Mobile Game Development Cost: What It Actually Takes to Ship

The honest bands: what a mobile game costs in 2026

Across 2,000+ delivered projects at Digital Heroes, mobile games cluster into three shapes. The gap between them is not "quality" in the abstract. It is content volume, whether there is a server, and whether the game is designed to be operated after launch or just shipped and forgotten.

Tier 1: focused single-mechanic game. $25,000 to $70,000. 3 to 4 months. One core loop, hyper-casual or casual puzzle, 20 to 60 levels or procedurally generated content, 2D art built on a purchased and lightly customized asset base, local save only, one ad network plus a single remove-ads purchase. Team is one Unity or Godot developer, a part-time 2D artist, a designer at 30 percent allocation, QA at 20 percent. Both iOS and Android come from one Unity codebase.

What falls out of scope at this tier, and you must accept it going in: no accounts, no cloud save, no leaderboards beyond the Game Center and Google Play defaults, no server-authoritative economy, no live events, no A/B testing, no localization past English, no custom character animation, no soft launch across geos. If a Tier 1 quote includes those things, someone is going to lose money on the project and it will eventually be you.

Tier 2: mid-core game with progression and an economy. $70,000 to $180,000. 6 to 9 months. A meta layer sits on top of the core loop, meaning currencies, upgrades, a shop, daily rewards, and a tuned difficulty curve. 80 to 200 handcrafted levels, or a content system that generates them. Custom 2D or stylized low-poly 3D art with rigged characters. A backend for accounts, cloud save, validated purchase receipts, and remote config so you can change balance without shipping a build. Analytics wired to events that matter, not just installs. Team is two developers, a backend developer at 50 percent, one artist plus one animator, a game designer, QA at 50 percent, a producer at 25 percent.

Tier 3: multiplayer or live-ops title. $180,000 to $400,000+. 10 to 18 months. Real-time or asynchronous multiplayer, matchmaking, server-authoritative state to stop cheating, seasons, a battle pass, live-ops tooling so a non-engineer can launch an event, and a content pipeline that keeps producing after launch. Five to nine people including a dedicated backend pair, a technical artist, and a live-ops producer. This tier does not end at launch. It starts there.

What actually drives the number

Six drivers, each with a price attached.

1. Art volume and style. 30 to 55 percent of total build cost. This is the single largest line and clients consistently underestimate it. Flat 2D vector art runs $150 to $400 per screen or asset set. Stylized 3D with rigging and animation runs $1,200 to $4,000 per character depending on how many animation states it needs. A game with 6 playable characters, each needing idle, run, three attacks, hit and death, is $15,000 to $25,000 in character art alone before you have drawn a single environment. Switching from 2D to 3D mid-project typically adds 35 to 50 percent to the whole budget, because it drags in a technical artist, level-of-detail work, and a device performance pass.

2. Server authority and multiplayer. $30,000 to $90,000 on top. A game where the client can be trusted costs almost nothing on the server side. A game where it cannot be trusted needs authoritative simulation, state reconciliation, matchmaking and anti-cheat. Real-time multiplayer with rollback is the expensive end. Asynchronous multiplayer, where you fight a snapshot of another player's team, gets you 80 percent of the social feel for roughly $25,000 to $40,000. Most clients should build asynchronous first.

3. Level and content count. $250 to $900 per handcrafted level. That covers design, build, tune, test. 150 levels is $37,500 to $135,000, and it is a real line item, not a rounding error. The alternative is a procedural or template-driven system, which costs $12,000 to $30,000 up front and then makes level 400 nearly free. The crossover point sits around level 60 to 80. Below that, handcraft. Above it, build the system.

4. Monetization depth. $8,000 to $35,000. A single remove-ads purchase is roughly $3,000 to $6,000 including store setup and receipt validation. A full economy with consumables, a shop with rotating offers, a battle pass, rewarded video placements, and server-side receipt verification against both Apple and Google is $25,000 to $35,000. Add about $6,000 if you need subscriptions, because renewal, grace period and refund states each need real logic and real testing.

5. Live-ops tooling. $20,000 to $60,000. Remote config, an event scheduler, a player admin panel, segmentation, and A/B testing. Skipping this is defensible for Tier 1. For any game you plan to run for two years, skipping it means every balance change becomes an app store submission, and you will pay for that in engineering hours and lost tuning speed within six months.

6. Platform breadth and device support. 10 to 20 percent. One Unity codebase to iOS and Android is already baked into every band above. Adding a genuinely low-end Android floor, meaning 2 GB RAM devices, adds 8 to 15 percent for the optimization and QA passes. Adding a web or Steam build adds 15 to 25 percent, because input, resolution and store integration are all separate work. Console is a different project, not an add-on.

Worked example: a mid-core squad battler

Asynchronous PvP squad battler. 8 collectible characters, upgrade progression, an energy economy, a shop, raids against snapshots of other players' teams, iOS and Android. This is a shape we quote often. Here is the arithmetic.

• Discovery, game design document, economy model, playable prototype of the core loop: $14,000
• Core combat engine, turn resolution, ability system, deterministic replay: $26,000
• Meta layer: progression, upgrade trees, energy, currencies, shop UI, daily rewards: $22,000
• Backend: accounts, cloud save, snapshot matchmaking, leaderboards, remote config: $31,000
• Character art and animation, 8 characters at roughly $2,200 each: $17,600
• Environments, visual effects, UI art kit, icons: $16,000
• Monetization: purchase catalog, server receipt validation on both stores, rewarded video, battle pass: $18,000
• Analytics and event instrumentation: $5,400
• QA across a 14 device matrix, balance passes, soft launch support: $16,000
• Store submission, store listing assets, store video, launch build hardening: $6,000
• Producer and project management across 7 months: $12,000

Total: $184,000. Add a 12 percent contingency and you brief the board at $206,000. That contingency is not padding. In games it gets spent on whatever soft launch teaches you, and if it does not get spent, you underused it.

Want the same game at $120,000? Cut to 5 characters (saves $6,600 in character art), push the battle pass to phase two (roughly $8,000 of the monetization line), use a bought environment art pack (saves about $9,000), trim the backend to accounts and cloud save without live remote config (roughly $11,000), and ship with a 6 device QA matrix (saves about $6,000). That lands near $143,000, and dropping the deterministic replay system plus two characters more gets you to $120,000. It is still a real game. It is just a smaller one, honestly scoped, with the cuts written down instead of discovered in month five.

The ongoing costs nobody puts in the quote

Building the game is a project. Running it is a business line. Budget both.

Store fees. The Apple Developer Program is $99 per year at published list price. Google Play developer registration is a one-time $25, also published. Both stores take 30 percent of revenue, dropping to 15 percent under their small business programs below $1 million in annual proceeds.

Backend hosting. A Tier 2 game with 20,000 monthly players runs $250 to $900 a month on a managed platform. A Tier 3 real-time multiplayer game at 100,000 monthly players runs $2,500 to $8,000 a month, because you are paying for concurrent connections and low latency across regions, not just storage.

Third-party services. Analytics, crash reporting, remote config and push together land at $150 to $700 a month at mid scale. Most have free tiers you will outgrow exactly when the game starts working, which is the worst possible time to be surprised by an invoice.

Maintenance: 15 to 20 percent of build cost per year. On the $184,000 example, that is $27,600 to $36,800 annually. It buys operating system compatibility, because iOS and Android both ship breaking changes every year, plus software development kit updates, store policy compliance, crash fixes, and support for hardware that did not exist when you launched. This is not optional. A game left untouched for 18 months often fails to build against current store requirements at all.

Year one changes the business will ask for. Budget 25 to 40 percent of build cost. Every game ships and immediately reveals what players actually do. New content, an economy retune, a rebuilt tutorial because a large share of installs quit before level 3, a new offer type. On the $184,000 example that is $46,000 to $73,600, and it is the money that decides whether the game earns anything at all. Vendors rarely raise it. Plan for it anyway.

How to not get burned on price

The cheapest quote is usually the most expensive outcome. A $40,000 bid on a $150,000 game is not a discount, it is a different and smaller scope that nobody wrote down. The vendor books it, burns the fee by month three, and every gap becomes a change request billed at a rate you never negotiated, because you have no leverage once your art and code live on their machines. In games this bites harder than in apps, since game scope is genuinely fuzzy until the loop is fun, and a team with no margin left will ship the loop that is cheapest, not the one that works.

Ask every bidder to price the same written scope. If one number lands 60 percent below the others, the explanation is almost never efficiency. Make them show the line items.

What a change request should cost. A change request is scoped, quoted in hours at your contracted rate, and approved in writing before work starts. A new character with a full animation set on an existing rig pipeline is 20 to 40 hours. A new game mode reusing the combat engine is 80 to 200 hours. If a vendor quotes changes in vague dollar lumps rather than hours times rate, you cannot audit it, which is exactly why it is quoted that way.

Contract terms that protect the number. Fixed scope with a named deliverables list and an explicit exclusions list, because the exclusions list is what ends the argument at month five. Intellectual property assignment on payment covering source code, art and audio, so a payment dispute never turns into a hostage situation. Source in your repository from day one, with commits landing continuously, not a zip file handed over at the end. Third-party asset licenses documented and transferable, because a purchased art pack with a non-transferable license is a legal problem you inherit silently. Payments tied to milestones you can inspect, meaning a playable build, not a status report.

How to brief a vendor so the quotes are comparable

Most quote spread is not vendor pricing. It is that every vendor guessed at a different game. Fix the brief and the numbers converge to within about 20 percent, which is the range where you can actually make a decision.

Send all bidders the same document. Name the genre and two reference games, and say specifically what you want from each, for example the combat feel of one and the meta progression of the other. State the core loop in three sentences. Give a number for content at launch, meaning levels, characters or maps, and say whether it is handcrafted or systemic. Specify art style with three reference images and say whether 2D or 3D is fixed or open. State the monetization model explicitly: ads, in-app purchases, both, or premium. Say whether multiplayer is real-time, asynchronous, or none, and if you genuinely do not know, say so and ask each vendor to price both. List the minimum device spec and the target frame rate. Name the platforms and the launch geos. State whether live-ops tooling is version one or version two. Give your target launch window and your budget band. Withholding the budget does not get you a better price, it gets you five quotes for five different games.

Then ask every vendor for the same four things: a line-item breakdown, team composition with named roles and allocation percentages, their hourly rate for change requests, and an explicit written list of what is excluded. That exclusions list will tell you more about a vendor than the total ever will.