5x the Pages, 70x the Citations, 1615x the Visitors

Since launch in October 2025, Grok has churned out a staggering quantity of content material: 737,930 revealed pages.

By comparability, Wikipedia has 3,725,102 crawled pages, or 5 instances larger. Given how rapidly Grok is amassing content material, it appears believable that the positioning might overtake Wikipedia’s complete web page rely.

Assuming Wikipedia’s web page rely stays static and Grokipedia publishes new pages on the identical fee (nearly 6,000 pages per day since launch), Grokipedia would overtake Wikipedia in roughly 500 days, or circa July 2027.

However though it’s simple for AI to pump out plenty of pages, it’s not really easy to earn natural site visitors.

We estimate Wikipedia’s natural site visitors at ~2,100,000,000 pageviews per thirty days, in comparison with ~1,300,000 for Grok: a distinction of 1,615x (requiring a log scale to plot on the identical chart):

Put one other method, for each single pageview a Grokipedia web page receives, Wikipedia receives 1,615.

This site visitors distinction is defined partly by key phrase rankings. Grokipedia ranks for 1,193,013 key phrases with a median key phrase place of 40.59. Wikipedia ranks for 45,573,106 key phrases, or 38x extra, with a median place of 12.26:

Wikipedia has 879,615,708 backlinks to Grokipedia’s (nonetheless very spectacular) 1,414,654 backlinks, or 622x extra. Dividing backlinks by complete web page rely, Wikipedia has ~236 backlinks per web page to Grokipedia’s ~2:

It’s an identical story for referring domains: 2,024,916 for Wikipedia and eight,016 for Grokipedia, or 253x extra. Wikipedia has 0.54 referring domains per web page to Grokipedia’s 0.01.

By all conventional search metrics—bar one, variety of pages—Grokipedia has a staggering quantity of catch-up to do.

We randomly chosen 100,000 Grokipedia-Wikipedia article pairs (two articles overlaying the identical subject) to check content material construction for each wikis.

Grokipedia pages common 3,942 phrases in contrast with Wikipedia at 3,495, a distinction of round 13%—barely longer, however much less of a distinction than I expectedm given how “low cost” generative AI has made content material creation:

Grokipedia launched in October 2025, so the typical age of its pages (as measured by the typical time since our crawler first detected the URL) is barely 67 days. Wikipedia, by comparability, has a median web page age of three,381 days—9.3 years, or nearly 51x older:

Surprisingly, Grokipedia has extra exterior hyperlinks in its reference part: a median of 68 per web page, in contrast with Wikipedia’s 40.

However Wikipedia, with over 5 instances as many complete pages, is a lot better at inner linking between associated ideas, with a median of 144 distinctive inner hyperlinks per web page to Grokipedia’s 83:

Photographs present the starkest distinction. Wikipedia’s human-written, human-edited content material incorporates a median of 24 pictures per web page. Grokipedia has a median of… zero pictures per web page:

Grokipedia is comparatively good at incomes AI citations.

Regardless of having have 1,615x the overall site visitors of Grokipedia, Wikipedia has “simply” 70x the variety of AI citations: 356,200 for Grokipedia versus 24,914,778, or a median of 0.48 citations per web page for Grokipedia, and 6.69 for Wikipedia:

Grokipedia has earned probably the most citations from ChatGPT and Google’s AI Mode, with Copilot and Gemini a distant third and fourth:

By comparability, the majority of Wikipedia’s citations come from AI Mode and AI Overviews, with ChatGPT and Perplexity in third and fourth:

We are able to additionally have a look at the relative distribution of citations throughout totally different AI platforms by evaluating each wikis on the identical chart (with totally different axes… please excuse my graph crime).

Most notable is the distinction in citations from AI Overviews and ChatGPT:

On condition that Wikipedia is an important supply of coaching knowledge for many LLMs, and Grokipedia is written totally by generative AI, I needed to know: how related are Grokipedia and Wikipedia pages?

To search out out, we in contrast 10,000 “subject pairs”: pairs of Wikipedia and Grokipedia pages overlaying the identical subject and referring to the identical entity.

The imply semantic similarity of those subject pairs was 0.791, and the median was 0.831:

That is very sturdy semantic overlap. For context, if we take Grokipedia pages and compute their cosine similarity with random Wikipedia pages (that’s, mismatched pairs), the imply cosine similarity is ~0.62.

Equally, if we take different pages that rank for a similar key phrases as Grokipedia and Wikipedia pages, the typical cosine similarity is ~0.61:

83% of the subject pairs we analysed had a cosine similarity larger than 0.62, indicating that Grokipedia pages share a really sturdy semantic overlap with Wikipedia pages describing the identical entity:

Grokipedia appears to symbolize a brand new mannequin for wikis: AI-generated content material designed for AI consumption.

Grokipedia has a comparatively sturdy affect on AI citations, relative to its measurement, however a comparatively weak affect on actual folks (as measured by estimated pageviews). Grokipedia pages are similar to Wikipedia pages on the identical subject, and it’s onerous to keep away from the picture of the ouroboros consuming its personal tail. Actually, the studying expertise appears extra meant for AI consumption than human.