How Long Does It Take to Make an Indie Game? Real Timelines

Stardew Valley was supposed to take six months. Eric Barone planned it as a resume-builder, something to show employers while he applied for programming jobs. He shipped it four and a half years later, after working on it solo for reportedly around 10 hours a day, seven days a week. If you’re scoping your first commercial game right now, that gap between the plan and the reality is the single most useful data point in this post.

There is no single average time to make an indie game, and I’d distrust anyone who quotes one. The “average dev time” stats floating around SEO content farms are unsourced. What developer postmortems and interviews actually support is a set of ranges by scope tier:

Why bother sorting projects into tiers? Because the market keeps getting more crowded. More than 19,000 games released on Steam in 2025 (Chris Zukowski’s January 2026 tally puts it at 20,282, a record), and roughly 48.6% of them ended the year with single-digit review counts. Your timeline decides two things at once: how much runway you burn, and how long your marketing window has to be.

A surprising number of famous indie games started as jam entries. Celeste began as a PICO-8 game that Maddy Thorson and Noel Berry built in three days in August 2015. Dome Keeper started as Dome Romantik, a 72-hour Ludum Dare 48 entry from April 2021 that placed 9th overall. Hollow Knight traces back to Hungry Knight, a Ludum Dare jam game from 2013.

The jam-to-commercial pipeline works because it answers the hardest question in game development before you invest years: is the core loop fun? A jam game that strangers play voluntarily, rate well, and ask for more of is a validated prototype. A design doc is a guess.

My take: if you’ve never shipped anything, your first project should be a jam game, not a commercial one. The weekend you spend is the cheapest scope test you will ever run.

This is the tier I push most first-time commercial devs toward. It’s long enough to build something sellable and short enough that your motivation, savings, and life circumstances will probably survive it.

Vampire Survivors took Luca Galante about one year from starting work in 2020 to its Early Access release on 17 December 2021. Dome Keeper went from jam prototype to full Steam launch in roughly 17 months. Both were small-scope games with tight loops, and both did enormous numbers relative to their dev time.

The trap at this tier is treating “small” as a quality ceiling. Vampire Survivors at Early Access was a small game by any production measure and became one of the biggest indie stories of the decade. Scope and ambition are different axes.

If you’re budgeting money alongside time, our breakdown of indie game development costs covers what projects at each tier tend to cost, and our roundup of famous indie game development budgets shows what 12 well-known games actually spent.

Most of the celebrated indie successes of the past decade live here.

Balatro took LocalThunk about two and a quarter years of solo work: he started in December 2021 and launched on 20 February 2024, and by his own timeline writeup he abandoned the project entirely for three months in the middle. Undertale took Toby Fox about 2.7 years from its mid-2013 Kickstarter to the September 2015 release. Celeste went from three-day prototype to full release on 25 January 2018, about two and a half years. Hollow Knight ran roughly three years (accounts conflict: the Kickstarter-to-release span is about 2.3 years, while some sources describe Team Cherry working full-time from 2013, closer to four).

Stardew Valley sits at the top of this band at four and a half years, and it only stayed there because Barone worked at a pace most people cannot and should not sustain.

What separates this tier from the one above it is usually content volume and polish depth, not mechanical complexity. Balatro’s core loop existed early; the years went into balance, feel, and the last stretch of refinement that makes a game seem inevitable in hindsight.

Animal Well took Billy Basso about seven years. He started in 2017 as a nights-and-weekends project alongside a day job and released on 9 May 2024. Along the way he cut at least 500 of the game’s planned encounters; 256 shipped. Cuphead also ran about seven years, from 2010 to its 2017 release. Hollow Knight: Silksong was announced on 14 February 2019 and released on 4 September 2025, roughly six and a half years from announcement to launch.

Here’s the thing about this tier: it can work. Silksong sold over 7 million copies by mid-December 2025. Animal Well and Cuphead are modern classics. A long timeline is not fatal if the game delivers.

But look at the funding structures. Basso kept a day job. Team Cherry had Hollow Knight revenue. If you have neither external income nor prior-game revenue, a five-plus-year project is a plan to run out of money. The GDC 2026 State of the Game Industry survey found 35% of developers are self-funded, and self-funding is exactly the model that long timelines punish hardest.

Two patterns are worth pulling out of that chart. First, the 2-to-3-year band is dense: Balatro, Celeste, Undertale, and Hollow Knight all shipped in that window, which is why I call 2 to 4 years the realistic mid-size range rather than an overrun. Second, the games at the long end were mostly not full-time for the whole span. Animal Well was nights and weekends for years. Calendar time and work hours are different measurements, and most famous-game timelines you read are calendar time.

Hofstadter’s Law says it always takes longer than you expect, even when you account for Hofstadter’s Law. Indie games are the purest demonstration of it I know.

Stardew Valley: planned at six months, shipped at four and a half years, a 9x overrun. Hollow Knight’s Kickstarter ran in November 2014 (it raised over A$57,000 from 2,158 backers) and the game shipped in February 2017, over two years later. Silksong’s announcement-to-release gap was about six and a half years. Even Balatro, a comparatively disciplined project, included a three-month stretch where LocalThunk walked away entirely.

The mechanism is simple: “making the game” is a minority of making the game. The best hard data I’ve found comes from Joseph Mirabello, who tracked every hour of Tower of Guns development with Toggl and published the totals: 3,850 hours over 600 days. Marketing alone consumed 983 hours, about 25% of the entire project. Add store page assets, trailer production, ports, playtesting, bug fixing, and business admin, and the feature work you mentally scheduled turns out to be half your hours or less.

Run this gut check on your own plan: take your feature-complete estimate, add marketing at roughly a quarter of total hours, then add the polish tail that turned Balatro’s working prototype into a two-plus-year project. If the result isn’t roughly double your original number, you’ve forgotten something.

This is the part most timeline articles skip, and for a commercial game it matters as much as the production schedule itself.

The standard guidance comes from Chris Zukowski: put your coming-soon page up at least 6 months before launch, ideally 12 or more, and post it as soon as you meet five criteria: genre locked, art style final, at least three unique environments to show, professional capsule art, and a gameplay trailer of at least 30 seconds. Our capsule design guide and trailer best practices cover the two criteria devs most often get wrong.

Valve points the same direction. The official Steamworks marketing docs tell developers to publish the coming-soon page “as soon as possible — as soon as you begin talking about your game publicly.” Data attributed to Valve (it circulates via summaries of Zukowski’s talks, not a primary Valve document, so hold it loosely) suggests pages live at least six months before launch saw roughly 300% more sales than pages that went up 30 days out.

What about announcing too early? Zukowski’s analysis of Cosmoteer and Song of Iron found conversion rates were statistically identical for wishlists added 4.5 years before launch and those added near launch. Wishlists don’t rot, so there’s no conversion penalty for an early page.

There is a nuance, though. GameDiscoverCo’s November 2025 wishlist conversion study found that games which exceeded sales expectations had an average pre-release page period of 214 days, while underperformers averaged 411:

I read that as correlation, not a command to announce late. Games that sit in Coming Soon for 411 days are often games that slipped or announced before they had anything compelling to show. The 214-day figure for hits lands almost exactly on Zukowski’s “at least 6 months” floor, which is a satisfying convergence from two independent data sets.

The practical rule, mapped onto your dev timeline: for a 6-to-18-month project, your page goes live near the midpoint, the moment you meet the five criteria. For a multi-year project, hold the page until you’re roughly 12 months out rather than announcing in year one; spend the early years building the game and a small community instead. Our coming-soon page guide walks through the setup itself, and how to get Steam wishlists covers what to do once the page is live. One benchmark before you announce: Zukowski’s medians for first-two-week wishlists run from about 100 (games that eventually gross under $10K) to about 7,000 (games that cross $1M).

Everything above compresses into a few planning rules I wish more first-time devs followed.

Pick a tier below your ambition. If you believe your game is an 18-month project, the record says it’s a 3-year project. If you can’t afford 3 years, design a 9-month game.

Ship a jam game first. Celeste, Hollow Knight, and Dome Keeper all grew out of jam prototypes that proved the loop before the multi-year investment.

Cut content, not quality. Animal Well shipped 256 encounters and cut at least 500 planned ones. Nobody reviews the encounters you didn’t ship.

Budget marketing hours like a feature. If Tower of Guns is representative, a quarter of your project hours go to marketing (the dollar side of that line item is covered in our marketing budget guide). Schedule them or they will eat your launch window instead. Steam Next Fest in particular needs a demo-quality build months before release, with submission deadlines that land roughly a month before each fest, so it has to live on the same timeline as your milestones.

Put a date on the page decision. The launch isn’t your only deadline. Work backward from your target release: page live 6 to 12 months out, demo and festival slots after that, launch beats in the final weeks. Our launch date planner maps those milestones against Steam’s event calendar.

How long did Stardew Valley take to make?

Four and a half years of solo development. Eric Barone originally planned it as a roughly six-month resume project, then reportedly worked on it around 10 hours a day, seven days a week, until it shipped.

How long did Hollow Knight take to make?

About three years, with a caveat: the Kickstarter ran in November 2014 and the game shipped on 24 February 2017 (about 2.3 years), while some accounts describe Team Cherry working full-time from 2013, which would put it closer to four.

How long does it take to make a game solo?

For a small commercial game, plan on 6 to 18 months of focused work; Vampire Survivors reached Early Access in about a year. Mid-size solo projects like Balatro (about 2.2 years) and Stardew Valley (4.5 years) define the band above that. Whatever you estimate, history says to double it.

When should my Steam page go live on my dev timeline?

At least 6 months before launch, ideally closer to 12, and only once you have a locked genre, final art style, three unique environments, professional capsule art, and a 30-second gameplay trailer. GameDiscoverCo’s data shows overperforming games averaged a 214-day pre-release window, so roughly 7 to 12 months is the target zone.

Before you commit to a timeline, sketch the whole arc: dev milestones, the date your page meets Zukowski’s five criteria, and a launch window that gives the page at least six months to gather wishlists. The launch date planner handles the calendar math, and the coming-soon page guide covers the page itself. The games in the chart above took anywhere from one year to seven; the ones that won planned for the timeline they actually had, not the one they hoped for.