How long does it take one person to make a game?

Classic solo games took years: Stardew Valley about 4.5, Animal Well about 7, Axiom Verge 5. With AI writing the code, a solo creator can now finish an arcade-scope game in a day — the years were mostly implementation time, which is what AI removed.

Can I make a game in one day?

Yes — an arcade-scope browser game via an AI builder takes minutes to generate and a few revision rounds to make genuinely fun. Game jams have proven the one-weekend game for decades; AI compresses it further.

How Long Does It Take to Make a Video Game? (2026 Numbers)

Q: How long does it take to make a video game?

It depends entirely on scope: AAA titles typically take 3–7 years with hundreds of people; commercial indie games 1–5 years; a solo hobby game weeks to months; and an AI-built arcade game about two minutes plus an afternoon of polish.

Anywhere from two minutes to seven years — and the difference is almost entirely one variable. Here are the honest 2026 numbers at every scale.

The short answer, by scale

Scale	Typical timeline	Team
AAA blockbuster	3–7 years	Hundreds to thousands
Commercial indie	1–5 years	1–20 people
Solo passion project (classic)	2–7 years	One stubborn person
Game-jam game	48–72 hours	1–4 people
AI-built arcade game (2026)	~2 minutes + an afternoon of polish	You + a model

The variable is scope — hours of content, fidelity of art, and how much of the work is implementation vs. design. Nothing else comes close.

Why the classic numbers are so big

The famous solo timelines — Stardew Valley ~4.5 years, Axiom Verge 5, Animal Well ~7 — hide a breakdown worth staring at: the majority of those years went to implementation: writing engine code, drawing every sprite, composing, debugging. Design — deciding what the game is and tuning what's fun — was the smaller slice. It just came bundled with the years of typing (we profiled all of them in games made by one person).

AAA is big for a different reason: content volume at high fidelity has no shortcut. A 60-hour photoreal world is thousands of person-years of asset work, and that's why studio timelines aren't collapsing the way hobby timelines are.

What AI actually changed

Text-to-game models attack exactly the implementation slice. On Arcade Sandbox, a one-sentence idea becomes a complete, playable browser game — loop, physics, input, scoring, sound — in about two minutes (here's how that works). What's left for you is the part the classic devs spent the minority of their time on but that made their games great: playing it and deciding what's wrong. A realistic AI-era timeline for a game people replay:

Minute 0–2: first playable build from your prompt.
Hour 1: three or four plain-English revisions — feel, difficulty, juice.
Day 1: friends playtest via a link; one more revision from what you watched; publish.
Week 1: read your plays and unlocks; build the sequel to what worked.

▶ TWO MINUTES, TIMED

Test the claim yourself — first game's free. New accounts start with ⚡ 5 tokens; a 2D build costs 4.

Start the clock →

2D builds 4 tokens · 3D builds 8 · revisions 3 · publishing free.

How long should YOUR game take?

Match the timeline to the goal, not the ego:

"I want to see my idea exist" → minutes. There is no longer any reason this takes longer (weekend plan here).
"I want strangers to play something of mine" → a weekend, including polish and publishing.
"I want to sell a premium game on Steam" → 1–3 years and a real engine; AI-built reps first will make that year far more likely to succeed (the career math).
"I want to make the next Hollow Knight" → years, by design. Scope is the point of such games; go in eyes-open.

The uncomfortable truth in the numbers

Most "how long does it take" searches are really asking "how long until I'm allowed to start?" The answer used to be "after months of tutorials." Now it's now. The two-minute game isn't the destination — it's the rep that starts the compounding, and people who start compounding today are years ahead of people still comparing engines.