What did Valve change in Steam's pre-release review process?
Valve published the change on November 3, 2026, in a Steamworks post titled "Updates to the Pre-Release Review Process." The post adds new items to the checklist developers must complete before a store page or build enters Valve's review queue. Valve says the changes are meant to help more games pass on the first attempt.
The most unusual addition is a playtime requirement. Developers must now launch their game through Steam and leave it running for at least five minutes. Any player's playtime counts, and it must fall within the past seven days, or since the last build if that build is older than a week.
Valve's explanation is blunt. "Believe it or not, sometimes we review builds that have not ever been played through Steam," the post says. Launching the executable directly can miss steps, and Valve says developers "might have missed basic steps that would impact a user installing and playing via Steam."
Playtime also has to be recorded before the checklist will clear. Valve warns of a delay of a couple of minutes after a player exits the game, which is enough to catch a team submitting at the last minute. Any user's session counts, not just the lead developer's.
Valve treats the store page and the build as separate submissions. The page review checks that descriptions, screenshots and features match what will ship. The build review checks that the game starts and works through Steam. Both normally take three to five business days, and either can produce a re-review.
The rest of the list covers common submission gaps:
- •Confirm which common redistributables Steam should install, such as DirectX, Visual C++, .NET or XNA, or declare that none are needed.
- •Check that achievements are configured correctly, including names, icons and duplicate entries.
- •Write a save file to Steam Cloud from a member of the partner group when cloud saves are listed as a feature.
- •Upload at least five screenshots categorized as gameplay, with menu shots, concept art and cinematics allowed only under their own labels.
- •Complete dynamic reminders at submission, such as language support and test credentials for multiplayer.
Why are Steam reviews taking up to 12 business days?
Valve's own FAQ sets the expectation. It says the team aims to test every build "typically within 3-5 business days." It then adds that Valve is "taking up to 12 business days" because of re-reviews.
Valve calls the package a starting point rather than a finished fix. The post says the initial improvements "shipped in the past weeks" and that more changes are coming to build review.
That 12-day figure is a backlog, not a new standard. As of November 5, Steamworks documentation still lists a three-to-five-day review and asks developers to plan for seven days. The checklist changes have not yet been written into that public page; Valve says a reference guide is "coming soon."
Business days exclude weekends, so 12 of them span more than two calendar weeks. For a studio trying to hit a fixed release date, that is enough time for a marketing plan to go stale. Steam's documentation also notes that adult-content pages and builds can take longer than the typical window.
The volume explains part of the strain. IGN reports that about 24,299 games have launched on Steam so far in 2026, against roughly 4,500 in all of 2016. SteamDB figures cited by Rock Paper Shotgun show almost 4,000 releases in October 2026, compared with just over 2,000 in October 2025.
Valve says it has been trying to keep up. The post says the company has "spent the last several weeks working to scale our process to match" the submission increase. It also acknowledges the human cost, saying Valve has "heard from a lot of developers" about waits that ran longer than expected.
Re-reviews compound the problem. A rejected build goes back into the queue, so a basic error can cost two waits instead of one. Valve describes the delay as "having to re-review games, which isn't great for you or for us."
Is AI slop to blame for the Steam review backlog?
Valve does not blame AI in the post. Eurogamer notes that the company "did not directly name gen-AI" as a cause. PC Gamer says Valve "did not comment on why" the logjam developed. IGN says the post did not address the AI issue at all.
The timing invites the question. Engadget cites a Reddit analysis finding that 49.9% of October 2026 releases carried an AI disclosure. Rock Paper Shotgun, citing Game Discovery Tool, puts the share above 50%, up from a little over 25% in January 2026. Valve has required AI disclosure since January 2024.
AI use shows up in demos as well. Engadget reports that nearly 20% of Next Fest game demos disclosed AI use in June. One developer told Engadget the spike appeared around September and October, pointing to "the overwhelming amount of slop now flooding Steam."
A disclosure is not a quality rating. It records that a developer used generative AI, not how much or how well. Rock Paper Shotgun warns that correlation is not causation, and Valve has not published data linking AI submissions to rejection rates.
The disclosure is part of the review itself. Valve's post says reviewers compare game content with "elements disclosed in the content survey, such as mature content and AI usage." A mismatch between the build and its survey answers is a review problem on its own.
Valve's public position has been permissive. A Valve developer recently said the company does not want to "be gatekeepers telling developers which tools they should or shouldn't use," according to PC Gamer. At a Steam Indie Connect event in Korea, Valve said it had no plans to ban AI games and suggested developers make their work harder to clone, IGN reports.
Disclosure has its own politics. IGN notes that Steam gives customers the option to "vote with their wallets" through those disclosures, and that Epic Games boss Tim Sweeney has called the approach a "political message." That argument shapes how many developers disclose, and therefore how useful the release data is.
Other platforms are moving differently. Itch.io recently deleted what it described as terabytes of AI slop, PC Gamer and IGN report. Steam's approach so far is to speed up review, not to narrow what can be submitted.
Which games have been delayed by the Steam review queue?
The delays are visible in public developer posts. Studio Daimon submitted Food Devils on October 12 and was still waiting on November 3, more than 20 days later. Food Devils had been scheduled to launch that day, according to Rock Paper Shotgun. The studio said it could "only assume" the queue was "clogged up by AI slop."
The studio also said it had raised the urgency of its case. It sent another ticket warning that it "might miss our release date because of this," according to Rock Paper Shotgun. The ticket, the studio said, was ignored.
The cost is momentum. Cubic Earth developer structmira told PC Gamer the game's trailer had more than 60,000 YouTube views and the page had more than 6,000 wishlists. "Now, our Steam Community page is rapidly filling with player complaints, and we've lost all momentum," the developer said.
PC Gamer also names Wheel Make It among games stuck in the same limbo. Eurogamer reports a further worry: some developers fear an AI-generated clone could reach Steam before the original clears review.
These accounts are self-reported, and Valve has not commented on individual cases. They do show where the queue bites: small teams with fixed launch dates, finished builds and no way to speed up the process. Players see the same delay from the other side, as announced dates pass without a release.
What do the new checklist items mean for indie developers?
The new checklist lands hardest on developers who handle their own releases. IGN notes that games from studios without a major publisher's track record must go through Valve's testing process.
Several items target failures that are cheap to fix before submission but expensive after rejection:
The documentation gap matters. The new items appear in the release checklist inside Steamworks, not yet on the public review page. Developers who plan from the old page may miss them entirely.
The old scheduling advice also needs revision. Valve has long asked for at least seven business days before launch to absorb feedback. In a 12-day queue, that buffer no longer covers the worst case, and a launch date announced months earlier can slip for reasons outside the studio's control.
Publishers can absorb a delay by shifting a campaign or a store placement. A self-funded studio usually cannot. Wishlists, press coverage and creator videos are timed to a date, and a missed date weakens all three at once.
- •Test the build through Steam, not just the executable, so launch options and depots are exercised.
- •Declare redistributables, because a missing DirectX or Visual C++ install can stop a game from starting on a clean PC.
- •Verify cloud saves from a partner account, with developer mode turned off.
- •Make sure every language promised on the store page is reachable in the build.
- •Categorize screenshots correctly; five gameplay images are now the minimum.
How can developers avoid a re-review?
Valve's FAQ points to three checks its reviewers apply. The first is whether the store page and assets are clear and complete; vague descriptions and illegible capsule logos are common failures. The second is whether the build matches what the page promises, including features and language options. The third is whether sensitive content, including AI usage, is disclosed in the content survey.
The practical lesson is to treat the store page as part of the submission, not marketing added later. Reviewers install and play the build, and they compare it with the page. A missing feature or an unreachable language option can send the build back to the queue.
Valve's reviewers check that promised languages can actually be selected. The post says testers ask whether they can "get to all the language options promised on the store page." A language listed on the page but buried behind an unfinished menu will fail that check.
Multiplayer creates its own trap. Valve's post says reviewers may need test accounts to reach online features before release, and the submission form now asks for credentials. A build that cannot be tested is a build that cannot be approved.
For updates, the rules are lighter. Valve's documentation says a game that has been reviewed and approved does not need review again, and developers can update freely. The bottleneck is concentrated at first release and at store page approval.
Will Steam review times go back to normal?
Valve says more changes are coming. The post notes that the company has spent weeks scaling its process and still has "a ways to go." It gives no date for a return to normal or for the documentation update.
Several signals will show whether the fixes work. The first is whether developers report queue times falling back toward three to five business days. The second is whether re-review rates drop once the new checklist beds in. The third is whether Valve publishes the promised reference guide and any staffing or automation changes.
AI disclosure is the wildcard. If the share of releases carrying a disclosure keeps climbing, review volume may keep rising even if rejection rates fall. Valve has shown no appetite for a ban, so the queue, not the rulebook, is where the pressure will appear.
Seasonal timing adds risk. A backlog in early November sits in front of the holiday release window, when many studios want to launch. If the queue does not shorten before then, more small games will slip into December or later.
The post does not say Valve will reject more games or accept fewer of them. It adds checks and asks for better-prepared submissions, which is a change in workload rather than in what Steam is willing to sell.
For now, the safe plan is the cautious one. Teams with fixed dates should assume a 12-business-day worst case, submit earlier than the old seven-day buffer, and treat the release checklist as a pre-flight test rather than paperwork. Valve's own framing is that the changes exist to help games pass review "the first time."
