What 428 Game Outsourcing Studios Actually Look Like: A 2026 Directory Analysis
Most writing about game outsourcing describes a market nobody has counted. This post counts one.
In July 2026 we queried our own directory and pulled every listed outsourcing studio: 428 of them, across 53 countries, along with the services, engines, platforms and art styles each studio publishes about itself. What follows is what that census says. It answers questions that are otherwise guesswork. How many studios actually sell QA. Whether Unity or Unreal wins. How many can ship to Switch. Where in the world this work is concentrated, and how old the studios are in each market.
Methodology, and what this is not
The numbers below come from a single query against our production database on 26 July 2026. The filter was: active, listed, not deleted, type OUTSOURCING_COMPANY. That returned 428 studios.
- Country coverage: 53 countries, after normalizing
UKandEnglandintoUnited Kingdom. 59 of the 428 listings have no country recorded at all and are excluded from every country figure in this post. Country tables therefore draw on a smaller base than 428. - Tags are self-reported. Studios choose their own service, engine, platform and art-style tags when they list. Nobody audits them. A tag tells you what a studio sells, not how well it does it.
- Multi-select. A studio can carry several services, engines and platforms, so those columns sum to more than 428.
- This describes the directory, not the world. It is a census of who has listed with us, which skews toward studios doing outbound marketing in English. Treat it as a large sample with a known bias, not as the global market.
- There is no pricing in this study. Only 19 of the 428 studios have any budget field filled in, and those values are unusable (one range runs from 100 to 10,000,000). So this is an analysis of composition, not of price. Any hourly figure elsewhere on this site is an editorial market estimate, not something we measured here.
Finding 1: this is an art market with a programming market attached
| Service | Studios listing it |
|---|---|
| 3D art | 238 |
| Programming | 206 |
| Game design | 171 |
| 2D art | 140 |
| Full production | 114 |
| Animation | 96 |
| QA testing | 57 |
| Audio | 49 |
| Marketing | 35 |
| Localization | 32 |
| UI/UX | 25 |
| Narrative | 21 |
| Porting | 10 |
| Co-development | 10 |
3D art at 238 and programming at 206 are the two anchors of the market. Then the numbers fall off a cliff. QA testing sits at 57, audio at 49, localization at 32.
Buyer takeaway: you are shopping in two different markets depending on what you need. Ask for 3D art and you have hundreds of candidates, genuine price competition, and the freedom to walk away from a studio that gets slow. Ask for QA, audio or localization and your shortlist is short before you have filtered for time zone, platform experience or budget. Search for the scarce discipline first and build the rest of the package around whoever you find.
One important caveat on the bottom two rows. Porting and co-development each show only 10 studios, and that badly understates real capability. Many of the 114 full-production studios do porting and co-development routinely without tagging either service. Read those tags as "sells this as a headline offer," not as "can do this."
Finding 2: Unity and Unreal are nearly tied nationally, and not tied locally
| Engine | Studios listing it |
|---|---|
| Unity | 280 |
| Unreal Engine | 260 |
| Custom / in-house | 113 |
| Other | 40 |
| Godot | 21 |
| CryEngine | 3 |
| GameMaker | 3 |
| Cocos2d | 2 |
Twenty studios separate Unity from Unreal across the whole directory. That near-tie disappears the moment you look at one country at a time.
| Country | Studios | Unity | Unreal | Custom |
|---|---|---|---|---|
| United States | 60 | 42 | 43 | 27 |
| United Kingdom | 35 | 28 | 28 | 19 |
| Ukraine | 28 | 26 | 24 | 3 |
| India | 25 | 23 | 23 | 5 |
| Japan | 24 | 7 | 5 | 4 |
| Poland | 17 | 12 | 16 | 5 |
| Canada | 17 | 15 | 12 | 5 |
| Germany | 14 | 8 | 7 | 6 |
| China | 10 | 9 | 9 | 1 |
| France | 10 | 6 | 8 | 2 |
Three patterns show up here.
Poland's cohort is Unreal-first: 16 of its 17 studios list Unreal against 12 for Unity. Italy is the same shape at smaller scale, with all 5 of its studios listing Unreal against 2 for Unity. Ukraine goes the other way, 26 Unity against 24 Unreal.
Custom-engine experience is concentrated in the expensive markets. The United States has 27 studios listing a custom engine, the United Kingdom 19, Germany 6 of only 14 studios. Ukraine has 3 out of 28 and China 1 out of 10. If your project lives inside proprietary tech and you need people who have worked outside a commercial engine before, that is a real geographic constraint on your search.
Japan is a tagging gap, not a finding. Its 24 studios produce only 7 Unity and 5 Unreal tags. That does not mean Japanese studios avoid commercial engines. It means engine tagging is sparse in that cohort, and you should ask rather than filter.
Buyer takeaway: do not pick a country and then hope the engine fit works out. Filter by engine first, country second. You can do that directly on the Unity and Unreal Engine pages.
Finding 3: PlayStation and Xbox are near-universal, Switch is not
| Platform | Studios listing it |
|---|---|
| PC | 308 |
| Android | 236 |
| iOS | 233 |
| PlayStation | 218 |
| Xbox | 218 |
| Nintendo Switch | 130 |
| VR | 55 |
| Web | 43 |
| AR | 27 |
PlayStation and Xbox land on exactly the same number, 218. Switch drops to 130. That is the single most actionable gap in the platform data: a studio that ships to two of the three big consoles is very likely to ship to the third, unless the third is Nintendo.
VR at 55, Web at 43 and AR at 27 are specialist pools. Those are not markets where you compare six quotes.
Buyer takeaway: if Switch is in your release plan, filter for it at the start of the search, not at contract stage. The same applies to VR and AR, where the pool is small enough that availability, not price, will decide who you get. Platform filters live on the console and VR/AR pages.
Finding 4: stylized outnumbers realistic
| Art style | Studios listing it |
|---|---|
| Stylized | 205 |
| Realistic | 151 |
| Cartoon | 75 |
| Anime | 19 |
| Pixel art | 16 |
| Low poly | 15 |
| Hand-drawn | 10 |
| Abstract | 6 |
| Isometric | 4 |
Stylized is the most claimed style in the directory at 205 studios, ahead of realistic at 151. It is also the least specific word in an art brief. Two studios both tagged stylized can be producing work with nothing in common.
The specific styles are the small pools: anime 19, pixel art 16, low poly 15, hand-drawn 10.
Buyer takeaway: "stylized" is not a brief. If your look is anime, pixel or hand-drawn, use the style filter as the entry point to your search, because those cohorts are small enough that a portfolio match matters more than anything else on your checklist. And when you write the brief, lead with reference images rather than adjectives.
Finding 5: studio age depends almost entirely on the country
| Country | Studios | Median founding year |
|---|---|---|
| South Korea | 4 | 2000 |
| Japan | 24 | 2003 |
| Vietnam | 3 | 2003 |
| China | 10 | 2006 |
| Sweden | 5 | 2007 |
| Germany | 14 | 2008 |
| Croatia | 3 | 2008 |
| Argentina | 4 | 2009 |
| Serbia | 7 | 2010 |
| Canada | 17 | 2012 |
| United States | 60 | 2014 |
| Ukraine | 28 | 2014 |
| United Kingdom | 35 | 2015 |
| India | 25 | 2015 |
| Poland | 17 | 2016 |
| Spain | 9 | 2016 |
| Turkey | 5 | 2017 |
| Israel | 3 | 2019 |
| Saudi Arabia | 6 | 2020 |
The oldest cohorts are in East Asia. Japan's median founding year is 2003, South Korea's is 2000, Vietnam's is 2003. The newest are in the Gulf and the eastern Mediterranean: Saudi Arabia 2020, Israel 2019, Turkey 2017.
Read the small-sample rows with care. South Korea, Vietnam, Croatia and Israel each have 3 or 4 studios, so a single listing moves the median by years.
Buyer takeaway: age is a proxy for one specific thing, which is how many console generations and platform cert cycles a studio has lived through. It is not a proxy for quality. A 2020-founded Saudi studio and a 2003-founded Japanese studio are different kinds of bet, and the question that separates them is not "how long have you existed" but "show me the titles you shipped, with dates and platforms."
Finding 6: 172 of 428 studios sit in five countries
| Country | Studios listed |
|---|---|
| United States | 60 |
| United Kingdom | 35 |
| Ukraine | 28 |
| India | 25 |
| Japan | 24 |
Those five countries hold 172 of the 428 listings, roughly 40% of the directory. Of the 53 countries represented, 20 have five or more studios and 32 have three or more. The rest are one- and two-studio markets.
The service mixes inside the top five are not interchangeable. Ukraine's top three are 3D art (21), 2D art (19) and programming (16): an art-heavy cohort. The United Kingdom's are programming (16), game design (16) and 3D art (12): a design- and code-heavy one. Japan's are 3D art (17), audio (5) and programming (4), and Japan is the only country in the whole table where audio appears in the top three services.
Buyer takeaway: concentration is good news for buyers of art and code and bad news for everyone else. If you need a discipline that does not appear in any country's top three, you cannot shop by geography at all. You have to shop by service tag and take the time zone you are given.
What this census cannot tell you
Being clear about the limits is the point of publishing the method.
- Tags are unaudited self-reports. A studio listing three services may be strong at one of them.
- 59 listings have no country. Every country figure above excludes them, so the country base is smaller than 428.
- Porting and co-development are undercounted. Ten studios each is a count of who markets the service, not who can perform it.
- No usable price data. 19 studios of 428 have any budget field set, with values that make no sense. We have deliberately published nothing about rates here.
- Team size is unusable too. It is stored as free text with overlapping buckets, with
11-50sitting alongside50-100and51-200, so headcount is left out of this analysis entirely. - Listing is voluntary. Studios that do no English-language outbound marketing are underrepresented, which almost certainly affects the country distribution more than any other number here.
How to use this if you are buying
Four things follow from the data if you are about to run a search.
Search scarce to abundant. Lock in the discipline with the smallest pool first. If your project needs 3D art plus localization, the localization studios decide your shortlist and the art studios do not.
Filter by engine and platform before geography. Country tells you about time zone and rate band. Engine and platform tell you whether the studio can do the job at all.
Treat "full production" as a capability signal. 114 studios list it. Those are the teams most likely to cover a scarce discipline in-house rather than subcontracting it, which matters when the scarce discipline is the one holding up your schedule.
Ask for shipped titles with dates and platforms. It is the only cheap check on a self-reported tag, and this whole analysis is built on self-reported tags.
If you want to run these filters yourself, browse the outsourcing directory and stack service, engine, platform and style filters until the list is short enough to email. For a shortlist that has already been narrowed, see best game outsourcing companies. And for the pricing side, which this study deliberately does not cover, our complete guide to game outsourcing costs works through market rate ranges and worked project budgets.
Data: gamedevoutsourcing.com directory census, 26 July 2026. n = 428 listed outsourcing studios across 53 countries. Self-reported tags. Browse the full directory and contact studios directly with no platform fees.
