How long it takes to make your first game with AI

Can you know how long your first game will take?

No, and anyone who promises a timeline is guessing. The clock on a first game depends on your choices, not on the tool: the size of the game, the art style, how many times you change your mind along the way, and how many hours are left in your week.

AI changed part of the math. Today it writes the code, the set of instructions that makes the game work, from what you describe in plain English. But you are still the one who decides what the game is, tests it, and says when it is good, and that part takes as long as it takes. See what AI does and what stays with you.

What you can do is look at how long real games took, understand what weighed on each one, and choose a first game that fits your life.

How long did it take people who made a game with AI?

The best documented case is Jonathan Bjerk's. He is an accountant in Norway, was 38, and worked full time. As for programming, he knew only a little HTML and CSS, the basics for building simple web pages¹.

On May 28, 2025, he started making Curve Clash with AI, a game where each person steers a snake that leaves a trail, and whoever does not crash wins. By August, he had a test version. By mid-December, the game was out on the web, iPhone, and Android: less than seven months from start to launch, without quitting his job¹.

But the road was not straight. Bjerk chose a game for several people to play together online, and spent September and October stuck on bugs that made players' devices see different things¹. Getting the feel of the game right, like the turning speed and the thickness of the trail, took hours of testing. And the part he calls the hardest came after the game was done: getting people to know about it¹.

That is why his advice for beginners is not to start with a multiplayer game: first, make the game good for a single player¹. According to him, more than 1,500 people played Curve Clash. He has not disclosed how much the game earned¹.

And how long did famous games made before AI take?

Anywhere from a weekend to four years. Three warnings before the table: none of these games was made with AI; almost all of their creators already knew how to code; and these are the cases that worked out among thousands that almost nobody played.

Notice what separates the two ends. 2048 is a single screen with a single rule: slide the numbers to merge matching ones. Stardew Valley is a whole farm, with animals, a town, fishing, and caves to explore, and its creator did the programming, music, and art by himself¹⁰.

Notice, too, what the table does not say. "One weekend" is the first version of a small game, made by someone who already knew how to code. Even a game that started fast can take years to finish: Minecraft had its first public version in May 2009 and only reached version 1.0 in November 2011¹¹. These stories are told in famous games that started small.

What makes the time vary?

Five things, and four of them are in your hands. This list is our take, based on the cases above.

The size of the game

This weighs the most. Every extra level, character, button, and screen is one more thing to make, test, and fix. One screen with one action is one project; a world to explore is dozens of projects together. Bjerk puts it his own way: everything that goes into the game needs maintenance, bug fixing, and tuning¹.

The art

Simple art gets done sooner. AI helps make the game's art, but a detailed style takes many rounds before all the pieces match. Simple shapes shorten this part: Geometry Dash, according to its creator, started as a square that jumped and crashed⁶.

Changes of mind

This is the one nobody accounts for. Every big change redoes part of what was already finished. Changing the color of a button is quick; switching from a racing game to a fighting game halfway through is almost starting over.

Testing

Testing means playing your game many times, looking for what is broken or boring. AI helps find bugs, but it does not know whether the game is fun. Bjerk puts it this way: AI does not tell you if the game is fun, other people do¹. And outside players play when they can, on their own schedule.

Your hours per week

Bjerk made his game without quitting his full-time job as an accountant¹. Balatro started as a side project, and its creator only quit his job about a year before launch⁹. The same idea takes very different amounts of time for someone with two hours a week and someone with twenty.

How long does it take to get the game into a store?

On Google Play, Android's store, at least two weeks after the game is done, if your publisher account is personal and new. This is the part of the timeline that depends on neither you nor the AI.

Anyone who opened a personal developer account after November 13, 2023, has to run a closed test before publishing: at least 12 people you invite, with the game installed, for at least 14 days in a row². The days have to be consecutive to count². A closed test means only the people you invited can play.

After the test, you apply to the store for production access and answer questions about the test, the game, and whether it is ready. According to Google, this review usually takes seven days or less, but sometimes takes longer. And if the store thinks there was not enough testing, it can ask you to keep testing².

Added up, that is two weeks of testing plus the review time (our math). In our view, the best way not to lose time here is to invite the 12 testers while the game is still being made. The full walkthrough is in how to publish a game on Google Play.

What size of game fits your time?

The smallest one you still want to finish. There is no official table saying "a small game takes so many months," and making one up would be guessing. What you can do is compare game sizes with known cases and with what weighed most in each one.

In our view, a first game does well in the first two rows. The kinds of games that fit there are explained in what your first game should be.

Does AI make everything faster?

It speeds up writing the code, which was exactly where people who do not code used to get stuck. The rest still takes human time.

Bjerk described the split well. In the interview, his words were "AI was my engineering team, not my creative director"¹. He was the one who decided the game, tested the feel of every turn, and chased down players.

Some parts of the timeline AI does not shorten:

• Deciding. What goes in, what stays out, and when it is good.
• Testing with people. People play when they can.
• The store's test. Google Play's 14 days in a row are 14 days for everyone².
• Getting the game found. Bjerk says it was the hardest part of all¹.

The other side

None of the cases on this page works as a timeline for your game. Bjerk already knew some HTML, chose a specific game, and had the hours he had. The famous games in the table were made before AI, almost all by people who already coded, and they are exceptions among thousands that almost nobody played. And finishing is not selling: getting into a store is one step, and getting people to find the game is another, which can take longer than making it.

How do you avoid giving up halfway?

By choosing a game that ends. These suggestions are our take:

• Start with the smallest game that still excites you. Your dream game does not go in the trash; it goes on the top shelf, for when you have finished one or two small ones.
• Write down what stays out. A "later" list holds back the urge to add one more thing to the game.
• Decide in advance what "done" means. Without that line, the game never ends.
• Show it early. One person playing reveals in minutes what you did not see in weeks.
• Recruit testers before you need them. Google Play's 12 need 14 days in a row, and that clock only starts once everyone is in.

The first step fits on a sheet of paper: write your game idea in three sentences, before opening any program.

Your action plan

Check off what you've done. It's saved on this device, so you can pick up where you left off.

1. Write your game in three sentences: what the player does, how they win, and why they would play again.
2. Cut the idea down until it fits on one screen with one action, and keep the rest on a "later" list.
3. Write down how many hours a week you really have for the game.
4. Define, before you start, what needs to be working for the first version to count as done.
5. List 12 people with Android phones who would agree to test your game for 14 days in a row.
6. Read the official Google Play page about the closed test for new personal accounts.
7. Choose the type of your first game by reading what your first game should be.