Will GTA 5 die when GTA 6 comes out? Roleplay says no

No — and the reason is a mode Grand Theft Auto VI will not ship with.

At 04:07–04:21 UTC this morning, 34.7% of everyone watching GTA live across Kick, Twitch and YouTube was watching roleplay: 123,061 people out of 354,178, on 877 of 2,032 live channels. On Twitch alone, roleplay was 52.9% of the entire GTA audience. Roleplay runs on FiveM, FiveM runs on PC, and GTA VI arrives on 19 November as a console game. Whatever happens to the old game’s player count, the audience that shows up to watch it is currently built on the one thing the new game cannot do on day one.

The prediction we are checking

This is not our thesis. It belongs to Dan Clancy, Twitch’s CEO, who was asked about it at The Game Business Live in June 2025 and said (reported by Dexerto and RockstarINTEL):

The interesting thing was it was going to be a little strange because whenever it does come out it’s coming out on the console and roleplay will not be part of the initial launch and so GTA V will still interestingly be popular on Twitch even as everyone’s playing GTA 6 for a while.

He said that fifteen months and two delays ago, before the date was 19 November. “Whenever it does come out” is doing a lot of work in that sentence, and it is why the quote survived the reschedule. It is also a falsifiable claim with a mechanism attached, which makes it worth more than most executive commentary: if roleplay is a rounding error in GTA’s audience, Clancy is wrong regardless of what platform GTA VI ships on.

It is not a rounding error.

The mechanism, re-checked this morning

We read rockstargames.com/VI directly rather than trusting our own back catalogue. It says “Coming November 19, 2026” and lists PlayStation 5 and Xbox Series X|S. The words “PC” and “roleplay” appear on that page zero times — a fact worth stating plainly, because a claim that Rockstar has “confirmed” roleplay for GTA VI has started circulating off the back of the NoPixel V announcement, and NoPixel V is a GTA 5 product. We traced the same kind of decay through the streamer-mode rumour four days ago: one person’s recollection became a stated feature in about five days.

So the chain we set out on 4 August still holds, link for link: no PC version means no FiveM, and no FiveM means no roleplay of the kind 877 channels were running this morning. That is a claim about hardware, not about Rockstar’s intentions, which are demonstrably warm.

This is a different question from whether Rockstar keeps the lights on, which we answered separately in August: GTA Online is not shutting down, and the evidence there was a heist shipped to PS4 and Xbox One in July. Here the question is narrower and more interesting — not whether the old game still runs, but whether anyone is still watching it.

What we measured

Fourteen complete censuses of every live GTA stream on the three platforms, taken between 04:07:02 and 04:21:01 UTC on 10 September 2026:

	viewers	share of viewing	channels	share of channels
Roleplay	123,061	34.7%	877	43.1%
Everything else	231,117	65.3%	1,156	56.9%
All GTA	354,178	100%	2,032	100%

Sixteen days earlier we ran the same measurement in almost exactly the same clock window and published it. The comparison is the part that surprised us:

Same 04:0x UTC window	25 August	10 September	change
All GTA viewers	198,679	354,178	×1.78
Roleplay viewers	36,944	123,061	×3.33
Roleplay share of viewing	18.6%	34.7%	+16.1pp
Roleplay channels	812	877	×1.08

Roleplay’s share of the audience roughly doubled in sixteen days, and it did not do so by roleplay channels multiplying — there are only 65 more of them. It happened because the existing ones got much, much bigger.

A trailer buys a spike. A server bought a week.

The 27 August Extended Look is the most-watched thing that has ever happened in this dataset: 992,344 concurrent viewers at 19:21 UTC. (Our write-up on the night reported 959,078 at 19:25, which was the highest five-minute sample; the peaks endpoint keeps the 60-second maximum and caught the apex four minutes earlier. On a spike that steep the two resolutions will always disagree, and both numbers are the truth about their own sampling rate.)

Now look at what that record did to the day it landed on. Nothing. 27 August averaged 268,342 viewers across its 24 hours, which is an unremarkable Thursday, and 28 August fell back to 236,904.

Then NoPixel V’s closed beta opened on 8 September, and the two days after it are the two largest days of GTA viewing in our entire record — 332,802 and 342,747 as daily means, against 242,982 for 1–7 September. The 9 September total came to 8,225,934 viewer-hours; the trailer day managed 6,440,213.

Daily means still mix a launch hour with a quiet one, so the cleaner comparison holds the clock still. Averaged over the same four hours, 00:00–03:59 UTC:

Same hours, UTC	mean viewers	vs 1–7 Sep
1–7 September	259,661	—
8 September (beta not yet open)	288,093	×1.11
9 September	487,114	×1.88
10 September	475,279	×1.83

The 8 September row is the control, and a useful one: Rockstar never published a time of day for the beta, but we ran this census at 04:00 UTC that morning and found not one stream on NoPixel V, so those four hours are genuinely pre-launch. They read ×1.11 — noise. The next two mornings read ×1.88 and ×1.83, twenty-four hours apart, the second 2.4% below the first. That gap is far too small to be a decay curve and far too large to call the two identical; what it looks like is a plateau. A trailer is an event — it spikes, then it resolves. Two days on, this has not resolved.

The number Clancy would care about

He runs Twitch, and Twitch is not where GTA’s audience normally lives — Kick usually carries the majority, which is the whole point of the platform comparison we published in August. Split this morning’s census by platform and the roleplay concentration is lopsided in a way that maps precisely onto his argument:

Platform	GTA viewers	of which roleplay	roleplay share
Twitch	163,904	86,715	52.9%
Kick	187,181	35,527	19.0%
YouTube	3,093	818	26.5%

More than half of what Twitch’s GTA audience watches is a mode that will not exist on GTA VI on 19 November. On Kick it is roughly a fifth. If you are Twitch, “roleplay will not be part of the initial launch” is not an observation about a niche — it is an observation about most of your GTA inventory.

The YouTube row deserves a caveat rather than a conclusion: on a base of about 3,093 viewers its roleplay share swung between 13.1% and 46.9% across fourteen censuses fourteen minutes apart. Treat it as noise, not as a level.

Rockstar is still building for the old game

Seventy days before shipping GTA VI, Rockstar published “Introducing nopixel V” on its own Newswire — an officially backed roleplay server, distributed through the Rockstar Games Launcher, for a game released in 2013. Twitch then ran GTA RP Week alongside it, with Drops worth up to GTA$1,500,000 for watching NoPixel V streamers between 8 and 30 September.

Two companies with a direct commercial interest in GTA VI’s launch spent the run-up to it promoting GTA V roleplay. That is not a leak or an inference; both posts are on their own websites. It is also the most straightforward reading of why Clancy sounded relaxed about the “strange” launch: he was describing something his own platform was already planning for.

What this does not show

The largest problem with the measurement above is the last paragraph. The Drops campaign pays people, in in-game currency, to leave exactly these streams open, it is Twitch-only, and it runs until 30 September. We measured the effect of a Drops campaign in August and it was large. So the honest reading of 52.9% on Twitch is that it combines real interest with a paid incentive, and nothing in our data can separate the two. The level is inflated. Whether the mechanism is real does not depend on the level, which is why the argument here rests on roleplay existing at scale and on PC, not on it holding 34.7% forever.

Four more limits, each of which cuts against the number we published:

Content type is inferred from stream titles, so 34.7% is a floor rather than an estimate. Our classifier walks an ordered keyword table and takes the first match, and first-playthrough is tested before rp — so a stream titled “PLAYING NOPIXEL RP FOR THE FIRST TIME!!”, which had 9,454 viewers this morning, is filed as a first playthrough and never counted as roleplay. Streams whose titles name NoPixel or roleplay while classifying as something else held another 11,029 viewers. Counting them would put roleplay at 37.9%. We publish the smaller figure.

This is one clock hour, and 04:0x UTC is late evening in North America, which flatters an English-language server. NoPixel is one, and it shows: 88.4% of this morning’s roleplay viewing carried an English language tag, against a category that is mostly not English the rest of the day. Roleplay’s share moves with the hour, so treat 34.7% as a reading and not a constant.

We count concurrent viewers, not people or players. For Clancy’s claim that happens to be the right metric — he was talking about what stays popular on Twitch — but “will GTA 5 die” in the sense most people mean it is a question about players, and we do not measure players at all. Nothing here forecasts what GTA VI will do on 19 November either. It measures the size of the audience GTA VI cannot serve, which is a different and smaller claim.

Roleplay is a plurality, not a majority. The biggest single stream in our window was xQc on Twitch at 40,850 viewers, and counting the Kick simulcast he was running at the same time he held 52,871 — under the title “GTA 5 IS BACK”, playing freeroam, not roleplaying. Nearly two-thirds of GTA viewing this morning was something other than roleplay, and that two-thirds is exactly the part GTA VI is well placed to take.

A prediction, so you can grade us

Clancy’s claim is testable and so is ours. Ours: on 19 December 2026, one month after GTA VI ships, roleplay will still be more than 20% of live GTA 5 viewing in this same 04:0x UTC window, and GTA 5 roleplay viewership will be above 25,000 concurrent. That is a fall of more than a third from this morning’s 123,061 and it would still leave Clancy right.

We think that is the likely outcome because the roleplay audience has nowhere to go: a GTA VI roleplay scene needs a PC version, a PC version has not been announced, and the console-to-PC gap on the last three Rockstar titles ran from seven to nineteen months. If roleplay falls below either threshold, we were wrong, and we will say so on the same URL.

The cleaner test comes sooner. The Drops stop on 30 September, and we have already committed to re-running this census in the same clock window when they do. That number — roleplay’s share with the subsidy switched off — is the one that actually settles whether this morning was a campaign or a shift.

Check us

Every figure above comes from fourteen complete censuses of every live GTA stream on Kick, Twitch and YouTube between 04:07:02 and 04:21:01 UTC on 10 September 2026, taken from data.luciastream.com. In all fourteen, the top-400 board and the tail file carried matching timestamps, the per-stream counts reconciled exactly to the published totals in every one, no channel appeared twice, and no platform reported degraded coverage. Every figure quoted above is the mean across the fourteen — 354,178 viewers on 2,032 channels — and the individual samples ranged from 343,106 to 365,365. Daily and hourly figures come from the same host’s 30-day hourly series, which stores hourly means — so a short spike reads lower there than its instantaneous peak, and the 992,344 record comes from the peaks endpoint instead.

The 25 August comparison was measured by us in the same clock window and published sixteen days ago, before we had any idea what this morning would look like, which is the only reason it is worth anything.

Collection is described on our methodology page, the live board is at /live/, the roleplay split by language is at /language/, and the permalink for citation is https://luciastream.com/blog/will-gta-5-die-when-gta-6-comes-out/.