Why the AI Revolution Hasn't Hit Game Development Yet

Dark mode

Why the AI Revolution Hasn't Hit Game Development Yet — And Why It's Coming Next

In his latest video, Why Games Now Take 6+ Years To Make, game development reporter Jason Schreier talks about why it takes current gen games up to and over 6 years to develop.

This got me curious.

AI has revolutionized programming!

Simple yet impressive demo games get built by a single prompt in under 10 minutes.

I myself have recreated Risk using AI in a day. (Source code)

I still remember slaving away day and night for 2 weeks, over 5 years ago, to bring Q3DM17 The Longest Yard into the browser using Three.js. (Demo / Source code)

Then recently I tried my luck with an always solvable Yukon Solitaire, where I cut my teeth on my first Rust project to find solvable boards.

I told the same to Claude Code and it implemented a solution that was better and faster in 3 prompts.

Yet still, Jason Schreier, knowing all of this current development full well, releases the above video. It's even a follow-up to his video from 4 days ago titled "Why Video Games Cost So Much To Make".

How is that possible? Shouldn't video games get cheaper and faster to produce, now that we live in "The Agentic Era"?

Well the answer is actually pretty simple:

What did we get so far from AI?

Complete song generators

Picture generators

Video generators

Text generators

Code generators

Text to Speech

Speech to Text

What do games need?

Sound effects and music that dynamically adapts to the action

Textures rather than complete pictures

3D engines rather than video

Dialogue trees with branching conversation options

Quests and storylines

3D models and levels

Animations

QA

So the simple answer is that large AI companies focused on the more common use cases first:

Large detailed images

Crazy ideas come to life videos

Entire genre songs

Chat and text to write your emails

Code generators for the everyday software engineer

Text to speech so we can finally listen to The Catcher in the Rye

Speech to Text so we can get the text of a Sean Carroll AMA video and skip to the parts we care about

Ok, got that. So when will these AI companies focus on game development and fix Starfield?

Basically now. In internal teams, this has been given as the next frontier.

Keep in mind that the games industry dwarfs the movie business.

The global video game industry generates roughly $190–250 billion per year, depending on the estimate and what is included (games, mobile, subscriptions, etc.).

Global movie box office revenue is usually around $30–40 billion per year.

Mobile gaming alone generates over $100 billion annually, making it larger than many entire entertainment sectors by itself.

Some examples that show the scale:

Grand Theft Auto V has sold over 200 million copies and generated billions of dollars in revenue over its lifetime.

Fortnite became such a cultural force that it hosted virtual concerts with millions of simultaneous viewers.

The delay of Grand Theft Auto VI was estimated by analysts to affect industry revenue by $2.7 billion in a single year.

The expectations for Grand Theft Auto VI are honestly on a scale the industry has never seen before. Some analysts estimate around $1 billion before release from pre-orders alone, $2–3 billion in its launch period. Market research firm DFC estimated roughly $3 billion in total sales during the first year, including pre-orders. For comparison, Grand Theft Auto V made about $1 billion in 3 days when it launched in 2013.

These are numbers large AI companies such as OpenAI, Anthropic, Google, Microsoft, Meta and Amazon won't ignore much longer.

Keep in mind that money maker Nvidia even started as a gaming hardware company. It would be more than natural for Nvidia to come full circle and provide the basis for AI generated games.

Amazon has the hardware and the media experience with gaming they need to go all in on this. They got bitten especially hard with exactly the problems Schreier describes in his videos, with their games New World and Crucible.

I won't talk much about Meta in this context, because… well we all know why. 🙂 Holy shit Zuck, get your act together man!

Google had an interesting experience in the world of gaming with Google Stadia — I still have 2 controllers lying around here somewhere. I loved the idea, but holy cow, having to buy the games again — how did that slip through in the business plan meetings?

In any case, Google is no stranger to the world of gaming. Yeah, they paid dearly for great tech that got destroyed by a stupid business plan, but who says they won't strike again given a good business plan. And there is one. At the very least Google will generate the hot mobile game of the month very soon. Why pay creators if you can gen it?

Tell an AI your game idea and out pops your game.

You think I'm crazy?

I guarantee this is coming next and right freaking soon.

Will it be one prompt to make...