Technology

A Huge Leak of Old Steam Data Has Exposed Canceled Games and Cut Content

Martin HollowayPublished 2month ago4 min readBased on 4 sources
Reading level
A Huge Leak of Old Steam Data Has Exposed Canceled Games and Cut Content
Photo by Tim Eulitz / CC BY 4.0

More than 12 terabytes of data from Steam, the popular PC game store, has leaked online. That is roughly the same as 12,000 standard DVDs worth of information. The data includes internal versions of games, content that was cut from finished games, and projects that were never released, from both Valve (Steam's parent company) and other game publishers (The Verge).

The data comes from old Steam servers that stored and delivered game files to players between 2003 and 2013, according to Tom's Hardware. There is so much data that it could take days just to figure out everything inside it (Game Rant; XDA).

The most talked-about discovery involves Half-Life 2: Episode 3, a game Valve canceled years ago. The leak includes material connected to that project, including something called the Weaponizer, a tool or feature that was apparently meant for Episode 3 (The Verge).

A version of a game called F-Stop also appeared in the leak. F-Stop is set in the Portal universe and was built around a camera mechanic, though exactly how it worked is still being figured out by people digging through the files. F-Stop had been mentioned in earlier leaks and fan discussions, but this is the first time anyone has had a playable version to examine.

There is more Portal material, too. People searching through the data found deleted dialog from Portal 2 that changes how players might view the game's Companion Cubes, along with alternate opening sequences and different versions of scenes involving GLaDOS, the game's artificial intelligence character. Other cut content from Portal 2 is scattered throughout the archive (The Verge).

The leak also contains early versions of games from other publishers. Builds of early Call of Duty and Resident Evil games are in there, along with Mirror's Edge, Batman: Arkham Asylum, and Dragon Age: Origins (The Verge).

The third-party content raises questions that go beyond Valve. Steam's servers from this period stored game files for every publisher using the platform, so the leak could expose early or internal versions of games from dozens of studios. This was the era when Steam was becoming the main way people bought PC games, and the systems used to store and protect that data were not as advanced as they are today.

For people who study and preserve video game history, this leak is a rare and valuable find. Cut content, alternate scenes, and canceled projects are usually kept inside game studios and never shared with the public. The Episode 3 material and the F-Stop build are the most exciting discoveries so far, but the archive is so large that sorting through it all will take the community a long time.

There is a security angle worth considering here. The fact that data this old was still sitting on servers where someone could take it points to a common problem. When companies stop using old servers, the data on them often stays put long after the people who built those systems have moved on. Security measures that were good enough when the servers were active may not be strong enough against today's methods. This is a reminder for any company, not just Valve, that old systems carrying old data can still be a target.

The community has focused mostly on the Valve material so far, which makes sense given how important Half-Life and Portal are in PC gaming. The Episode 3 material is especially notable because the game was canceled quietly, without any official announcement, and fans have been guessing for over a decade about what it might have looked like. Seeing actual game files rather than concept art or rumors is a completely different kind of evidence.

What is not yet clear is how much of the third-party material in the archive is genuinely unreleased or internal, versus finished versions of games that just happened to be stored on Steam's servers at the time. That distinction matters for the studios whose early work is now public, and sorting it out will take time.