Revealed: What publishers can expect AIs to pay for words, pics and video
The market for premium content is headed to $30 billion, but publishers must navigate the tricky balance between dollars and sense...
Welcome from magnificent London, which is grey, cold, historic, and fabulous. I had beers last night with friends at a 16th Century pub frequented by Dickens.
Welcome to new subs from Yahoo, Apple News, Bauer Media, The Conversation, Britain’s MyLocal and the Yarrawonga Chronicle, among others. Hey there :)
I just presented to the Professional Publishers Association on the risks of Google SGE, leading to new subs from EssenceMediacom and Dentsu, strategists Stratechmedia, Pushly, Sloop Media, and more.
Great to have you all along. Let’s kick off with a glance ahead to reporting season before we get to the big stuff...
Major media and Big Tech will all be opening the kimono on their latest numbers and performance over the coming days.
Think of it as your old school report, only with trillions of dollars, and the future of world order at stake.
Microsoft, Meta, Alphabet, Amazon, and Apple will all step up to be judged by the markets. So will bellwether media stocks like News Corp. It’s popcorn season 🍿
Meta was first to report, and revealed continued and astonishing revenue growth, up 27 per cent to $36.5 billion for the quarter, but shares crashed 16 per cent as investors punished Zuck for big spending on AI.
Google reports in the next few hours. It’s expected to show some recovery after recent shellackings, and YouTube’s smashing it after a $9.2 billion last quarter, but the market remains twitchy on its AI vision.
Microsoft also reports shortly. Wall Street sees it as AI’s leader, and is expecting Azure cloud and Microsoft 365 growth. Microsoft already trades on a hefty 37x revenue.
Apple’s still having a tough time, with iPhone sales in its pivotal China market falling 19 per cent in the March quarter, its worst performance since 2020. It’s also facing antitrust, leading to headlines like this: Say Goodbye To Your iPhone Maybe If Antitrust Regulators Have Their Way. Oof.
Many analysts are predicting that Amazon will perform best this time around. Its AI bet Perplexity, just grew to unicorn status after a funding round.
We’ll find out soon but note that all this investor optimism comes against a backdrop of tightening privacy, antitrust action, escalating fines, and regulation.
First reports of a tech bubble bursting are emerging, with Reuters quoting UBS predicting a 15.5 per cent collapse within a year. Ouch.
Never mind, you can always rely on Bitcoin. No, wait, hang on…
Right, let’s get to the main event.
It’s still numbers, but these are the stats that every publisher and content owner needs to know.
Kudos to Reuters who pulled these together. I’m keynoting at Reuters’ Canary Wharf HQ here in London in a few hours.
Its reporters collated what Big Tech and independent content brokers are paying publishers for text, images, and video, to train AI.
Before we dive in, remember you need to know your RAG from your trainers.
Also, go into it with your eyes open.
Big Tech wants to buy vast archives of text, images and video, to train their AIs to usher in the GenAI future.
They want this future because AIs will then use your work to create content for free and at unlimited scale where ads can be placed on a near 100 per cent margin, from which you earn nothing.
Many have dived in and done deals, including AP. I’m seeing them on Friday.
If tech is pushing you there, ask yourself why.
My recommendation is to focus on RAG. This is where AI is going to replace search. There is more detail in here.
The quick version is this: If you ask Gemini or Microsoft Copilot a question about anything less than two years old, it retrieves the information from the open web.
It then generates and delivers an AI response, that feels conversational. Retrieval And Generation, R.A.G. Hence, RAG.
That’s the future of search, and that’s where the real war for value is being fought. Microsoft and Google are battling over what has become a $2 trillion market sector that AI is disrupting.
Remember, Big Tech wants all the money. It’s the landgrab I’ve been writing about.
So, before you decide what to sell, and to whom, for what purpose, and for how much, do your homework and decide if you want a cheque now and sell your future, or sell for RAG to build an ongoing revenue stream.
But I know you’re hungry for the numbers, so these are the ones the Reuters team dug up.
Please read the full article as this will have taken the journalists a lot of work, and I don’t want to steal their thunder, but this is what you need to know.
Analysts estimate that the new licensing market created by AI is worth $2.5 billion already and headed to $30 billion.
If you make content, and/or own it, this is a new revenue opportunity that’s potentially larger than your digital ad revenue today.
The tech companies are demanding scale. They might not take your call unless you have billions of words, tens of millions of images, or millions of videos.
Photobucket traded 13 million photos and images at rates of 5c and $1 per photo, and more than $1 per video.
Shutterstock has done deals worth $25 million to $50 million with Meta, Google, Amazon, and Apple, for hundreds of millions of images, videos, and music files.
Freepik has allowed two unnamed tech firms to license 200 million images at 2c to 4c per image.
Content brokers are also beginning to emerge. They typically charge 30 per cent to cut deals.
Defined.ai says it licenses to Google, Meta, Amazon, and Microsoft, at $1-2 per image, $2-$4 per short form video, and $100-300 per hour for longer films, with text was $0.001 per word.
Not all images are valued equally. Photos of nudity and violence are trading at premiums of $5 to $7 as they can train AIs to avoid NSFW content.
Crime scenes, conflict violence and surgery attract the highest fees, and war correspondents in Gaza are being approached directly for their footage.
There you are.
Those are the numbers. It’s a murky market right now, where the value of what’s being traded is largely unknown, in a legal vacuum, and with a negotiating imbalance.
The question I’m being asked most at the events I am hosting is how much will I get? Now you know.
The question I should be being asked is, for what? I will organise an open chat ewhen I am back so we can do a Q&A on this.
In a surprise development, the Zuck and I agreed on something this week.
“The thing that I think is going to be more valuable is the feedback loops rather than any kind of upfront corpus,” Zuckerberg said in an interview with newsletter Command Line.
Feedback loops, or human reinforcement learning as geeks love to call it, let AIs know when they make a mistake, so they can learn and improve.
Zuck said: “Having a lot of people use it and then seeing how people use it and being able to improve from there is actually going to be a more differentiating thing over time.”
We learned this at my video AI company Oovvuu. We initially trusted tech to embed videos that were contextually relevant to articles.
What we learned was that voice to text and natural language processing (NLP) tech, which were early AI tools that were talked up at the time, were not up to the job.
They were only 80 per cent accurate, which was probably OK for the wonky good-enough-to-sell-an-ad on our wild west web approach of Google, but it failed for premium publishers who set a far higher bar.
Instead, Oovvuu pivoted to use AI tech to recommend the video to a professional journalist, and then used the journalists’ decisions to teach the AI to make smarter recommendations.
We called it Loop. Zuck’s saying the same. He has three billion people on his platforms to train his AI. It’s his differentiator.
But…
Publishers have more than five billion readers collectively. We are the bigger elephant. That means we can do it better if we work together.
I’ll return to this theme and reveal a $4 billion opportunity for publishers shortly.
Phew. That’s a lot of numbers. Sorry if it has your head spinning. Here’s one last number: 38. That’s how many minutes I was able to stay awake at the Royal Albert Hall last night before jetlag got me… See you next time.