Your Licence Survives the Licensor's Collapse. Your Hosted Catalogue Doesn't.

When you license HTML5 games, the contract usually survives a licensor's insolvency. The files, the hosting and the update pipeline are the parts that don't.

This is the failure almost nobody models. A portal operator runs a quality pass on the catalogue, argues about territory for a fortnight, negotiates the revenue share down two points, signs, launches. Two years later the licensor is gone — absorbed into an acquirer and quietly folded, wound down, or just no longer answering email. The operator's first instinct is to ask a lawyer whether the licence still stands. In most cases the answer is yes. It is also, operationally, beside the point.

Supplier stress in this industry is not hypothetical. GDC's 2026 State of the Game Industry report, drawn from more than 2,300 industry professionals, found that 28% had been laid off in the previous two years, rising to 33% in the United States, and that half said their current or most recent employer had run layoffs in the previous twelve months. That measures job losses, not company failures — I could not find a credible count of games-company closures over the same window, so read it as a signal about sector conditions, not a supplier failure rate. But if you are signing a three-year catalogue term, "will this counterparty still be operating and staffed in year three" deserves a clause, not a hope.

⚖️ Rejection Is a Breach, Not a Rescission

Start with the good news, because it is better than most buyers assume. If a US licensor enters bankruptcy and its trustee rejects your licence agreement, that rejection does not cancel your rights.

Two mechanisms do the work. The first is 11 U.S.C. § 365(n), which lets a licensee of intellectual property elect to retain its rights for the balance of the term — including renewal periods it can extend — instead of treating the contract as terminated. The trade is that the licensee keeps paying royalties and waives setoff rights and certain damage claims against the estate. The second is Mission Product Holdings v. Tempnology, decided 8–1 in May 2019, in which the Supreme Court held that a rejection "breaches a contract but does not rescind it", so the rights the licensee already received stay in place and the licensee is left with a damages claim on top.

Tempnology mattered because § 365(n) leans on a closed definition of intellectual property at § 101(35A): trade secrets; inventions, processes, designs and plants protected under title 35; patent applications; plant varieties; works of authorship protected under title 17; and mask works. Game code and game art are works of authorship, so a generic catalogue was always inside that fence. Trademarks are conspicuously absent from the list — which is exactly why the branded side of a games licence needed a Supreme Court case before anyone could be confident about it.

Two honest caveats. This is United States insolvency law; a supplier incorporated in the UK, the EU, India or Singapore sits under a different regime with different answers, so ask counsel in the licensor's jurisdiction, not yours. And none of it compels performance. Section 365(n) preserves your right to keep using what you already have. It does not oblige an estate to host anything, patch anything, deliver the next batch, or answer a support ticket.

Which is the whole point. The legal layer is the survivable one. The operational layer is where portals go dark.

🔌 The Real Exposure Is "Where Do the Files Actually Live?"

Every catalogue deal lands in one of three delivery shapes, and they have completely different survival characteristics. Most buyers do not know which one they bought until they need to.

1. The licensor hosts, you embed

You received a list of iframe URLs. Your CMS stores links; your servers store nothing. This is the fastest integration on the market and the most fragile asset you can own. The games run on the licensor's origin, behind its CDN contract, resolved by its DNS, paid for by its card on file. When the company stops paying its infrastructure bills — which happens weeks before any formal insolvency, and months before anyone tells you — your entire catalogue 404s at once. Your § 365(n) election is a piece of paper describing games you cannot serve.

2. The licensor delivers builds, you host

You received archives. Each title is a folder of HTML, JavaScript, assets and a manifest sitting on your own origin or your own CDN. If the licensor disappears tomorrow, nothing changes for your players today. You stop receiving new titles and you stop receiving fixes, which matters — but the catalogue you have keeps earning. This is the shape you want for anything load-bearing.

3. You host the build, and the build phones home

The dangerous middle. The files are yours, so the deal looks like shape 2 on paper, but at runtime each title fetches an ad configuration, a licence check, an analytics endpoint, a remote-config JSON or a shared SDK from a domain the licensor controls. Sometimes the call is optional and the game degrades. Sometimes it is blocking, and the loading bar sits at 90% forever. You will not find out from the contract, because the contract does not describe network behaviour.

🧪 The One-Hour Test That Tells You Where You Stand

This is cheap enough that there is no excuse for not having run it.

1. List every third-party domain your game pages contact. Load twenty titles on staging with the browser's network panel open, or pull a week of egress logs, and collect the hostnames that are not yours.
2. Separate them into licensor-controlled, ad-stack, and everything else. The ad stack is a different continuity problem with its own answer.
3. Null-route the licensor-controlled domains at your edge or in your staging hosts file. Not blocked at the browser — blocked at the network, the way a lapsed domain registration behaves.
4. Play the same twenty titles. Record which load, which load without sound or without saves, and which never reach a first frame.
5. Extrapolate honestly. If four of twenty die, roughly a fifth of your catalogue is a hostage.

Run this before renewal, not after a scare. The result is also the single most persuasive thing you can put in front of a licensor when you ask for delivered builds instead of embed links: not a legal argument, a list of titles that stop working.

📦 Escrow Sounds Like the Answer. Read the Trigger.

Source-code escrow is the reflexive suggestion here, and it is a reasonable instrument that is routinely bought in a form that will not help.

The standard release conditions — vendor bankruptcy or insolvency, ceasing to trade, failure to maintain, material SLA breach — sound comprehensive until you notice how much interpretation each one invites. Vague triggers stall release exactly when speed matters, and the insolvency trigger is the weakest of them precisely because an insolvency process is the thing most likely to sit between you and the release. Counsel at Lowenstein Sandler, writing on escrow in the cloud era, make the sharper point: for hosted software, "the actual, practical transitioning of the application, data center and hosting environment in the event of a release condition may be more catastrophic than the downtime itself." A games catalogue served from someone else's CDN is exactly that kind of hosted software.

Three further limits worth knowing before you pay for escrow on a catalogue:

• Escrow delivers code, not rights. Getting a zip out of an agent does not grant you a licence to publish what is in it, and it does not touch any third-party brand rights layered on top.
• A thousand titles is a documentation problem. Escrow only functions if the package includes build instructions specific enough for someone else to rebuild. For a large catalogue assembled over years from multiple engines and studios, that documentation frequently does not exist internally, let alone in escrow.
• Third-party dependencies may not be escrowable at all. A build that relies on a licensed middleware SDK, a font, or a hosted service is not something the licensor necessarily has the right to deposit.

Escrow is worth buying for a small number of titles you genuinely could not replace — a bespoke branded build, a title carrying your own IP. For a broad arcade catalogue, the cheaper and more effective control is insisting on delivered, self-hostable builds from the start. If you are buying at the source-code level rather than as a licensee, you are already most of the way there.

🏷️ A Branded Title Has Two Contracts Above It

Everything above concerns a straight two-party licence. Branded content adds a layer that behaves differently and is easy to miss.

When you license a game featuring a licensed character, your agreement is with the games licensor — but the games licensor's right to grant it comes from a separate agreement with the brand owner. You hold a sublicence. Sublicences generally cannot outlive the head licence they hang from, and brand owners routinely reserve termination rights that insolvency, change of control or reputational events can trigger. So the reassuring US position on rejection can hold for your direct contract while the upstream right that made the deal possible falls away independently.

Three questions to put in writing before signing anything with a character on it: does my sublicence survive termination of the head licence, and for how long; if the head licence ends, do I get a sell-off or run-off period to wind the title down rather than pulling it the same day; and who holds the approvals file, because if the licensor goes quiet and the brand needs an approval you cannot evidence, the title comes down regardless of what your contract says. The commercial mechanics of these upstream deals are covered in more depth on our licence and pricing page.

📄 Six Clauses That Decide Whether You Survive Your Supplier

1. Delivery, defined as files. The licensor delivers self-hostable builds within a stated number of days of execution — not links, not access to a player. If the commercial deal is hosted, negotiate the right to take delivery of builds on notice.
2. Assignment and change of control. Say what happens if the licensor is acquired. At minimum, you want notice; ideally, a right to terminate and a right to continue on identical terms, at your election rather than theirs.
3. Survival on licensor default. Spell out that your right to continue exhibiting delivered titles for the balance of the term survives the licensor's insolvency, dissolution or cessation of business. It should not require you to litigate § 365(n) to find out.
4. A self-host step-in right. If the arrangement is hosted, the contract should grant you the right — and the technical means — to move the titles onto your own infrastructure on defined triggers.
5. Sublicence survival and run-off for branded content. As above, in writing, per brand.
6. Notice and transition period. A minimum notice period before any service withdrawal, with an obligation to hand over builds, asset manifests and per-title metadata during it. Metadata matters more than people expect: rebuilding category, thumbnail, orientation and age data for a thousand titles by hand is a quarter of someone's year.

🚫 Five Ways Buyers Get Continuity Wrong

• Treating the legal question as the whole question. Confirming the licence survives and stopping there. The licence surviving is necessary and nowhere near sufficient.
• Buying embed links because integration week was easier. A two-day integration that creates a permanent single point of failure is not a saving, it is a deferred cost with interest.
• Assuming a big supplier is a safe supplier. Size correlates with staying power, not with your specific product line staying open. Healthy companies discontinue catalogues during reorganisations, and the effect on you is identical.
• Never testing the dependency. The one-hour test above is not run at a majority of the portals I have seen, and the first time anyone learns which titles phone home is during an outage.
• Ignoring the metadata. A catalogue is also a database. Ask for the metadata export as a scheduled deliverable, not on request.

🎯 What a Direct Licensor Hands Over When You License HTML5 Games

Continuity is mostly decided by what physically arrives at signature, which is why it is worth being specific about how licensing works when you buy direct rather than through a feed or a reseller. 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.

A licence conversation covers scope and territory, what formats you receive, and how the titles are delivered — HTML5 builds you can host on your own infrastructure, APK builds for Android distribution, source where applicable, branding, and hosting options if you would rather not run the serving yourself. Because HTML5 and Android sit inside the same catalogue and the same conversation, a buyer who needs both formats is not managing two suppliers with two sets of continuity risk. If you are scoping a portal, a campaign or an app, browse the catalogue or license HTML5 games by category and come to the call with the titles you actually want, plus the delivery shape you need them in.

🧸 Licensing Branded Games for Campaigns, Portals, Events and Apps

Forestry Games also develops HTML5 games in-house and works with branded IP, with brand partnerships including Disney, Nickelodeon, Cartoon Network and Warner Bros. Businesses can license branded game content through Forestry Games for marketing campaigns, white-label game portals, live events and mobile apps, alongside the generic catalogue.

Branded work carries the extra contractual layer described above, so scope it early rather than at the end of a campaign timeline. The practical next step is to send the use case — surface, territory, term, and whether you need branded or generic titles — and ask for a licence scope in return. If you need to see the delivery shape before committing, ask for a portal demo, or look at what a catalogue membership includes.

🧭 Do This Before the Next Renewal, Not After the Next Scare

There is no version of a licensor's collapse where you improvise your way to a working catalogue in a fortnight. The window in which you can fix any of this is the window in which the licensor still wants your renewal.

So: run the one-hour dependency test this week and write down which titles fail. Find out which of the three delivery shapes you are actually in, rather than which one you believe you are in. Then take two clauses into the next redline — delivery defined as self-hostable files, and survival of your exhibition rights on licensor insolvency. Both are ordinary asks that a direct licensor with its own catalogue should be able to agree without drama. A supplier that cannot agree to either has told you something useful about what you are buying, and the time to hear it is before the term starts.