A game can miss its release date without a single bug. That is what happened to Australia Did It: End-of-the-World Railway Defense Business, a turn-based train-defense game from Rami Ismail, published by Mystic Forge. It was due on Nov. 10. The date came and went while the game’s build sat in Valve’s review line. According to Make Tech Easier, the publisher posted on Saturday that the build had finally been approved and the game would arrive Nov. 17.
One late indie game is a footnote. A pattern is a story, and this month several small studios have said the same thing: Steam’s approval process, which used to take a few days, now takes weeks. Valve has confirmed part of that. It has not confirmed the explanation many developers have settled on. Below I separate the documented facts from the theories, then say what I think they add up to.
How long are the waits? Four games, one ruler
Valve’s published target is three to five business days for a build review, and its own documentation advises submitting a store page at least seven business days before you want it live. On Nov. 3, Valve said reviews were taking up to 12. Here is how four reported cases measure up, counted in plain calendar days.
Two of the four sit right around Valve’s ceiling. The other two blow past it. I think that spread matters more than the headline number. A ceiling of 12 business days is something a studio can plan around. A queue where some builds go well beyond it, and where a support ticket gets a reply telling you to wait, is not.
That second part comes from Alfred Kamon, director of Studio Daimon, who spoke with Kotaku. He filed a support ticket on Oct. 24, as Steam’s advice suggests. Valve replied on Oct. 26 with a concern about sexual content in Food Devils. He answered right away, and a week later he was still waiting. Other studios describe a similar feeling of silence. Ismail told reporters his team’s build failed over a minor controller bug, that they fixed it the same day, and that they have been back in the queue ever since.
My take: a one-line fix should not cost the same as a first review. Whether it does is one of the many things Valve hasn’t said, and it is the detail developers care about most, because it decides how risky a first rejection is.
Valve’s explanation, graded claim by claim
Much of the online argument mixes four different claims. They are not equally well supported, so here they are one at a time.
- Reviews are taking up to 12 business days.Confirmed
Valve said so itself in a Steamworks post on Nov. 3, and called its usual three-to-five-day target the norm it is currently missing.
- Re-reviews are most of the delay.Valve’s explanation
Valve says much of the wait comes from games that fail at first submission and have to be looked at again. That is its stated cause. It has published no numbers to test it.
- AI slop is clogging the queue.Developers’ theory
Studio Daimon and the publisher 2 Left Thumbs suspect it. Valve’s post does not say it, and nobody outside Valve can see the queue to check.
- The new checklist will bring waits back to normal.Unproven
Valve hopes more games will pass on the first try. It has not said when waits will return to three to five days, and it has given no review-team figures.
Notice the pattern: the two claims that rest on Valve’s own words are the least controversial, and the two that most people repeat are the least proven. That does not make the AI theory wrong. It means the evidence stops at the edge of Valve’s data.
What the five-minute rule actually checks
Valve’s main complaint is almost funny. It says it sometimes reviews builds that were never played through Steam. A game can run fine from a developer’s own executable and still fail for a player who installs it through the client, because a launch option or a content depot is wrong. Each of those failures sends the game back for another round, which is exactly the loop Valve says is eating its time.
- Five minutes of playtime through SteamThe build has to be launched through Steam, and any user’s playtime counts if it falls within the past seven days or since the latest upload. It catches wrong launch options and missing content depots.
- Required redistributablesDevelopers must list things like DirectX, Visual C++, .NET or XNA, or confirm that none are needed.
- Achievements and Cloud savesThe checklist flags incomplete achievements, including missing names, missing icons and duplicate names. If Steam Cloud is listed as a feature, a save file has to reach the cloud.
- Five gameplay screenshotsStore pages need at least five screenshots labeled as gameplay, with separate labels for menus, concept art and in-game cinematics.
- Reminders at submissionPrompts change with the features a page advertises, such as promised language options or login details for multiplayer that reviewers can’t otherwise reach.
These are sensible, cheap fixes. They are also fixes for the developers’ side of the problem. What they can’t do is solve the failures that are about judgment. One developer, 2 Left Thumbs, said Wheel Make It was rejected because reviewers couldn’t work out how to reach its local co-op mode, which turns out to mean sharing a keyboard. That account is the studio’s own, and no checklist item covers it. Coverage of Valve’s announcement also notes that Valve gave no figures for the backlog or for how many people review builds.
My take: it’s a good change that moves work onto developers while leaving reviewer capacity untouched. I would not expect it to shorten the line by itself.
Is it really AI slop? What the numbers can and can’t say
The theory has real material behind it. Kotaku reports that by mid-October more games had been released on Steam in 2026 than in all of 2025, and that around 4,000 games came out in October alone, with nearly half disclosing some use of generative AI in development. Those tallies come from third-party trackers that Kotaku cites. October’s Next Fest also drew more than 8,000 demos, which prompted complaints about vibe-coded filler. Valve has acknowledged a low-quality problem and says it doesn’t want to tell developers which tools to use.
I find the theory plausible. I also think three things should slow anyone down before calling it the answer.
- Disclosure is not quality. Steam’s AI disclosure covers any generative use in art, writing or code. A studio can tick that box for a small assist. I wrote about one such case in my Big Walk story, where the developer said AI touched well under 0.1% of its code.
- Volume is not the same as slop. Valve itself points at the surge in submissions and at re-reviews. A flood of rushed builds that fail basic launch checks would fit that description, but so would a record year for ordinary first-time indies.
- Valve controls the policy. Ismail said slopware is likely part of the issue, but that its place on Steam is Valve’s own policy. He also said the other big storefront isn’t a way out for him, because it would be, in his words, “trading the one billionaire for the other.”
The honest position is that AI-assisted volume is a strong suspect and that nobody has the evidence to convict. To settle it, Valve would need to publish submissions per month, how many fail their first review and the current median wait. It has said none of that so far.
Why November made this worse
Timing turned a slow queue into a painful one. Kotaku notes that dozens of indie games were trying to launch before the month’s heavyweights arrived: Ocarina of Time on Nov. 5, then GTA 6 on Nov. 19. Kamon’s team had planned to release Food Devils on Nov. 5 and moved the date up a day to avoid Zelda. Then the approval never came.
Behind that calendar entry is a nine-person studio in Italy that has worked on the game for four years. Kamon told Kotaku the team needs the revenue before year’s end, and that Steam pays out a month after sales. Delaying into December, right behind GTA 6, wasn’t realistic. He also asked for something simple: an honest line on the submission screen. He wants Valve to say it may take weeks instead of promising three to five days, because developers build hype around dates they are told are safe.
What I would do about it
If you play
- When a wishlisted indie slips by a week or two, check the developer’s own posts before assuming trouble. The cause may be the queue.
- Be gentle with small teams that go quiet around a launch date. A missed date is not a promise broken on purpose.
- Remember that a game can sit on Steam as “Coming soon” while its build is still in review, even after the studio says it is ready.
If you build
- Submit earlier than Valve’s seven-business-day advice and run every item on the checklist first.
- Explain hidden or unusual features, such as shared-keyboard co-op, in your review notes.
- Announce a date after the build is approved. Ismail said his team gambled on Valve coming through and lost.
Quick answers
How long does Steam game review take right now?
Valve said on Nov. 3, 2026 that store page and build reviews are taking up to 12 business days. Its usual target is three to five business days. Valve has not said when waits will return to normal.
Why are Steam game reviews taking so long?
Valve says it has seen a significant jump in submissions and that much of the delay comes from re-reviewing games that fail at first submission. It has not published submission numbers, review-team size or a recovery date.
Is AI slop causing the Steam review delays?
Some indie developers believe so, but Valve’s announcement does not name AI as a cause, and no one outside Valve can see the queue. Valve has separately acknowledged a problem with low-quality and AI-made games on the store. The link between the two is a theory, not a confirmed finding.
What is Steam’s five-minute playtime rule?
Valve’s updated release checklist now requires at least five minutes of playtime launched through Steam before a build can be submitted. Playtime from any user counts if it falls within the past seven days, or since the latest build was uploaded. The goal is to catch broken launch options and missing content before reviewers do.
How early should developers submit a game to Steam now?
Valve’s documentation still advises submitting at least seven business days before the date you want to go live. With waits of up to 12 business days and the chance of a re-review, that buffer looks thin. My advice is to plan for the ceiling plus a second round, and to announce a date only after the build is approved.
The delays are real, documented and costly for a small studio that lives on launch revenue. Valve has given us the symptom, up to 12 business days, and the mechanism it blames, re-reviews. It has not told us what is driving the volume. I’d treat AI-made games as a strong suspect, not a verdict. The five-minute rule is sensible, but until waits shrink I would plan as if the ceiling were the average.
