For content owners and buyers
What AI companies pay for content: published examples.
A published percentage is not an archive valuation. These dated examples show what particular providers and agreements disclose, with their sources and limits.
Fiund scopes each proposed collection separately. Source fit, authority, usable quantity, preparation and the buyer’s actual needs determine whether there is a proposal to discuss.
Source-backed examples
18 examples across 12 guides. Sources checked 2026-11-05.
| Published example | Scope | Source and date |
|---|---|---|
| 25% of net licensing revenue | LANDR artist pool Allocated among qualifying recordings included in a licensed dataset; not 25% of all LANDR revenue. | LANDR Fair Trade AI program and FAQ Source checked November 5, 2026 |
| Promotion-specific | LANDR recoupable advance Eligibility, amount and participation conditions depend on the accepted offer. | LANDR terms: Fair Trade AI participation and advances Source checked November 5, 2026 |
| 20% average corporate royalty rate | Shutterstock data-license royalty rate The contributor help page describes this rate and allocation proportionate to included content and metadata. | Shutterstock Contributor Fund explanation, September 2026 Source checked November 5, 2026 |
| $60–95 per audio hour | SpeechData.ai catalog buyer price Provider-advertised catalog range, not a Fiund quote or contributor payout. | SpeechData.ai buyer guide, updated October 17, 2026 Guide updated October 17, 2026 |
| $5,000 per participating title | Reported HarperCollins offer The Authors Guild reports a 50/50 author/publisher split. | Authors Guild: HarperCollins AI licensing offer November 2024 offer |
| Approximately $3,000 per eligible work | Anthropic settlement comparison Before applicable fees, costs and allocation; not a licensing price or guaranteed individual payment. | Authors Guild settlement update, October 29, 2026 Authors Guild update October 29, 2026 |
| Financial terms not disclosed | OpenAI–News Corp announcement The announcement states the scope and multi-year nature of the partnership without a price. | OpenAI and News Corp partnership announcement, May 2024 May 22, 2024 |
| $203 million aggregate contract value | Reddit arrangements entered in January 2024 Multiple arrangements over two to three years, as disclosed in the IPO prospectus. | Reddit 2024 IPO prospectus, SEC filing March 2024 prospectus |
| At least $66.4 million expected in 2024 | Historical revenue-recognition expectation A projection in that filing, not a statement of cash received or later actual results. | Reddit 2024 IPO prospectus, SEC filing March 2024 prospectus |
| 25% of net licensing revenue | LANDR pool Subject to qualifying track inclusion and allocation. | LANDR Fair Trade AI program and FAQ Source checked November 5, 2026 |
| 20% average corporate royalty rate | Shutterstock fund Allocated proportionately to included content and metadata. | Shutterstock Contributor Fund explanation, September 2026 Source checked November 5, 2026 |
| Promotion-specific | LANDR recoupable advance The accepted offer and terms control; no fixed current per-track advance is assumed. | LANDR terms: Fair Trade AI participation and advances Source checked November 5, 2026 |
| Limited-period and perpetual arrangements | LANDR partner-dataset terms The program distinguishes refresh and payment mechanics for these two arrangements. | LANDR Fair Trade AI program and FAQ Source checked November 5, 2026 |
| Non-exclusive buyer licenses | Fiund contributor agreement The agreement distinguishes Fiund’s exclusive representation from the non-exclusive licenses granted to buyers. | Fiund contributor agreement Source checked November 5, 2026 |
| Qualifying rights required | LANDR participation condition The program describes eligibility conditions for master and publishing rights. | LANDR Fair Trade AI program and FAQ Source checked November 5, 2026 |
| Included content and metadata | Shutterstock allocation basis Payment allocation concerns material included in purchased datasets. | Shutterstock Contributor Fund explanation, September 2026 Source checked November 5, 2026 |
| Allocation depends on included recordings | LANDR pro-rata example The share of the pool depends on recordings included relative to the dataset; it is not a fixed payment per track. | LANDR Fair Trade AI program and FAQ Source checked November 5, 2026 |
| 10–15 business days, as an example | Suggested acceptance window A buyer-guide example to negotiate, not a guaranteed duration for a deal. | SpeechData.ai buyer guide, updated October 17, 2026 Guide updated October 17, 2026 |
By topic
Payment and license structure
Video, photos and motion
Writing and text
Category preparation notes
These categories have no verified public payout benchmark in this guide. The notes describe what to inventory and review before a source assessment.
- Podcast archives: what can be assessed before pricing
- Interview archives: source records before a quote
- Call recordings: permissions and privacy before pricing
- Voice recordings: permitted uses and identity restrictions
- Stems and samples: preserve the session and its rights
- Sound-effect libraries: describe source and permitted use
- Video archives: usable originals, context and permissions
- Photo libraries: image-specific permissions before pricing
- Drone footage: coverage, telemetry and reuse authority
- Dashcam footage: define a lawful, useful collection
- Gameplay and simulation data: actions, outcomes and rights
Use the right comparison
Some examples describe a whole corporate contract, some a provider’s selling price and others a contributor allocation. A settlement award is a separate legal comparison. None can be divided into an assumed payment for your archive.
See publicly reported licensing agreements or read Fiund’s payout mechanics.
Frequently asked questions
Are these Fiund rates?
No. Each example describes its named provider, agreement or disclosure. It does not establish a Fiund quote or a guaranteed payment to an archive owner.
How should I compare the figures?
Keep the unit, date, license scope and payment basis attached. A corporate contract total, a buyer’s dataset price and a contributor’s share are different measures.
Why do some categories have no payout figure?
This guide has not established a verified public payout for every category. Those research notes explain preparation questions without inventing a rate.