NoPixel viewer count: 20,000 one day, 127,000 the next

This morning NoPixel drew between 87,609 and 88,300 concurrent viewers across Kick, Twitch and YouTube — 37.5% to 38.8% of every live Grand Theft Auto stream on the planet, across three complete censuses between 04:09:46 and 04:13:45 UTC on 23 October 2026.

That is the number, and you should not trust it for long. Eleven days ago the same instrument, pointed at the same minute of the same clock, read 20,212. Eight days ago it read 127,364. Nothing structural happened to NoPixel in between. What changed was which three or four people happened to be awake.

This is the problem with every NoPixel viewer count you will find, including the ones we have published. A single morning’s reading of a roleplay server is not a measurement of the server. It is a measurement of a roster.

What we measured this morning

Three complete censuses of every live Grand Theft Auto stream on Kick, Twitch and YouTube. Each census reconciles exactly against our own population total — same publish timestamp on all three endpoints, no duplicate channels.

Census (UTC)	All live GTA	NoPixel viewers	NoPixel channels	Share
04:09:46	232,535	88,300	122	37.97%
04:11:45	233,550	87,609	121	37.51%
04:13:45	226,424	87,861	121	38.80%

Four minutes apart, the share moves by 1.3 points — and almost all of that is the denominator, not NoPixel. The server itself is steady inside a single morning. It is between mornings that it falls apart.

The same measurement, ten times

Every reading below was taken the same way, in the same 04:0x–04:2x UTC window, by matching nopixel or no pixel in the stream title across the full live population:

Date (2026)	NoPixel viewers	Share of live GTA	Largest single channel
9 Sep	125,031	39.2%	—
22 Sep	116,168	36.2%	—
12 Oct	20,212	16.2%	—
15 Oct	127,364	49.7%	—
17 Oct	92,133–92,250	39.7%	xQc (25.6–26.0%)
18 Oct	38,088–38,569	21.5%	omie (28.2–28.3%)
20 Oct	96,139–101,686	43.6%	xQc (39,065 on two platforms)
21 Oct	66,361–67,051	30.3%	buddha (15,629, 23.6%)
22 Oct	83,821–84,729	43.7%	xQc (36,063 on two platforms)
23 Oct	87,609–88,300	37.5–38.8%	xQc (35,236–35,975, 39.9–41.0%)

Read down the share column and there is no trend to find. It goes 39, 36, 16, 50, 40, 22, 44, 30, 44, 38. Between 12 and 15 October the audience grew 6.3-fold in three days. Between 15 and 18 October it fell 70%. Those are not corrections to each other; they are the same instrument, working correctly, on four different mornings.

A caveat on that table worth stating before anyone quotes a row: 9 September was day one of the beta with Twitch paying GTA$1.5m in Drops to anyone who left a stream open, and 12 October was the lowest live-GTA reading in our thirty-day window — 116,908 against 232,535 today. Some of the swing is the whole category breathing. Most of it is not.

One person is 40% of it

Here is the mechanism, and it is not subtle. Of the 88,300 viewers watching NoPixel at 04:09:46 this morning, 35,236 were watching xQc — who was streaming the same session to Twitch and Kick simultaneously, under the same title, and so appears in our data as two rows:

Channel	Platform	Viewers	Share of NoPixel	Share of all live GTA
xQc	Twitch	24,859	28.2%	10.69%
buddha	Kick	14,725	16.7%	6.33%
xqc	Kick	10,377	11.8%	4.46%
omie	Kick	5,195	5.9%	2.23%
AnthonyZ	Twitch	4,260	4.8%	1.83%

One person holds 39.9% to 41.0% of what anyone would call “NoPixel’s audience”, and 15.2% to 15.9% of all live GTA viewing anywhere, on his own, across the three censuses. The top three rows — two of which are the same man — are 56.6% to 57.1% of the server. The top ten are 77.7% to 78.3%.

Strip xQc out and NoPixel read 51,886 to 53,064 this morning. That is a different answer to the same question, asked at the same minute, and the only thing separating them is one person’s streaming schedule.

The identity of that person keeps changing, which is the part that makes a single reading useless. Across the last six mornings the largest channel on the server has been xQc, omie, xQc, buddha, xQc and xQc. On 18 October xQc was not live on V at all and the server read 21.5%; on 21 October buddha led it and xQc was not in the top ten; today he is back on top for a second day running — the first time in this series the leader has repeated.

It is a roster question, not an audience question

Our thirty-day board makes the same point from the other direction. Of the 100 largest GTA channels by average concurrent audience over the thirty days to 04:04 UTC today, 23 are roleplay channels. Here are the ten largest of them, and whether they were actually on NoPixel this morning:

Channel	Platform	30-day avg viewers	30-day airtime	On NoPixel now
paulinholokobr	Kick	39,535	31.3 h	offline
Jynxzi	Twitch	31,693	7.2 h	offline
xQc	Twitch	27,643	265.1 h	24,859
buddha	Twitch	13,911	226.8 h	offline
Jerma985	Twitch	10,806	53.7 h	offline
xqc	Kick	9,660	267.4 h	10,377
buddha	Kick	9,229	310.9 h	14,725
lordkebun	Kick	7,870	144.1 h	offline
Chatterbox	Twitch	5,782	233.0 h	offline
AnthonyZ	Twitch	5,263	269.6 h	4,260

Four of the twenty-three are on the server right now. The rest are asleep, playing something else, or on another server — and each of them is capable of moving the headline figure by five to forty thousand viewers the moment they go live. buddha is the clearest case on the table: his Kick channel is carrying 14,725 viewers this morning while his Twitch channel, which averages 13,911 over thirty days, is dark.

So when a figure for NoPixel’s audience swings six-fold in four days, nothing has happened to NoPixel. The right question is not “how many people watch NoPixel” but “how many of a dozen very large creators are logged in”, and those are different questions with different answers every morning. The structural version of this argument is in our October piece on whether NoPixel V is dead; eleven readings later, it is the part that held.

What this does not show

We count streams, not players. NoPixel V is invitation-only at roughly 450 players. Everything here is viewership of people streaming it, which is a much larger and completely different population.

Multistreaming double-counts people, on purpose. xQc’s 35,236 is two simultaneous audiences summed. That is the correct answer for “concurrent viewers” and the wrong one for “humans” — someone watching him on Twitch is not the same person watching on Kick, but we cannot tell you how many switched tabs. Our September measurement of simulcasting put the simulcasters at 28.2% of everyone watching GTA, so this is not a one-creator artefact.

The server is identified by stream title. A NoPixel stream whose title does not say so is invisible to us, and a title saying “NoPixel” while the streamer plays elsewhere is counted. Matching on nopixel/no pixel is the loosest sound rule we have found; filtering further — to titles containing “V” or “5” — undercounts by about 13%, because it drops buddha’s Kick title.

Ten readings is ten mornings. Every census here is from one narrow window at 04:0x–04:2x UTC, which is American evening and European night. We have no equivalent series for 18:00 UTC, and the composition of a roleplay audience almost certainly differs by hour. The readings on 9 and 22 September and 12 and 15 October come from earlier runs of the same instrument, recorded with their census times; 17 October onward are daily.

We make no causal claim. We can show that the figure tracks which large creators are live. We cannot show that this is the only thing moving it, and we have not tried to separate a creator going live from whatever made them go live.

Citing this

Permalink: https://luciastream.com/blog/nopixel-viewer-count/. The figures are three full-population censuses at 04:09:46, 04:11:45 and 04:13:45 UTC on 23 October 2026, plus nine dated readings back to 9 September, all taken in the same clock window by the method described in our methodology. Current numbers are always on the live board.

If you want one number, use a range and date it. Any single figure for a roleplay server’s audience — ours included — is a statement about one morning’s roster, and it has a half-life of about a day.