A Market Map of Web Games Found 13 Categories. When You License HTML5 Games You Are Buying One of Them.

When you license HTML5 games you are buying one category out of thirteen, and the other twelve — hosting, analytics, ads, payments, localisation — land on you. That is not a metaphor. In June 2026 the web games industry got its first published market map, and the exercise of drawing it produced the most useful procurement document I have seen this year: a picture of how many separate things have to exist before a licensed catalogue turns into a product somebody plays. The catalogue is one box. Here is what the other twelve cost, which of them you can defer, and how to tell before you sign which boxes your supplier can actually close for you.

🗺️ Thirteen Categories, and the Catalogue Is One Box

On 11 June 2026, web games distributor Playgama published a market map of the web games ecosystem covering 120+ companies and services across 13 categories, as reported by PocketGamer.biz. Trade coverage described it as the first attempt anyone had made to lay the sector out on one page. The categories run: game developers (split into three subgroups), game engines, platforms, distributors, analytics, monetisation solutions, authorisation and localisation services, in-game payments infrastructure, in-game ad providers, and cloud hosting.

Playgama founder Dmitry Kachmar framed the reason for making it plainly: "The web games market is huge, but there's still no shared understanding of how the ecosystem fits together." That is a fair description of what buyers run into. The map's value to an operator is not the company names on it — those change. It is the shape. Thirteen boxes exist because thirteen distinct jobs have to get done, and a licence agreement, however good, covers a narrow slice of them.

Most portal budgets I have seen are built the other way round. There is a line for content, a line for development, and a rounding error labelled "infrastructure". The map says the content line is one-thirteenth of the surface area. The rounding error is where the year goes.

📊 The Market Size Estimates Are Nearly Three Times Apart

The same coverage put existing estimates of the global HTML5 and web games market at anywhere between $8 billion and $23.4 billion annually, with projected growth of 3–10% CAGR. That is close to a threefold spread on the size and a threefold spread on the rate. Both figures were reported alongside the map by GamesBeat and by PocketGamer.biz.

Do not treat that gap as sloppiness. It is definitional. Whether a WeChat mini game counts, whether an instant game embedded in a news site counts, whether you measure platform gross or developer net, whether ad-funded play in an app webview belongs in "web" at all — every one of those choices moves the number by billions. Nobody has settled them, which is exactly Kachmar's point about shared understanding.

The practical consequence for a buyer is small and worth stating anyway. When a supplier deck opens with a market size, ask which of the thirteen categories the number counts and whether it is gross or net. If the answer is vague, the number is decoration. Build the business case on your own traffic, your own fill rate and your own cost per acquired session instead. Those you can measure in week one; the TAM you cannot measure at all.

🏭 15,000 Games Shipped in One Quarter. Supply Is the Cheap Part.

The supply side is not tight. More than 15,000 new web games were released in Q2 2025 alone — a 2.7x increase year on year, according to Playgama research reported by PocketGamer.biz in July 2025. Whatever the market is worth, it is not short of titles.

That has two honest implications, and one of them is uncomfortable for anyone selling games.

• Generic content is not scarce, so it should not be priced as if it were. If a supplier wants a premium per title for undifferentiated casual content, the release volume above is your negotiating position. Use it.
• What is scarce is everything around the title. Clean chain of title. Consistent build packaging across a hundred games instead of a hundred one-off zip files. Metadata that matches from title to title. Territory and platform rights you can put in front of a client's legal team. A named person who answers when a build breaks on a Chrome release. None of that shows up in a release count, and all of it is what you are actually paying for.

So the question "where do I buy HTML5 games cheaply" has an obvious answer and a useless one. The question worth asking is which supplier closes the most boxes on the map at an acceptable price. That is a different shortlist.

🧩 The Twelve Boxes Nobody Puts in the Budget

Here is the map redrawn as a purchase order. For each one: what it is, and whether you can put it off.

You cannot defer these

• Cloud hosting and CDN. Someone serves the builds. If the licensor hosts, your origin is theirs and your caching, your service worker and your offline story are constrained by that. If you host, you inherit the bandwidth and the cache invalidation on every title update.
• Analytics. A licensed build reports plays. A business needs sessions per user, day-7 return, and revenue per session by title. That is your event pipeline to build, because the game will not send it.
• Ad monetisation. Ad network account, mediation or waterfall, consent management, ads.txt, invalid-traffic hygiene. This is three or four vendors before it is one line of revenue.
• Compliance. Age signals, privacy notices, regional content rules, accessibility obligations depending on your market. Not optional, not cheap to retrofit.
• Ops and release cadence. Browsers ship every few weeks. Titles rot. Somebody owns the regression sweep and the removal of a title whose licence expired.

You can usually defer these

• In-game payments infrastructure. Most licensed casual catalogues have no purchase surface inside the build at all. Charging happens at the portal level — a subscription, a credits balance, an ad-free tier — and that can wait until you have retention worth charging for.
• Authorisation and accounts. Anonymous play works fine for a first quarter. Accounts become urgent the moment you want cross-device progress or a paid tier, not before.
• Localisation beyond the shell. Translating your own navigation is a week. Translating text baked into a hundred game builds is a licensor conversation, and it belongs in the licence, not the sprint.
• Intrinsic in-game ad formats. Interesting, and irrelevant until you have inventory at scale.

Two boxes sit awkwardly between the lists: distribution and discovery. You can defer them in the sense that nothing breaks. You cannot defer them in the sense that a portal without a traffic plan is a hosting bill with a homepage. Decide the surface before the builds land.

🔧 Unity Is 55% of New Web Games. That Sets Your Hosting Budget, Not Theirs.

The same research run gave the engine mix for new web games in Q2 2025: Unity at 55%, Construct at roughly 16% and climbing from about 12% in early 2024, Cocos at 8.1%, Phaser at 7.1% and LayaAir at 5.3%, per the PocketGamer.biz report linked above.

That distribution is downstream of you and it lands on your infrastructure. A Unity WebGL export and a Phaser 2D build are different products from a delivery point of view: different payload sizes, different memory behaviour on a four-year-old Android handset, different first-frame timing over a mobile network, different SDK surfaces to integrate with an ad provider. If your catalogue skews to one engine, your hosting bill, your low-end device support burden and your loading experience all skew with it — and you did not choose it.

The fix is a column in the catalogue sheet. Ask any licensor for engine and compressed build size per title before you sign, not after. It costs them nothing to provide and it lets you model the two numbers that decide whether a portal feels fast: median payload and time to first interaction on your worst supported device. If a supplier cannot produce that column, they are handing you files rather than a supply agreement.

🧭 Three Scenarios, Mapped to the Boxes They Actually Need

The stack is not one size. Count the boxes for the job in front of you.

1. One branded game for a trade-show booth. Boxes needed: the game, a kiosk shell, offline delivery. That is close to three. No analytics pipeline, no mediation, no accounts. If the ask is genuinely one bespoke game with a custom mechanic and nothing else, a single studio commission is often the better route and I would say so — the catalogue advantage does not apply to a sample size of one. Where a catalogue does help here is choice and speed: picking a proven title and branding it beats briefing a build six weeks before the show.
2. 200 titles live on a carrier portal in six weeks. Boxes needed: nearly all of them. Content, hosting, analytics, monetisation or billing integration, localisation, compliance, distribution, plus an ops rhythm. This is the scenario where a file handover fails and where sourcing from four suppliers fails harder, because the integration work between them is yours and six weeks does not contain it.
3. An APK catalogue for an OEM preload. A different set again: store or preload compliance, an update pipeline, age ratings, signing and packaging. CDN barely matters; release engineering matters enormously. Buyers who assume the HTML5 stack transfers to Android find out in week three that it does not.

🚫 Five Ways Operators Under-Budget the Stack

• Budgeting content and calling the rest "integration". The twelve other categories exist because they are twelve jobs. Give each a line and an owner, even if the line is zero this quarter.
• Assuming hosting is included because a demo URL worked. A demo on the licensor's origin is not a hosting agreement. Get the origin, the caching rules and who pays for egress in writing.
• Buying breadth before deciding a traffic surface. A thousand titles and no acquisition plan produce the same revenue as forty titles and no acquisition plan.
• Treating metadata as an afterthought. Categories, tags, orientation, input method, supported languages, thumbnail sizes. Miss them and you rebuild the taxonomy by hand across the whole catalogue, once, badly.
• Signing before the compliance boxes are named. Age signals and privacy obligations are cheap at contract stage and expensive at launch minus two weeks.

📋 A Stack Audit to Run Before You Sign Anything

Nine questions. Any licensor worth signing with can answer them in one document.

1. Which formats are in scope — HTML5 builds, APK builds, or both under one agreement?
2. Where do the builds live: your origin or ours, and who pays egress?
3. What metadata ships with each title, and in what schema?
4. What engine and compressed size per title?
5. What does the build call out to at runtime, and can those calls be disabled?
6. Which territories and which platforms does the grant cover, enumerated?
7. Is source available for the titles where we need to change something?
8. Can our branding go on the shell, and on the titles themselves?
9. What is the release cadence for new titles, and who removes an expiring one?

Notice that eight of the nine are stack questions rather than price questions. That is the point. Price is easy to compare and rarely what goes wrong.

🎯 What You Get When You License HTML5 Games From a Direct Licensor

A licence from a direct licensor is a supply agreement, not a file drop. Forestry Games has operated since 2017 and licenses a catalogue of 1,049 titles covering HTML5 games and Android APK games, with titles also published on Google Play and the Apple App Store, and develops HTML5 games in house. In stack terms that closes more than the content box: HTML5 builds and APK builds settled in one conversation rather than two supplier relationships, source where applicable, your branding on the shell, and hosting either on our infrastructure or on yours.

Catalogue depth is what makes the sizing question answerable — a broad, light selection for an operator bundle, a narrower higher-quality set for a paid tier, both from one agreement. White-label portal delivery covers the shell, and distribution and monetization support means the boxes after launch are part of the conversation instead of your problem alone. To see the range first, browse the catalogue, or start from the HTML5 game licensing options — or the APK game licensing side if the preload scenario above is yours — and ask for a licence scope written against the boxes you actually need closed.

🧸 Licensing Branded Games for Campaigns, Portals and Events

Branded content is a separate conversation from a generic catalogue licence, and it adds boxes rather than removing them: approvals, compliance review and campaign windows all take calendar time. Forestry Games works with branded IP and has brand partnerships including Disney, Nickelodeon, Cartoon Network and Warner Bros. Businesses can license branded game content through Forestry Games for marketing campaigns, white-label portals, events and apps.

If branded titles are part of the plan, raise them at the start rather than once the launch date is fixed — the approval calendar, not the build, sets the schedule. Ask for a licence scope with the territories and campaign window named, or request a portal demo to see how branded and generic titles sit together in one storefront.

✅ Draw Your Own Map Before the Next Supplier Call

The market map is useful because it is uncomfortable. Thirteen categories, 120-plus companies, a market that credible sources size somewhere between $8 billion and $23.4 billion, and 15,000 new titles in a single quarter — that is an ecosystem where content is the abundant input and assembly is the scarce one. Nobody sells you the assembly on a per-title price list.

So do the cheap thing first. Draw the thirteen boxes on one page, mark the ones your project genuinely needs in the first six months, and write a name next to each: your team, a vendor, or your licensor. Then take that page into the next supplier conversation and ask which boxes they close. A supplier who can answer for content, formats, hosting, branding and territory in one document has closed five of them before you have hired anyone. If that is the shape of what you need, ask for a licence scope built around those boxes rather than a title count.