ICODA tested 36 crypto queries across 4 AI engines to find which media they actually cite. This is the ranking, why the top outlets win, and what to do about it.

Crypto.news, CryptoSlate and CoinDesk are the crypto media outlets AI engines cite most, and only crypto.news and CoinDesk were cited by all four engines.

Anyone planning crypto PR wants to know which publications AI pulls from when a user asks about crypto. ICODA ran 36 queries on ChatGPT, Perplexity, Claude and Gemini with live web search, collected 144 answers and 1,137 citations, and counted every crypto media outlet that appeared. The list of outlets that matter is short enough to build a generative engine optimization plan around.

Which Crypto Media Outlets Do AI Engines Cite Most?

Crypto.news ranks first with a 13.9% citation score, followed by CryptoSlate (12.5%) and CoinDesk (11.8%) across 36 queries and four AI engines.

An outlet’s score is the share of the 36 queries where it was cited, averaged across the four engines, so Perplexity (about 23 sources per answer) does not drown out ChatGPT (about 3). In total, 49 crypto media outlets appeared among 355 cited domains and took 24% of all citations.

Engine columns show how many of the 36 queries cited the outlet. Rank range is the 95% interval from resampling the 36 queries 2,000 times. Ranks 10 to 12 tie on score.

Read the rank ranges before the ranks. The top three have the best scores, but their ranges overlap with each other and with ranks 4 to 8. In practice, five or six outlets form the top tier, with a long tail behind them. The top five hold 44% of all crypto media citations.

Decrypt, Blockworks and BeInCrypto were cited in none of the 144 answers. BeInCrypto is a large outlet: Ahrefs counts about 5,300 AI citations for it across six platforms. ICODA’s 36 questions did not pull it in, so the ranking reflects this query mix and says little about the wider market.

Why Do ChatGPT, Perplexity, Claude and Gemini Cite Different Crypto Media?

Each engine reads a different slice of the web: crypto media is 34% of Claude’s citations, 24% of Perplexity’s and 23% of Gemini’s, but only 11% of ChatGPT’s.

Perplexity: The Widest Net

Perplexity searched on all 36 queries and cited the broadest set of crypto media, 33 different outlets.

It averaged 23 citations per answer, so most mid-table outlets owe their place in AI search citations to Perplexity alone.

Claude: The Most Media-Friendly

Claude gives crypto media the highest share of its citations (34%) and cited a source in 26 of 36 answers.

It named 26 different outlets, with crypto.news first.

Gemini: Rare but Media-Heavy

Gemini cited a source in only 12 of 36 answers, but 10 of those 12 included a crypto outlet.

It named 10 different outlets. Twelve answers are too few for a stable ranking.

ChatGPT: Regulators and Exchanges First

ChatGPT cited a source in 27 of 36 answers, yet only 6 of those 27 included a crypto outlet.

Regulators (SEC, ESMA, investor.gov) made up 39% of its citations and exchange help pages 29%. The media it did cite were The Block (3 queries), CoinDesk (2) and Cointelegraph (2).

When Does an Engine Search at All?

Explainer questions such as “What is blockchain?” rarely trigger web search, so no outlet gets cited.

Claude and Gemini cited nothing on all six explainer queries, and ChatGPT cited a source on one. Only Perplexity searched every time.

Which Crypto Media Outlet Wins Which Type of Query?

Query intent decides the winner: crypto.news leads on news, bitcoinfoundation.org on “best of” questions, and no outlet gets above 4% on explainers.

Each group has six queries, so one query moves a score by about 4 points. Differences of a few points are noise.

News is the one group where every engine cites sources every time, and it is where crypto media is strongest. Founder queries are the ones a project asks when it wants coverage, such as “Where can I distribute a crypto press release?”. Perplexity answered them with press-release sections from crypto.news, coingape.com and cryptoslate.com.

On founder queries, 31% of all citations went to press wires and sponsored-content networks, and 22% to agencies’ own sites. Crypto media took 26%. If you sell a service, the ranked lists on those domains help shape how AI describes your category.

What Kind of Page Do AI Engines Cite?

Fewer than four in ten cited pages are news articles: most of the rest are homepages, tag pages, directories and review pages.

ICODA classified the 158 distinct URLs cited from the 15 top-ranked outlets:

• News or analysis articles: 59 (37%). Mostly dated stories on ETF flows, prices and regulation.
• Homepages, tag and section pages: 35 (22%). For news queries the engines often cite a front page or a tag page such as /tag/bitcoin, which updates all day.
• Directory, best-of and review pages: 32 (20%). CryptoSlate’s wallet and exchange directories and company profiles are the clearest example.
• Press-release or sponsored pages: 7 (4%). Six of the seven are section index pages. One is an individual release.
• Learn and guide pages: 3 (2%). Explainer content is rarely cited because explainer queries rarely trigger search.

Why Do These Outlets Win?

The pattern points to four reasons: freshness, list-friendly structure, a findable press-release section and fit with each engine’s habits.

ICODA did not test causes directly. These are readings of the pattern.

• Freshness. News leaders publish constantly and keep indexable front and tag pages, which live search lands on.
• Structure for lists. Directories, wallet and exchange rankings and company profiles get cited on commercial and founder queries, where the engine wants a list.
• A findable press-release section. On “where do I distribute a press release” queries, outlets with a labelled, crawlable release page get named.
• Engine fit. ChatGPT cites outlets mainly on news-style questions, so The Block and CoinDesk benefit most.

Size is not on the list. For the 19 outlets ICODA also pulled from Ahrefs, citation score had a rank correlation of 0.20 with Domain Rating (p = 0.42) and 0.07 with monthly organic traffic (p = 0.77). Neither is statistically clear: bigger outlets were not reliably cited more.

How Can a Crypto Project Get Cited by AI Engines?

Choose outlets by the question you want to be found for, then earn the page types AI actually cites: editorial coverage, directory entries and list placements.

Generative engine optimization for crypto PR comes down to five steps:

• Pick outlets by question. For launch and market news, the engines cite crypto.news, CoinDesk, Cointelegraph and The Block. For “best of” and review queries, look at CryptoSlate, Coin Bureau and bitcoinfoundation.org. Check each outlet’s current rules for your category first.
• Plan for more than a press release. Only 7 of 158 cited outlet URLs were press-release pages. Ask your crypto PR agency for the profile, directory entry or editorial piece as well as the wire post.
• Treat ChatGPT as a separate target. Your own site, exchange listings and regulatory filings matter there as much as another media mention.
• Get into the lists for your category. Check which ranked lists and agency directories show up in AI search citations for “best” queries, and whether you appear in them.
• Re-test on your own queries. Run your 20 most important questions on all four engines before you spend on placements.

How Was the Test Run?

ICODA is a crypto PR agency that arranges paid placements, but the ranking rule was fixed before data collection and uses no commercial relationship.

Queries came from Ahrefs Keywords Explorer matching terms (US volume) and were fixed before any run. Each ran once through the OpenRouter API with web search forced to each provider’s native search tool. Before scoring, ICODA resolved Gemini’s redirect links, removed tracking parameters, counted each domain once per answer and sorted every cited domain into one of eleven source types. Domain Rating, traffic and AI citation counts come from Ahrefs, pulled 3 October 2026. ICODA’s own site was cited once, by Perplexity.

ICODA’s earlier study, Google’s AI cites crypto media four times less than rivals, used a prompt set ICODA could not see. This one uses a known query set and adds Claude.

Check whether AI cites your project

The free AI Visibility Checker shows whether ChatGPT, Perplexity, Claude and Gemini name your project today. Check if AI cites your project with the free AI Visibility Checker.

Appendix: The 36 Queries

All 36 queries, in six groups of six, ran on every engine.

• News: latest bitcoin news; latest Ethereum news; top crypto news stories; why bitcoin is moving today; bitcoin ETF flows; SEC on crypto ETFs.
• Explainer: blockchain; crypto; staking; mining; immutability; consensus.
• Commercial: best crypto to buy; wallet; exchange; app; app to buy crypto; place to buy crypto.
• Regulation: US crypto tax; CLARITY Act status; GENIUS Act; CLARITY Act vote date; latest crypto regulation; latest stablecoin regulation.
• Safety: Coinbase; Binance; Coinbase Wallet; Robinhood; bitcoin; crypto investing.
• Founder: best crypto marketing agency; finding a crypto PR agency; what an agency does; writing a press release; distributing a press release; best distribution services.

FAQ

Which crypto media outlet do AI engines cite most?

In ICODA’s test, crypto.news ranked first, then CryptoSlate and CoinDesk. Only crypto.news and CoinDesk were cited by all four engines.

Does ChatGPT cite crypto media?

Rarely. Crypto outlets appeared in 6 of its 27 answers that had citations and made up 11% of its citations, while regulators and exchanges made up 68%.

Does a high Domain Rating make an outlet more likely to be cited?

Not in this data. The rank correlation between Domain Rating and citation score was 0.20, which is not statistically clear, on 19 outlets.

Are press releases cited by AI?

Seldom. Press-release pages were 7 of 158 cited URLs from the top 15 outlets, and six of those were section index pages.

Was this writing helpful?

Disclaimer and Risk Warning

The content featured on Coinpedia's press release page is provided for informational purposes only. Coinpedia does not endorse, verify, or take responsibility for the accuracy, completeness, or reliability of any press releases or associated materials. Any views, opinions, or statements expressed in these press releases are those of the respective issuers and do not reflect the opinions or positions of Coinpedia. Coinpedia is not liable for any content, products, services, or actions mentioned in the press releases. Readers should independently verify the information before taking any actions related to the subject matter of the releases.