Before GTA 6 Trailer 3 drops, 400 channels hold 94% of viewers

At 16:03 UTC on 31 July 2026 there were 264,439 people watching 2,668 live Grand Theft Auto channels across Kick, Twitch and YouTube. Four hundred of those channels held 93.9% of the viewers. The other 2,268 channels — 85% of everyone broadcasting — split 16,256 viewers between them, an average of just over seven people each.

That is the number worth keeping in mind while everyone waits for Trailer 3.

When will GTA 6 Trailer 3 come out?

Rockstar has not said, and nobody outside Rockstar knows. What is actually confirmed is thin: GTA VI ships 19 November 2026, and Take-Two has scheduled its first-quarter call for the morning of 7 August — the first results covering the pre-order window that opened 25 June. Everything else is inference. Insiders including Tom Henderson and NateTheHate have pointed at August for the next big beat, and fans catalogued several quiet changes to the GTA VI website on 29 July. Treat both as unconfirmed. Website tweaks have preceded exactly nothing before.

The wait has itself become a format. Sitting tenth in our board at that moment was Cross on YouTube, 3,505 concurrent viewers, streaming in Portuguese under the title “PLANTÃO GTA 6: TRAILER 3 DIA 6? ONDE ESTÁ O AGRESSIVO MARKETING PROMETIDO?” — GTA 6 news desk, trailer 3, day six, where is the aggressive marketing we were promised? Someone is counting the days on stream, and three and a half thousand people are counting along with them.

How concentrated is the GTA streaming audience?

Severely. Here is the split between the 400 channels our board ranks and everything below them:

Segment	Channels	Share of channels	Viewers	Share of viewers	Avg per channel
Top 400	400	15.0%	248,183	93.9%	620
Everyone else	2,268	85.0%	16,256	6.1%	7
Total	2,668	100%	264,439	100%	99

A single channel, absi on Kick, was drawing 34,307 viewers — 13.0% of the entire global GTA audience from one Arabic-language broadcast. The top fifteen streams together held 44.5%. If Trailer 3 lands and you want to know how the audience reacted, you are really asking about a few hundred channels, and mostly about a few dozen.

Every Kick GTA channel ranks in the global top 400

This is the cross-check that convinced me the concentration is real rather than an artefact of how we rank. Comparing the full platform aggregate against the ranked board:

Platform	Live channels	On the board	Board’s share of its viewers
Kick	154	154 (100%)	100%
YouTube	231	38 (16.5%)	97.5%
Twitch	2,283	208 (9.1%)	77.4%

Every single Kick channel streaming GTA at that moment was large enough to place in the top 400 worldwide. Not most of them — all 154. Meanwhile 2,075 Twitch channels sat below the cut, sharing 14,785 viewers.

The two slices agree, which is the point of running them: the aggregate totals come from our summary feed, the board from the ranked stream rows, and they reconcile to the same 264,439. Where they differ — Twitch’s board covering only 77.4% of its viewers — the difference is the long tail, exactly where it should be.

Is GTA streaming mostly roleplay?

No, and this surprised me more than the concentration did. Roleplay gets the coverage, the documentaries and the drama threads. It is not what most people are watching.

Content type	Viewers	Share	Channels
Freeroam	189,762	71.8%	1,984
Roleplay	50,654	19.2%	452
GTA Online	14,429	5.5%	63
Mods	4,636	1.8%	58
Just Chatting	2,735	1.0%	10

Freeroam — someone messing about in Los Santos with no server, no character, no storyline — is nearly four times the audience of roleplay. That matters for Trailer 3 because freeroam is the format with nothing to lose. An RP server has a running narrative and a cast to coordinate; a freeroam streamer can pivot to reaction content the minute a video goes up, and 1,984 of them are already live on any given afternoon.

I would not over-read the split. Content type is inferred from titles, tags and category metadata, and the boundary between “freeroam” and “GTA Online” is genuinely fuzzy — some of that 71.8% is people in public Online lobbies who would describe themselves differently. The gap is far too wide to be classification noise, but the exact figure is softer than it looks. Our methodology page sets out how the tagging works.

What this actually predicts about Trailer 3 week

Less than you would hope, honestly, and I would rather say that than dress a single reading up as a forecast.

This is one snapshot, taken at 16:03 UTC on a Friday. It is not a trend line. Viewership rotates hard through the day as time zones wake up, and a mid-afternoon UTC reading flatters some regions and buries others. We have been collecting since 29 July, which is nowhere near enough history to say what a normal Tuesday looks like, let alone what an abnormal one would.

What it does establish is a baseline, and baselines are the useful thing to have before an event rather than after. When Trailer 3 arrives, the questions worth asking are whether that 93.9% concentration tightens or loosens, whether the freeroam share moves as reaction content floods in, and whether the long tail of 2,268 tiny channels finally gets a night where it matters. Those are all measurable against this. You can watch the same numbers move on the live board, check the leaderboard for who is climbing, or see the full language breakdown, which is its own strange story.

My honest expectation is that a trailer concentrates attention rather than spreading it — the biggest channels absorb the spike and the tail stays a tail. I would like to be wrong about that. We will have the numbers either way.