Every Perplexity figure here is a dated claim of the party that made it. Publisher revenue-share terms could not be read on Perplexity's own pages when checked on 2026-11-09 and 2026-11-16, so none is stated.
This page is part of an independent reference on AI training data licensing that covers every buyer by the same method.
Each number carries its source and check date, and how DataPacts sources and dates every figure is set out in the methodology.
How Perplexity Acquires Content: Answers, Not Model Training
Perplexity acquires publisher content in two documented ways, by crawling the web for its search index and by licensing content from named publishers and platforms to display in answers, and it states that its search crawler does not collect content to train foundation models.
Its crawler documentation, re-read on 2026-11-16, names PerplexityBot, which surfaces sites in search, and Perplexity-User, which fetches pages when a user asks. The licenses on record come from a newspaper, an image library, a newspaper group and an encyclopedia API.
Perplexity states that PerplexityBot "is not used to crawl content for AI foundation models". No statement on foundation-model training data and no dataset disclosure are on record, so this page does not say whether or how Perplexity trains models.
For a licensor the route matters: content acquired for retrieval and display is a different grant from a training license, and the difference between licensed vs scraped training data decides what a grant allows.
Perplexity Company Identity
Perplexity's legal entity is Perplexity AI, Inc., a United States company named as the defendant in each of the US suits on record and operating at perplexity.ai.
The table lists the identity attributes on record, as of 2026-11-16.
| Attribute | Value | Source |
|---|---|---|
| Legal name | Perplexity AI, Inc. | Complaint, S.D.N.Y. 1:24-cv-07984-KPF, Doc. 1 |
| Brand | Perplexity | Perplexity documentation |
| Website | perplexity.ai | perplexity.ai |
| Country | United States | NYT Form 10-Q, period ended 2026-09-30 |
| Type | AI answer engine company | Perplexity documentation |
| Registry number | not verified | n/a |
| Founding year, founders, executives, address, listing status | not stated | n/a |
The sources checked here do not give a registry number, founding year, founders, executives, headquarters address or listing status, so none is written.
Perplexity Publisher Program and Comet Plus: What the Record Holds, Dated
Perplexity's publisher revenue-share terms are not verified: its own program pages could not be read when checked on 2026-11-09 and 2026-11-16, so the table lists each term a publisher would check and what the record holds for it.
The table compares each publisher-program term with what Perplexity's own pages show, as of 2026-11-16.
| Term | Record | Checked | Source |
|---|---|---|---|
| Publisher revenue share (Comet Plus) | announced on Perplexity's blog; the page returned HTTP 403; terms not verified | 2026-11-09, 2026-11-16 | Perplexity blog post |
| Percentage, pool and subscription price | not verified | 2026-11-16 | no readable primary page |
| Publisher Program advertising share | not verified; the post and the publishers page returned HTTP 403 | 2026-11-16 | perplexity.ai/publishers |
| Eligibility and application route | not stated | 2026-11-16 | no readable primary page |
| Reporting, payment timing and minimums | not stated | 2026-11-16 | no readable primary page |
| Company data program | none on record | 2026-11-16 | perplexity.ai |
| Data-holder payout and company referral fee | none on record | 2026-11-16 | no program page |
| Crawler statement | PerplexityBot "is not used to crawl content for AI foundation models" | 2026-11-16 | Perplexity bot documentation |
Perplexity announced its publisher revenue share on its own blog, in the Comet Plus post, and that page returned HTTP 403 on 2026-11-09 and again on 2026-11-16. The Publisher Program post and perplexity.ai/publishers also returned HTTP 403 on 2026-11-16.
Figures for the share, a pool and a price circulate in press coverage only. A buyer's terms are written here only from the buyer's own page, so none of those figures appears.
This row reads "not verified" beside the other programs in the program terms compared across buyers.
Perplexity's Licenses on Record: Four Agreements by Date
Four licenses with Perplexity as licensee are on record, each announced by the counterparty, and none states a value.
The table lists each license with the date, content and use stated by the counterparty, as of 2026-11-16.
| Date | Counterparty | Content | Use stated | Value | Source |
|---|---|---|---|---|---|
| 2025-05-14 | Le Monde | newspaper content | answers only, training excluded | not disclosed | Le Monde |
| 2025-07-31 (disclosure date) | Gannett, now USA TODAY Co., Inc. | news content | not stated | not disclosed | Q2 2025 earnings release, 8-K Ex. 99.1 |
| 2025-10-31 | Getty Images Holdings, Inc. | images | display in AI search and discovery tools; training not stated | not disclosed | Getty Images newsroom |
| 2026-01-15 | Wikimedia, LLC (Wikimedia Enterprise) | Wikipedia and Wikimedia content | paid API access; scope not stated | not disclosed | Wikimedia Enterprise |
Le Monde announced a multi-year agreement on 2025-05-14 and writes that it "ne prévoit pas l'utilisation de nos contenus pour entraîner un LLM", meaning it does not provide for using its content to train a language model. It applies only to answering questions.
Getty Images announced on 2025-10-31 a global multi-year license for display of its content across Perplexity's AI search and discovery tools, and says nothing on training.
Gannett's chief executive called the deal "fair value of our content" in the earnings release of 2025-07-31, which gives neither an amount nor a signing date. Wikimedia Enterprise named Perplexity among its new paid API partners on 2026-01-15 and discloses no pricing.
No party states a value for any of the four. The deal tracker of AI data licensing deals dates each one with the party that stated it.
Perplexity Payouts: What a Publisher Can and Cannot Learn
Perplexity states no payout that can be verified on its own pages as of 2026-11-16, so this profile gives no percentage, pool or per-visit amount for publishers.
No data-holder payout or company referral fee is on record either, and the licenses above carry no stated value.
To evaluate a usage-based revenue share, a publisher needs the base the share is calculated on, the reporting it receives, any minimum, the term and the termination rights. Perplexity's record answers none of these in a primary source that can be read.
Because revenue depends on usage, a publisher cannot know it in advance. How a revenue share compares with a license fee or a sale is covered on the monetization models page.
Perplexity Reputation: Crawling Dispute and Lawsuits on Record
Perplexity's record holds a public crawling dispute with Cloudflare and six US lawsuits by publishers and platforms, all pending as of the dates in the table.
The table lists each suit with court, docket, filing date, claim type and latest docket entry in the knowledge held as of 2026-11-09.
| Case | Court and docket | Filed | Claim type | Status (as of 2026-11-09) | Source |
|---|---|---|---|---|---|
| Dow Jones & Company and the New York Post publisher v. Perplexity AI | S.D.N.Y. 1:24-cv-07984-KPF | 2024-10-21 | copyright and trademark | discovery; motion to compel granted in full 2026-10-29 (Doc. 227) | Doc. 227 |
| Encyclopaedia Britannica and Merriam-Webster v. Perplexity AI | S.D.N.Y. 1:25-cv-07546-JLR-SLC | 2025-09-10 | copyright and trademark | discovery; Opinion and Order 2026-10-27 (Doc. 152) | Doc. 152 |
| Reddit v. SerpApi, Oxylabs, AWMProxy and Perplexity AI | S.D.N.Y. 1:25-cv-08736-PAE | 2025-10-22 | DMCA section 1201 circumvention and related claims | motions to dismiss of SerpApi and Perplexity predominantly denied 2026-07-31 (Dkt. 104) | Complaint |
| Chicago Tribune v. Perplexity AI | S.D.N.Y. 1:25-cv-10094-LAP | 2025-12-04 | copyright | pending; coordinated with the New York Times case (source code protocol 2026-10-21) | Complaint |
| The New York Times Co. v. Perplexity AI | S.D.N.Y. 1:25-cv-10106-LAP | 2025-12-05 | copyright and trademark | renewed partial motion to dismiss filed 2026-04-17; stipulated source code protocol submitted 2026-10-21 (Dkt. 81) | Dkt. 81 |
| Cable News Network v. Perplexity AI | S.D.N.Y. 1:26-cv-04427-LAP | 2026-05-28 | copyright | motion to dismiss (ECF 30) fully briefed; oral argument requested 2026-10-23 (Dkt. 42) | Dkt. 42 |
Cloudflare's report of 2025-08-04 de-listed Perplexity as a verified bot. Cloudflare measured the declared crawler (user agent Perplexity-User/1.0) at 20 to 25 million requests a day and undeclared crawling at 3 to 6 million a day, and it reports a crawl-to-refer ratio of 118:1 for 2025-08-01 to 2025-08-07. These are Cloudflare's measurements of its own network, not findings of a court, and its post does not say whether PerplexityBot is included.
The suits concern copying for answers or circumventing access protections, not a breach of a license, and no court has ruled on infringement or fair use. Claims and rulings in Dow Jones and NY Post v. Perplexity are set out on its own page, which tables every Perplexity suit.
Is Perplexity a Reliable Licensee for Publishers?
Perplexity is an identifiable licensee with four counterparties on record, but a publisher cannot check its revenue-share terms on a primary page, and it is a defendant in six pending US suits.
Documented are a named legal entity, four counterparties and public crawler documentation. Missing are readable program terms, any stated value and any audit or reporting term.
The litigation shows that publishers that did not license sued. The Dow Jones complaint describes a licensing offer of July 2024 that received no response, and the case page records that account.
How other publishers structured their deals is covered where news publishers licensing content to AI companies are compared, with each deal's stated use.
Benefits, Risks, Downsides and Alternatives to Licensing Content to Perplexity
For a publisher, licensing to Perplexity can bring display revenue under a license that excludes training, against terms that cannot be read on a primary page and a crawler record that is disputed; the alternatives include other publisher programs and not licensing at all.
Benefits
- A display-only license is possible: Le Monde's agreement excludes training (Le Monde, 2025-05-14).
- Counterparties of different types are on record: a newspaper, an image library, a newspaper group and an encyclopedia API.
- Perplexity states that PerplexityBot does not crawl for foundation models (checked 2026-11-16).
Risks
- Revenue-share terms are not readable on a primary page.
- Perplexity says Perplexity-User "generally ignores robots.txt rules".
- Cloudflare reports undeclared crawling, which Perplexity's documentation does not address.
- Six US suits are pending.
Downsides
- No stated value, rate or minimum exists for any license.
- No company data program is on record, so holders of non-content data have no route.
- Revenue that depends on usage is unknown in advance.
Alternatives (alphabetical, each with its stating party, checked 2026-11-09)
- Cloudflare Pay per crawl, in closed beta: the site owner sets a price per zone, and Cloudflare is Merchant of Record (Cloudflare).
- Google News AI pilot: Google says it pays participating news publications, with amounts not published (Google).
- Microsoft Publisher Content Marketplace: Microsoft states publishers "will be paid on delivered value" (Microsoft).
- ProRata, an intermediary: "Payment depends on the relevant rights, licenses, and commercial agreements" (ProRata).
- RSL Collective, a licensing collective: royalties "whenever an AI application uses your content to generate a result", with rates not published (RSL Collective).
- TollBit, an intermediary: publishers set their own rates (TollBit).
- Not licensing at all, and blocking crawlers that honor robots.txt.
The wider question of should you sell or license your data at all is covered on its own page.
Perplexity Among Frontier Labs That Acquire Data
Perplexity is one of the frontier labs that acquire content for AI, and its licenses on record cover answers, with no training use stated by Perplexity. It is compared with other frontier AI labs as data buyers on the parent page.
- Who buys AI training data: every buyer and program, compared
- OpenAI, Google, Microsoft and Amazon: other labs with content licenses
- Cloudflare pay per crawl and content licensing platforms for publishers: priced alternatives
- AI crawlers and pay-per-crawl and crawler licensing: how to allow, block or charge crawlers
- Getty Images AI licensing deals and news archives as AI training data: the image and news deals
- AI copyright lawsuits: the case tracker
- Choosing a buyer for your data and alternatives to selling data: the seller side
More About Perplexity as a Data Buyer
Where Perplexity's Crawlers, Cases and Deals Are Covered
Perplexity's crawlers, the suits against it and each of its licenses have their own pages, so this profile summarizes them in one line each.
- PerplexityBot and Perplexity-User: documentation and the stealth-crawling dispute
- Reddit v. Perplexity, SerpApi and others: the scraping suit
- Britannica and Merriam-Webster v. Perplexity: the encyclopedia suit
- News publishers v. Perplexity (New York Times, Chicago Tribune, CNN): the news suits
- Le Monde AI licensing deals, USA TODAY Co. (Gannett) AI licensing deals and Wikimedia Enterprise AI deals: the deal pages
- Data buyer incidents: the dated log, which lists the Reddit suit
Frequently Asked Questions About Perplexity and Publisher Licensing
Does Perplexity pay publishers?
Perplexity has announced a publisher revenue share, but its program pages could not be read when checked on 2026-11-09 and 2026-11-16, so no verified percentage, pool or amount exists on this page. No license on record states a value.
Does Perplexity train AI models on publisher content?
Perplexity states that PerplexityBot is not used to crawl content for AI foundation models, and Le Monde says its 2025 agreement excludes training. No broader training statement is on record.
Can a company license its data to Perplexity?
No program for company data is on record: Perplexity's documented licenses cover publisher, image and encyclopedia content for answers, checked 2026-11-16.
Who runs Perplexity?
Perplexity is operated by Perplexity AI, Inc., a United States company; the sources checked here do not record its founders, executives or headquarters address. The entity is the defendant in each US suit tabled above.
This page is information, not legal advice.