5 Takeaways from the AI Content Licensing Deals I'm Tracking (size of avg. deal , market, etc.)

Here's what I've learned from analyzing the first 34 AI Content Licensing Deals

Hi and welcome!

I created Media & the Machine to help content owners grow leads with AI.

If you own content or build AI, I think you’ll love it.

10-Second Summary of this Post

• AI companies are paying an average of $24M per publisher for content.

• News Corp’s $250M deal = 2.5X their 5-year net income

• Big 4 AI buyers: OpenAI (53%), Google (12%), Microsoft (9%), Meta (6%)

• Market size: $816M per year → $2.92B in total commitments

• More deals coming — Smaller publishers are next

In this first post, I share key insights from a spreadsheet I made on the first wave of AI content licensing deals (it’s up to 34 right now).

News Corp’s $250M OpenAI deal = 2.5X their 5-year net income.

I started the spread May 22, 2024 when I read this article in the Wall Street Journal: “OpenAI and News Corp Strike Content Deal Valued at Over $250 Million”.

That got my attention — $250 mil. (over 5 years) is big…even for a giant like News Corp.

Captain Obvious Note: Content licensing is almost pure profit — I should know. In 1994, I closed some of the first web content licensing deals at CMP Media. My partner Delyn Simons & I built a $2M-a-year biz with 90% profit—our only cost was $200K salary (a top 3 most profitable biz out of 20! Gravy Baby!).

Back to News Corp. Consider this — they made only $99 mil in total net income over the last 5 years. Total!

With a guaranteed $250 mil. from OpenAI, News Corp. 2.5X’s their income in one move.

I know why News Corp. CEO Robert Thomson took the deal—Gravy, Baby!

My 5 Takeaways on the First 34 AI Content Licensing Deals

Disclaimer — I’m not Gartner Group. I just track LLM/content deals using press releases, SEC filings, earnings transcripts and analyst estimates.

OK, on to my first 5 takeaways from the first 34 deals:

1) AI Companies are Paying an Average of $24 Mil. per yr to the First Wave of Content Publishers

AI companies are paying publishers an average of $24 mil. per year.

The $24 mil. average is a total estimate — I’ve based it on the 16 deals (see above) whose revenue estimates I got from places like:

• Earnings Call transcripts

• Quotes from a CFO

• Anonymous Sources in reputable publications like The Wall Street Journal and Reuters

“Fixed” Upfront Versus “Variable” Backend Licensing Fees

For now, my revenue estimates do not break out the two components of some AI Content Licensing deals:

1. a “Fixed” upfront payment

2. a variable backend payment.

The sharp Adweek Senior Reporter Mark Stenberg brought up the fixed/variable scenario when he broke the news that OpenAI was paying IAC’s Dotdash Meredith a “minimum” of $16 mil. per year from OpenAI — he noticed from an earnings call that IAC’s COO/CFO Chris Halpin mentioned that there would also be “variable components”.

My spreadsheet includes one licensing fee number per year (which is either fixed (in the case of Dotdash Meredith) or may be total deal size in other cases).

A $24 Mil. Average Deal Size Might be Artificially High (or maybe not)

You could argue that my $24 mil. deal size estimate is artificially high.

OpenAI & all may be paying top dollar only in these early days to the publishing giants with huge troves of data.

You may be right.

We need to watch out for the next wave of deals.

And there will be plenty more.

News Corp. and Hearst are the only publishers in the first 34 deals I tracked that rank in the top 20 largest media companies in the world.

Smaller Content Publishers (also in the $millions?)

I don’t have any good #s on AI Content Licensing deal size for smaller publishers.

But, the CFO of a $50 mil./yr. content company recently shared that they expect $20 mil. in AI content licensing revenue in 2025. It’s likely that will be from multiple LLMs.

Perhaps that small content co. is talking about 5 LLMs they license to. That’d still be $4 mil average deal size.

I’ll share more on that story in another post.

2) The Top LLMs Paying for Content

The LLMs that are most active in content licensing are below:

The Big 4 LLMs — The first 34 deals have been dominated by:

1. OpenAI has dominated (52.9% of deals) — This makes sense. They have the largest treasure chest of the pure-play LLMs. And they are open about announcing new things.

2. Google/Gemini (11.8%). — Google coming in at #2 makes sense and might underrepresent their marketshare because they are more guarded about announcing things. Note: I’m calling this group “Google/Gemini” because Google still seems to be rejiggering its corporate structure (e.g. Gemini is part of Google DeepMind which reports directly to the Alphabet parent; and Google.com is separate and also reports to Alphabet. There are likely content licensing agreements that are for just one or the other. But I expect Alphabet to clean this up — if not, I might separate out Gemini and Google.com in the future.

3. Microsoft (8.8%) — This is also not a surprise. They’ve taken a backseat to their strategic partner OpenAI and have access to their data. But I expect Microsoft to announce more and more of their own deals going forward as the 2 frenemies compete more with each other.

4. Meta (5.9%) — Meta is a bit different. Their LLAMA LLM is primarily trained on publicly available data sources. However, they’re spending a bunch of time in court these days about claims they used pirated content (such as LibGen). Mark Zuckerberg was deposed.

Note: Three of the first 34 deals were done by a “large tech company” (likely OpenAI, Google, Microsoft, Meta or Apple).

3) The Most Active Publishers Licensing Content to AI Companies

This is a breakdown of the most active content companies licensing their data to AI co.s (out of the 34 deals I’m tracking):

Some early learnings on these content licensing deals:

• Specialized & Niche Audiences are Valued— Examples: Future (200+ niche brands like PC Gamer), Stack Overflow (developer-focused content), IAC/DotDash Meredith (lifestyle and wellness brands).

• Image Content is Popular — AI co.s are paying for images. Shutterstock is by far the most active (with 6 deals) and Freepik is tied for 2nd. It’s no surprise Getty Images paid $3.7 bil. for Shutterstock.

• Non-Exclusive — Shutterstock, Reddit, Axel Springer and Associated Press all did deals with more than one LLM. Axel Springer (licensed to OpenAI and Microsoft), The Financial Times (OpenAI) and The Atlantic (OpenAI) have separately agreed to let ProRata license their content to other LLMs in a 50/50 revenue share. I’ll publish a list of ProRata deals in a separate post.

• International Content in Demand — Even though top AI LLMs are U.S.-based, they are striking deals overseas (i.e. with Axel Springer (European-centric), Prisa Media (Spanish-language media), GEDI (Italian newspapers) and Agence France-Presse (international news)).

4) The Total Market Size for AI Content Licensing: $816 Million per Year

I haven’t seen Gartner, IDC, Forrester or publish the size of the AI content licensing market.

So I’m gonna take a crack at it.

Based on the 35 deals I’ve tracked (below), I estimate that AI companies spent $816.7 Mil.. on content licensing in 2024.

Here’s my basic math:

• The average deal size I’m seeing is $24 mil. ($24,020,833 to be exact) (see the above 16 deals I tracked)

• There are 34 deals that have been reported on

• 34 * $24 million= $816,708,333

This is far from perfect.

• The $24 mil. is probably a bit on the high side

• But, the 34 deals I list is clearly on the low side:

  • some publishers likely have not announced their deals (they might want to be low-key about this until they figure this AI thing out).

  • some AI co.s might want to keep their deals private as they don’t want to tip their hand to rivals.

5) AI Companies Have Committed $2.92 Billion to Content Publishers

I estimate that LLMs Have Committed $2,92 Billion in Content Licensing Fees to Publishers (as of Jan. 28, 2025).

These “Commitments” of $2.92 bil. are larger than the$816 mil. because most deals are in the 1 to 6 year range.

Here’s my math:

• $24 mil. Average Deal Size — The average of the 16 deals I’ve got decent rev #s on is $816,708,333.

• 3.5 Year Average Deal Term - I’ve only got 6 deals with an estimated term length — they average 3.5. There are another 3 deals that are “multi-year” which it seems safe to assume are in the 2 to 5 years (again, a 3.5 year average).

• Math = 34 Deals *$24,020,833 deal size *3.5 term = $2.92 bil.

Final Takeaways

For Content Owners:

If you own valuable content, this is your moment. AI companies are spending billions to license data, and deals are getting more frequent. Even smaller publishers are securing multi-million dollar agreements.

If you’re not tracking AI content licensing, you might be leaving money on the table.

For AI Executives:

The content licensing market is heating up fast. OpenAI, Google, Microsoft, and Meta have committed nearly $3B so far—and that’s just the first wave.

If you want access to premium, proprietary content, now’s the time to lock in deals before prices go up or regulatory hurdles get tighter. Expect competition for high-value publishers to increase.

What I’ll share next:

I’m working on a few things for my next Media & the Machine posts:

• What the other large content co.s are doing — i.e. the ones who have not announced content licensing deals (Elsevier, New York Times, Gannett, Pearson, Fox, CNN, Warner Brothers, Disney etc.).

• What smaller content publishers are doing to monetize their content via AI. My table of 34 deal above are mostly media giants ($500 mil. to $10 bil. in $rev). Some smaller content players are generating substantial AI licensing revenue too.

• The Other LLMs — I’ll cover what DeepSeek, Perplexity, Anthropic and other AI companies are doing with content publishers.

• More on Deal Terms — Content Licensing deals have fixed and variable payments. LLMs pay different fees based on text versus images versus video.

…there’s a lot to cover — I’m psyched to do it.

Thanks for reading!

Rob Kelly, Creator & Host of Media & the Machine

p.s.: How to reach me:

If you want to reach me, the best way is to subscribe below and reply to any of my emails. There’s a free option — and your reply comes directly to my inbox.