AI video
India Simulator: the Viral GTA-Style Game That Does Not Exist
A GTA-style India simulator pulled 9.35 million plays in four days, and there is no game behind it. Every scene is AI video. Here is what actually happens in the clip, who is credited for it, how far it spread, and what it says about how good AI video got in 2026.
Someone built a GTA-style India simulator, and it racked up 9.35 million plays in four days. At first glance it really had me believing it was a real game. But after some digging, I found out it was fake. There's no game, and there's nothing to download. Every scene you are about to read about is AI-generated video, cut together like gameplay.
This whole thing came out of the Chinese internet. A gameplay video credited to the creator Dr.Clown PhD on X blew up on Chinese platforms in June 2026 with 9.35 million plays in four days, every mission prompt on screen written in Chinese, then picked up millions more in August when reposts crossed over to X. It looks like an AAA open-world game set in India, but it's video generation with a game HUD composited on top, made by one creator rather than any studio.
If you want to see it for yourself, the creator's page is x.com/DrClownPhD, and the June upload that started everything is on bilibili. I watched it once at full speed and then again slowly, and it held up scarily well both times.
What actually happens in this game?
You start out as an NPC working for a delivery company. In the opening scene your guy runs alongside a moving train, swipes a phone clean out of a passenger's hand, and takes off. The HUD plays it completely straight. There's a mission card that translates to "Extreme Train-Grab: seize your chance, jump on the train," precision-hit markers popping +50, and a status line that reads "phone dropped."
Then it keeps going, one absurd system at a time. The main character pulls out a bottle literally labelled "Ganges Water, Special Supply Edition," and it's a collectible item. The corner of the screen rates it: "Hygiene rating: unknown."
There are dance battles to Indian songs with other NPCs, triggered by a mission that translates to "Rooftop dance-off: when words fail, dance it out." The river drains your HP when you swim in it. At one point you're dodging a police roundup by crossing the river on the back of a giant lizard, through water carpeted edge to edge in floating rubbish.
Then comes the boss fight. After a street-food scene, the stall owner comes at you swinging a wok the size of a bin lid, under a mission update that translates to "The stall owner is enraged: dodge the attacks, defeat the stall-owner BOSS."
It ends the only way it could possibly end: the main character dancing on top of a moving train. In my reel I left that whole dance battle untouched, original audio and all, because nothing I could have cut was going to beat it.
How far did it actually go?
It travelled in two separate waves, and the numbers are real.
The first wave was June 2026, on the Chinese internet. Chinese gaming press 17173 covered it on 9 June with the headline number: 9.35 million plays in four days. That is the wave my reel reacts to.
The second wave came ten weeks later, in mid-August, when a reshare on X sent it around the English-speaking internet and added roughly five million more views. This is the version most people outside China saw, and you can watch it right here:
For the record, that account is a resharer and not the creator. Credit for the original sits with Dr.Clown PhD, whose watermark the footage carries and whom the June press coverage names.
It sparked some genuinely heated debate online too. Plenty of people found it hilarious, some people hated it and called it racist, and that argument is probably half the reason it travelled this far. My own take is that the video punches at the game genre it's parodying, and I react to it on those terms. You can judge it for yourself with the original in front of you.
There is also a separate project going around called GTA Saint Delhi, which layers a GTA-style UI over real footage of Delhi streets, and people keep mixing the two up. That one is real video wearing a fake game skin, made by someone else entirely, while the one in this article is fully generated.
Who made it, and with what?
The HUD is what gives the production away. If you look at those frames again, the mission cards hold the same clean rounded box from scene to scene, the HP bar sits steady at the top right, and the +50 hit markers stay crisp the whole way through. Video models in mid-2026 couldn't hold steady UI text like that across shots, and dense text is still the thing they garble most. So someone composited that HUD on top by hand, the same way a motion-graphics layer goes onto any normal edit, and to me that's the clearest fingerprint of how the whole video was put together.
So the honest answer to how this was made is that an AI video model generated the scenes, someone cut them together like any normal edit, and a designed HUD layer went on top to sell the game fiction. The August wave adds one more data point: ByteDance's Seedance 2.5, which launched at the end of July, made this style dramatically easier to reproduce, and the mid-August reshare wave rode exactly that. The June original predates it, which is part of why authorship chatter around this video stays messy. What's verifiable is the credit the footage itself carries, and that credit is Dr.Clown PhD.
The part I keep coming back to as a marketer is what it actually took to get fourteen million plus views across two languages. One creator with consumer AI tools, a hand-composited UI layer and a genuinely sharp joke, and for about a week it travelled harder than most real game marketing does. I used to be able to distinguish AI from real videos, but as the models get better, it's getting harder for me to do so, and this is the first clip that beat me at full speed.
What I think this means for the rest of us
If a fake game can pull bigger numbers than real game marketing for a week, then AI video is clearly good enough already and that argument is over. What I keep turning over is what this looks like when someone points the same recipe at my industry, or yours, and whether that someone ends up being us or somebody quicker.
I'm not saying everyone should go make parody games. I'm saying one person with a sharp idea and a couple of subscriptions can now produce the kind of attention that used to need a whole production team, and I think most marketing teams are underestimating how fast that floor is moving. I'd be happy to be wrong about that one.
