The promise is right there in the job ad. It sits between the requirements for Unreal Engine 5 and C++. A Game Designer (LLM) will be responsible for "creating AI-driven characters that feel alive and react to the world and player actions in meaningful ways" [welevel (Personio)]. It is a single line of text in a sea of open roles, but it is the entire premise of welevel. This Munich studio has spent four years quietly building a game and the proprietary AI to power it from the ground up. The debut title is a medieval survival city-builder called SolidLake (or Solid River), aiming for an early access launch next year [Triple Point PR, March 2025] [Steam, 2026]. The bet is that the studio's own generative tools for world and character creation will be the wedge that lets a small team build something that feels vast.
The procedural wedge
Welevel is not the first to chase the dream of a truly dynamic, player-responsive world. The survival genre, from Valheim to Palworld, is built on loops of gathering, building, and overcoming environmental threats. The city-builder genre thrives on intricate systems management. Welevel's stated goal is to fuse them. Players progress from scrappy survivor to lord of a sprawling settlement in a world that is procedurally generated and populated by AI-driven NPCs [BITKRAFT Ventures Job Board, 2026]. The studio's differentiation is not just the feature set but the production method. By developing its own generative AI models for terrain, assets, and character behavior, the team believes it can drastically reduce the cost and time of creating AAA-quality content, turning a 25-person team into a much larger virtual workforce [GamesWirtschaft.de, Sep 2024].
Why BITKRAFT wrote the check
The $5.7 million seed round, led by gaming specialist BITKRAFT Ventures in March 2025, signals a belief in this technical wedge [Triple Point PR, March 2025]. The investor's portfolio is a who's who of next-generation game studios and infrastructure. They see welevel's in-house AI development as a foundational capability. The angel list is equally telling, blending gaming industry veterans like Riot Games co-founder Daniel Weinand and former Oculus co-founder Nate Mitchell with German football star Mario Götze [Triple Point PR, March 2025]. This mix points to a round built on both technical credibility and local network strength in Munich's growing tech scene.
The execution cliff
The counter-bet here is steep. The AAA gaming space is brutally competitive and expensive, even with AI cost-saving measures. Welevel's tools are unproven at scale. The dream of "alive" AI NPCs has been a graveyard of ambitious promises and disappointing, robotic results. The studio must not only build a fun game, a monumental task in itself, but also prove that its proprietary AI stack delivers a tangible, player-perceivable advantage over studios using off-the-shelf engines and assets. Furthermore, the 2026 timeline lands in a market that will be crowded with other AI-native gaming experiments, all vying for the same player attention and talent.
The next twelve months
For welevel, 2025 is the year of proof. The milestones are clear: solidify the game's core identity, ramp up the team from its current estimated 25 members, and likely begin controlled external testing of its AI systems [GamesWirtschaft.de, Sep 2024]. The real cultural question the product is implicitly answering is not about technology, but about scarcity. In an era where player time is the ultimate currency, can a game built by algorithms to be endlessly novel finally solve the problem of running out of things to do? The studio's entire premise rests on the hope that the answer is yes. The tools they build to get there will become their most valuable asset. The first glimpse of that answer should arrive on Steam next year.