Ask HN: Why isn't AI spawning profitable indie games? | Hacker News Reader

Ask HN: Why isn't AI spawning profitable indie games?

7 pointseveningsun13 comments

Everyone’s hyped about Claude Code, Nano Banana models, and how AI can now “build a game in 30 minutes for $ 20.”

If that’s true—why don’t we see a flood of ultra-low-cost, high-quality tower defense games (like Kingdom Rush) built with AI, raking in serious revenue?

After all:

Art? → Image generators + style locking. Code? → LLMs can scaffold Unity/Godot logic. Balancing? → Maybe RL or even manual tuning on a tiny scope. Audio? → AI voice + music tools exist.

And yet… the App Store isn’t overrun with polished, profitable Kingdom Rush–style games made by solo devs using AI. Why?

P.S. Yes, this post was AI-generated—my first attempt at market research before launching Kingdom Rush: Nano Banana Edition.

View on HN

The same reason why 300 million games with hundreds of programmers flop. Game “fun” is much more complex than just coding or art.

The other thing is that games are getting more complex and audience is demanding. Even if you can use AI for your solo project the bar is higher than 10 years ago.

because 90% of people raving about 1000% productivity boost from LLMs are students and junior-level programmers, whose productivity was nonexistent in the first place. for programmers with 5+ years of experience with the tools they're using, LLMs are a 10-20% boost, at most, and I'm saying that as a very AI-positive individual.

Because as soon as the indie audience catches wind that a game uses AI for any code, artwork, or writing, there goes the "indie" cred. It's not cool anymore. Worst case, some ludds launch a vocal public crusade against you.

Rest assured, there is plenty of AI being used, but the game devs know better than to talk about it.

Fair point. But whether a game is indie or not isn’t really my main concern here. What I’m wondering is this: you probably know how much money simple games like Doodle Jump have made. It seems entirely feasible for a somewhat experienced developer to whip up a shameless clone—with a fresh coat of paint—in just a few days. Sure, it wouldn’t make a fortune, but it should still earn some money—enough to be worth the effort.

So why haven’t we seen a flood of these kinds of games on the App Store?

We have, look at Playstation store, App store, in the new games section. None of them are successful so it doesn't break your main feed. Someone can estimate no of game coming out with AI in them, but I am not that guy, I can give you a percentage and its like 70%.

Thats some bullshit, if your game is good and uses AI, users seem to not care about, for eg Arc Raiders.

But vast majority of games are shit and even more the ones where they use AI, and there is already enough slop on the market with AI, that users know they need to stay away from those games.

AI can't innovate, it can't make anything new that doesn't already exist.

So most games made by AI are just tutorial rehashes as that all the LLMs have to generate off.

The pipeline and incentive structure probably isn't there yet. I've hung out with game designers who are either bullish on AI but haven't integrated it, or just aren't tapped in at all.

Reminds me of the web3 wave a few years back, lots of people trying to make games with no actual understanding of what makes games fun and popular, not just profitable. The tech isn't the bottleneck, the taste is. Same applies to AI art

I saw this YouTube channel where this animation designer created a game. What I've noticed, even with the coding part fully scripted with Claude Code, his game still will take years to create because the animations won't be done by Claude Code. It looked great too.

He wouldn't be creating a game if it wasn't for Claude Code, so there's that.

I'm sorry that I can't find the source anymore.

AI cant make complex games yet (complex here means more than one level consistently, its a very low bar, a game dev with 20min experience can achieve this). When you use a tool like Unreal or Unity or Godot, its not just code, its also thinking about lighting, music, art, surfaces, shaders, materials and so much more. And also a lot of things have to be done in a visual way (connecting nodes, placing models)

Creating a game, even when you have capable team of 2-3 members, its very very broad objective. You have to be master of so many things. During the vibe coding jam, there were so many games build by AI, but none of them were any good, AI was making three.js games mostly and it was placing objects so stupidly that the term ASS (artificial stupidity sourced) code is a term on discord indie scene. For eg, in a car came it was placing objects on the road.

You can see https://geminimakesrally.vercel.app/ and judge for yourself. Its terrible and its funny for now. It uses the latest Gemini 3 Pro model from December. Can you sell this game? Maybe on mobile if you want to add to the slop store more, I personally dont want to post slop games like this. Removed it from my itch too as soon as we completed the

And before you say its me being shit at vibecoding, yes fair, but I am fairly competitive in non game vibecoding: https://plonkedin.vercel.app/ (made in couple mins)

1) Steam hates AI. Steam is at least 80% of the PC market. If you were doing multi-platform, it's just too big of a loss.

2) The indie community is hostile towards AI. Expedition 33 was disqualified for barely any AI use. Even things like brainstorming titles or generating textures will get you exiled.

3) Nano banana tier AI doesn't seem consistent enough to do assets properly. You need to remove the backgrounds. You need to wrestle it into following instructions. It does save time and you can do things like draw 5 guns and have it generate 4 more in the same style, but that's not much better than Diablo 2 recoloring units.

4) While not outright hostile, most people below 20 recognize AI art with a glance. Capsule and ads are immediately indicative of the game being AI slop even if it's not. If you have something AI-looking in the screenshots, people simply move on to the next game.

5) While Elevenlabs produces some impressive sounding audio, they're still not at a voice actor tier. They're only good enough for scaffolding and trigger uncanny valley.

Pretty sure y'all remember when levelsio was making his slop vibecoded "game" lmao and somehow it became viral...

The game didn't become viral because of its good quality, it became viral because it was made with AI, and the AI community (and HN) gassed it up. The author made money because he stuck ads in it while it was viral. The exact same game made entirely by a human being would be dismissed as garbage.

Reply on news.ycombinator.com