these agencies is failing. Their content businesses are simply not where the money is going.

The mechanism is arithmetic rather than malice. Agency content revenue is roughly flat to declining. The number of assets competing for it nearly doubled through AI submissions. Average earnings per file fell, and would have fallen even if every agency had left royalty rates untouched.

What does a contributor actually earn now?#

Rates did not stay untouched, though the big cut predates the AI wave. Shutterstock moved from a flat minimum per download to a percentage model in June 2020; the Wikipedia entry on Shutterstock records the entry-level rate dropping from a 25-cent minimum to 10 cents, or 15% of the sale. That tiered structure, 15% to 40% of the sale price depending on annual download volume, still stands, and the tier resets to the lowest level every January 1. The newer Contributor Fund pays small amounts, typically single dollars to low tens of dollars per payout, when content is used for AI training or licensed to AI companies. It is a fund for the thing that is displacing you.

The lived version comes from contributor forums. One long-time Adobe Stock contributor, cited in our research, reported monthly earnings falling from $1,000 to $1,200 down to $250 to $300 as rejection rates rose and AI assets filled the search results, with buyers increasingly applying a persistent "exclude generative AI" filter. That filter is the one hopeful detail in the story, and we will come back to it.

The aggregate picture for someone starting today: roughly $50 to $300 a month after six months with fewer than 500 images, rising to $400 to $1,200 a month with 1,000-plus images once established over one to two years, with only a small minority reaching $2,000. Those are ranges from our research, not promises, and the trend inside them is down. Building a 1,000-image portfolio is a year or two of unpaid evenings for a channel whose per-file payout is shrinking while you build.

That is the P&L. The other cost is the one nobody puts in a spreadsheet: every hour spent keywording for an agency's taxonomy is an hour not spent listing the same frame where it sells as a photograph.

Where does a real photograph still sell as a photograph?#

Two channels remain where a buyer pays for a specific image because they want that image, and neither is a stock agency.

The first is digital downloads on Etsy. A photograph sold as a file has no base cost and no print. Etsy's fee stack on a $10 download is a $0.20 listing fee, 6.5% of the sale, and payment processing of 3% plus $0.25, so the seller keeps about $8.60 before any Offsite Ads charge. Compare that to a 15% royalty on a stock subscription download, where the sale price itself is often under a dollar. The catch is that Etsy sends no traffic by default: a listing competes in search, and the metadata has to describe a room and a mood rather than a lens. It is a different craft from stock keywording, and a learnable one.

The second is print-on-demand for the images that suit it, with the caveat that Redbubble's 50% platform fee on Standard accounts leaves about $2 on a $24 shirt. Fewer dollars than a download, but it is a channel a buyer already visits.

There is one more thing your stock portfolio is good for. That "exclude generative AI" filter is a buyer telling you what they want. The frames that were selling to a filtered search are the ones with a human in the light, a place that exists, a moment that did not come from a prompt. Those are exactly the photographs that sell as wall art. The portfolio is a shortlist.

This is also why SpreadPix does not upload to Adobe Stock or Shutterstock, and will not unless paying users prove us wrong. Both would need browser-driven uploads, since neither offers a contributor API for third-party tools, and we could build that. We decided the effort belonged on the marketplaces where the buyer pays for the photograph. Our terms describe what the product does today; we would rather be narrow and honest than list logos.

Worked example: moving one stock image to a download listing#

Take one image from your agency portfolio, ideally one with a real sales history under the AI-exclusion era. This takes about ten minutes and does not touch your stock account.

1. Export the original raw again: sRGB, 6000 pixels on the long edge, JPEG quality around 85 so the file stays under Etsy's per-file cap.
2. Upload it to SpreadPix. The metadata step returns a title, a description, and 13 tags. Read them next to your stock keywords. The stock set says "mountain, sunrise, scenery, nature, travel". The download set says "misty mountain print, calm bedroom wall art, digital download". Same frame, different buyer.
3. Connect your Etsy shop once, then send the image. SpreadPix creates the listing as a draft through Etsy's official API, in the Digital Prints category with the file attached and a line in the description stating it is a real photograph.
4. Publish the draft from your Etsy shop manager. SpreadPix stops at the draft on purpose; the listing going live is your decision.

Then wait a month and put two numbers side by side: what that image earned on stock over the same period, and what it earned as a download. For most images the second number will be zero, because one listing does not create demand and we will not pretend otherwise. But the images where it is not zero are the ones to move next, and the stock account already told you which they are.

Keep the stock portfolio. It costs nothing to leave up and pays a little for no work. Stop building it. The question in the title has a different answer for the hours you have already spent and the hours you have not, and the second answer is the one that matters. If you want to run the ten-minute test on your own catalogue, the signup page currently asks for no card.