Did Stock Photography Really Collapse 98%? I Checked the Numbers

A viral article has been doing the rounds claiming that stock photography has collapsed by 98%, from $1.47 billion down to $31 million. Buries the whole industry in one headline. When I first read it my honest reaction was: there’s no way that’s true. So I did a little digging, and I want to walk you through what I found, because the truth is more interesting than the headline. The numbers don’t check out, but the anxiety behind them is real, and it’s worth talking about both halves honestly.

The short version: The 98% collapse claim doesn’t survive basic scrutiny; going from $1.47 billion to $31 million is not a decline, it’s an extinction, and the industry visibly still exists. What IS true: per-download earnings have fallen since AI images arrived, and nobody should deny that. But AI tooling has also made uploading dramatically faster, which offsets a lot of the loss. Real contributor data backs this up.

The claim

The article says the stock photography market fell from $1.47 billion to $31 million. That’s a 98% collapse. As soon as you start digging into those numbers, it doesn’t look to be entirely accurate. Think about what a 98% collapse would actually mean: every major agency insolvent, every full-time contributor gone, the entire supply chain dead. That is not the industry any of us are looking at. Agencies are still paying out, contributors are still uploading, buyers are still buying. Even if you’re pessimistic about stock, going from 1.47 billion to 31 million is a crazy claim, and I don’t think it’s true.

What’s genuinely true in it

Here’s where I want to be fair to the doomers: there has definitely been a fall-off in earnings. No one can deny that. Ever since the rise of AI images, per-asset and per-download income has come down, and pretending otherwise helps nobody.

But there’s a counterargument that the collapse articles never mention: you can now upload assets with such ease and speed that it massively counteracts the losses. The bottleneck in stock was never shooting; it was the fiddly keywording, titling and submitting. AI has all but removed that bottleneck. Lower earnings per asset, far more assets per hour of work. That’s the actual trade happening in this industry, and it’s a very different story from “98% collapse”.

What real contributor data shows

Around the same time as the collapse article, a stock photographer opened his books publicly: every agency, every sale since 2008, roughly $500,000 in total commissions. No pitch, no course to sell, just data. A few things jumped out at me:

• The peak was 2018, and it’s been trending down since. That matches the earnings decline everyone feels.
• Shutterstock used to generate around $12,000 a year for him, and the chart shows it going down, down, down. Yet it’s still his biggest earner.
• Adobe Stock is the one killing it. That matches my own books too; Adobe is where I make the most.
• He built that on about 25,000 assets over 18 years. Uploaded manually, that’s a monumental amount of work. But here’s the encouraging bit: I’ve got 19,000 assets and I’ve only been doing this for about two years, because AI handles the fiddly metadata side for me. What took him 18 years of manual grind is now achievable in a fraction of the time.

So the honest reading of the data: commissions per download are not what they were, and you probably won’t see 2018-era rates again. But the volume you can realistically produce has exploded, and that’s the redeeming factor of the stock photo and video industry right now.

My take

Should we be concerned? About lazy headlines, yes. About the industry, cautiously no. The market is changing shape, not disappearing. The people who will struggle are the ones running a 2015 workflow: shoot, manually keyword, upload a handful of clips a week. The people doing fine are treating stock like a volume-and-systems game. I’ve written up my own real numbers, month by month, in my road to £10k with stock footage if you want to see what the income actually looks like from the inside. And the tool I built to do the cutting, tagging and metadata locally on my own Mac is Clip Up Pup; it’s the reason 19,000 assets in two years was possible.

What do you think? Is the article right and I’m in denial, or is it full of bogus numbers? I read every reply, and stories like this go out every week in the newsletter: join the Dispatch.

Frequently asked questions

Did stock photography really fall 98%?

Almost certainly not. The viral claim of $1.47 billion falling to $31 million doesn’t survive scrutiny; a decline that size would mean the industry had ceased to exist, and it visibly hasn’t. Agencies are still paying contributors every month.

Are stock earnings really declining?

Per-download earnings, yes. Public contributor data and my own experience both show a peak around 2018 with a steady decline since, accelerated by AI imagery. Total income is a different story, because AI tooling lets you upload far more assets in far less time.

Which stock agency pays best right now?

In both the public dataset discussed here and my own books, Adobe Stock is currently the strongest performer, with Shutterstock still significant but declining year on year.

Is stock photography still worth starting in 2026?

I think so, if you treat it as a volume-and-systems game rather than a handful of manually uploaded clips. I built a five-figure portfolio in about two years working that way, and I’ve documented the real numbers on this site.