The disc is still in the case and the console still turns on, but the game will not start. Somewhere between the purchase and the present, a server that authenticated the copy was switched off, and a title that shipped on physical media now behaves like a rental that expired without notice. Anyone who has tried to replay a decade-old game has met some version of this moment. The reflex is to read it as a technical failure. It is not. Its data is intact, emulators run it well, and a disc image copies to a hard drive in seconds. What stops the copy from being useful is not whether it can be made, but whether making it is allowed.
Preservation turned into a legal problem when the acts required to preserve a game became the acts copyright reserves. The statute gives a rights holder control over reproduction and distribution, so an archive that dumps a cartridge and serves the file to researchers is doing precisely what the law forbids, even if nobody has sold the original in years. Section 1201 of the Digital Millennium Copyright Act adds a second layer: it makes circumventing a technical protection measure unlawful on its own terms. Defeating an authentication check to read the data on a disc can therefore be a violation independent of any copying. That is why the legal risk arrives before the technical question is even answered.
The law does leave doors open, but they are narrow and they move. Libraries, archives and museums can claim exemptions for some preservation work, and the Copyright Office has granted exemptions for specific activities, including letting preservationists access the server software behind abandoned online games. What matters is the structure of the process. Exemptions come out of a rulemaking that recurs on a fixed cycle. Each one is scoped to a defined use and a defined class of user, and every round requires advocates to demonstrate harm again from the beginning. The cost of that cycle falls on whoever is willing to do the paperwork. Preservation is not a settled right, then, but a permission slip that has to be reissued.
Abandonment has no legal definition. A game does not die on a date the law recognizes; it simply stops being sold.
Follow the incentives and the resistance stops looking irrational. A back catalog is not a dead asset to a publisher; it is a stock of options. A remaster, a bundle, a sequel's re-release, or a slot in a subscription catalog all gain value when the original is hard to buy legally, and all lose a little when a working copy circulates freely. Broad preservation exemptions would also hand the same tools to people who never intended to preserve anything. No rights holder wants to build that road on the assurance that the traffic will behave.
A related shift is underway inside the industry itself. As physical media fades and storefronts delist titles for commercial or licensing reasons, the working canon of older games is increasingly whatever sits in the rotation of a few subscription services. That keeps what the platform finds worth keeping and drops the rest, on commercial grounds that have nothing to do with the archive.
A title that stays licensed is available indefinitely. A title that does not is gone, no matter how many people bought it.
Even a publisher that wants to keep a game alive faces an assembly problem unrelated to the code. A single release can bundle licensed music, an engine from one vendor, middleware from others, actors' performances, and brand or sports licenses that expire on their own clocks. Each is a separate contract with a separate holder, and any one of them can block a re-release by holding out or by being impossible to locate. This is why reissues sometimes surface with swapped music or without a licensed car. The barrier is often not appetite but the number of parties who must agree.
The archive's problem is that ownership of a copy was never the thing being sold.
Costs settle on the people least able to carry them. Volunteer communities maintain emulators, document hardware behavior, and verify disc dumps, all under legal exposure that a studio or a platform holder would never tolerate. Institutional archives have counsel and narrow exemptions. Hobbyists have neither, and the difference shows in what gets attempted. Developers land on the same side of the line, since many cannot legally keep or exhibit the work they built once a studio closes and the assets are divided among creditors. Discs rot, drives fail, and hardware that depends on a dead online service degrades with no replacement path. Historians of the medium increasingly write about games they are not permitted to run, which is a strange foundation for a field that claims to study them.
A second front has opened that sidesteps copyright almost entirely. When a publisher shuts down a game's servers, modes that required a connection stop working, and the grievance is not about copying at all. It is about what was sold. Consumer protection law addresses that directly: if a product was marketed as a purchase and later disabled, the question becomes whether the buyer received what the seller promised. This framing has made the issue legible to legislators with no appetite for rewriting copyright, and it is why proposals now center on requiring publishers to leave games in a playable state when support ends. Different jurisdictions draw the line in different places, so an archive's exposure can depend on where its servers happen to sit, an accident of geography rather than principle.
Real change would mean moving the question out of case-by-case permission and into statute. Codifying the museum and archive exemptions so they stop lapsing would erase a recurring cost that buys nothing. A workable orphan works regime, or a defined legal standard for abandonment, would give archives something to act on. Narrowing Section 1201 along the lines the Copyright Office has recommended would let preservationists do technical work without carrying a separate legal risk. Deposit requirements, under which a publisher hands source and assets to an archive before support ends, are routine in film and would translate directly. End-of-life rules forcing a playable offline build when servers close would cover the consumer side. None of this is exotic. Each has been proposed, and each has stalled for the same reason.
Reform stalls because the people who care about preservation are diffuse and the people who profit from the status quo are not. A publisher defending a catalog is protecting revenue and a licensing business. A preservationist is protecting a shared record, a benefit spread across everyone and paid for by nobody in particular. That asymmetry has shaped copyright policy for decades, and it is why the practical victories have been incremental, scoped, and reversible.
What the fight has not settled is whether a game is property or a service. If it is property, a purchase should outlast the seller's exit, and some obligation to keep it usable follows. If it is a service, the buyer always held a license and the ending was written into the terms, whatever the storefront page implied. The law has drifted toward the second answer while buyers have gone on assuming the first, and preservation lives in the gap between them. No emulator will close it.
The question is who is responsible for the record of a medium its makers remain free to switch off.