stadiongaming.eu

Deadlock just out-peaked Team Fortress 2 - and it isn't even released yet

Valve's invite-only hero shooter hit 262,145 concurrent players on Steam, edging past TF2's 253,997 all-time record. The trigger: a chess wizard called Solomon. The real story: a game with no release date is already Valve's third-biggest thing.

Rene Sjeverac

Sunday, October 11, 2026

Deadlock just out-peaked Team Fortress 2 - and it isn't even released yet

262,145. That's the number Deadlock put on SteamDB this week, and it matters because the game you need to beat on Valve's own storefront to get there was Team Fortress 2 at 253,997 - a record TF2 set in 2023 after sixteen years of grinding away. Deadlock did it while still technically invite-only, with no release date, no marketing campaign, and a store page that for most of its life said "you can't buy this."

The immediate trigger is a hero. Solomon is a wizard who turns a chunk of the lane into a chessboard: a knight piece that drops barriers for your team, rooks that punish anyone standing in the wrong square, and a queen ultimate that wipes the grid. On paper it's a gimmick. In practice it's the kind of ability kit that gets clipped a thousand times on day one, and every clip is free advertising for a game Valve isn't advertising.

But Solomon is the spark, not the fuel. The fuel was the "City Never Sleeps" update at the end of September, which already pushed Deadlock to a fresh peak in the 228k range and buried its previous record of roughly 171k from the September 2024 hype wave. Two record-breaking peaks inside two weeks is not a content drop doing well. That's a game crossing from "alpha people check out" into "thing people actually main."

Why I care more about the retention curve than the headline

I've watched enough shooter launches to be allergic to peak CCU as a metric. A peak is one screenshot. What tells you whether a competitive game has legs is what the curve does three weeks later, when the clips dry up and the only reason to queue is that the game itself is good. Deadlock's problem in 2024 was exactly that: huge spike, then a long bleed as the matches turned into 40-minute slogs and the balance swung hard with every patch.

The difference this time is that the floor is higher. The game that hit 262k is a visibly different build - reworked map, a swollen hero roster, and a matchmaking system that has had two years of punishment. If the September-to-October climb holds instead of cratering, Valve has a third live pillar alongside CS2 and Dota 2, and it's a MOBA-shooter hybrid that every studio who tried the genre before them failed to make stick.

The esports angle is the one I'm watching. Dota 2's entire competitive ecosystem grew out of Valve letting the community run with it for years before The International made it official. Deadlock already has grassroots cups and a scene that organizes itself around patches. If the player base stabilizes anywhere near these numbers, a Valve-backed circuit isn't a question of if.

And a small personal note for the config tinkerers: Source 2 at these player counts means the game's performance profile is about to get scrutinized the way CS2's was. I'd expect launcher option threads, frametime comparisons and the first "Deadlock stutter fix" guides to show up within the month. I'll be logging 1% lows the moment I get a match in.

Image: Sergey Galyonkin / CC BY-SA 2.0, source: https://commons.wikimedia.org/wiki/File:Steam_Machine_from_Gigabyte_and_Steam_Controller.jpg