Finnish game studio MiTale showcased its RPG-in-development "C.L.A.Y. - The Last Redemption" at gamescom 2026. What drew attention was its integration of Qiskit—the software framework originally developed by IBM—into game production, under the banner of "quantum computing in a commercial game." However, no quantum processor is actually speeding up the game. Instead, quantum circuits are simulated on a classical computer, and their probabilistic outputs feed back into the map, character relationships, and story branches.
This distinction changes how the technology should be evaluated. Tracing C.L.A.Y.'s approach reveals a realistic entry point for bringing quantum computing into games: rather than starting with a race for speed, the developers chose areas where some variability in results wouldn't break the experience, while keeping a classical fallback path intact. What MiTale has spent more than six years testing is not the performance of quantum hardware, but this design methodology.
Feeding Quantum Circuit Outputs Back Into Unity
C.L.A.Y.'s implementation splits into two parts: the game itself, running in Unity, and a Python/Qiskit backend running on a local server. The Unity side, written in C#, triggers pre-selected quantum circuit simulations in response to player actions. The backend returns the computation results as structured data, which the game then reflects in changes to the map or events.
Natasha Skult's doctoral dissertation, submitted to the University of Turku in March 2026, revealed this architecture. Skult is MiTale's Creative Director and a co-designer of C.L.A.Y. The research began in 2019, when James Wootton, then at IBM, visited MiTale and introduced procedural generation using Qiskit. The following year, in 2020, Wootton published a paper presenting map-generation procedures suited to near-term quantum devices.
Qiskit is a software framework for writing quantum circuits and running them on either simulators or quantum processors. C.L.A.Y. used the former. Even in IBM's current documentation, the Qiskit SDK's reference implementation is described as a local state-vector simulator, clearly distinguished from execution on actual hardware.
So there is no quantum processor sitting inside a player's PC, nor is the game connecting to cloud-based quantum hardware every time it runs. Instead, circuit behavior governed by quantum mechanics is computed on a CPU or GPU, and values usable by the game are derived from the resulting measurement distributions. What's novel here isn't the computing resource itself, but the fact that quantum circuits were treated as a component of the production pipeline.
Starting From the Parts Least Likely to Break
The first application was maps and exploration ranges shaped by a "fog" mechanic. Changing the placement of terrain, items, and encounters is unlikely to disrupt the main storyline. Wootton's earlier research had also covered map generation, making this the lowest-risk implementation for MiTale. The dissertation states that this component worked as intended.
From there, the team moved into two areas with greater impact on the game. One was character growth and relationships, which shift depending on the player's choices and behavioral patterns. The other was a puzzle system called the Ombrascope, in which players communicate through symbols with a broken, AI-controlled machine.
| Application area | What Qiskit's output determines | Design role |
|---|---|---|
| Map exploration | Placement of terrain, items, and encounters | Lowest risk; the variation itself becomes part of the gameplay |
| Character relationships | Relationship shifts based on action history | One of the higher-impact areas for the game |
| Ombrascope | Responses to symbols and branching outcomes | One of the higher-impact areas for the game |
Character relationships and the Ombrascope carry more weight on the game than the map system does. In the Ombrascope, the same symbol won't always produce the same answer. Once past actions and play style are factored in as inputs, what counts as the "correct" answer in one scenario can become "incorrect" in another context. The development team leaned on the game's premise—that the entity you're communicating with is a malfunctioning AI—so that unexpected responses register as a character trait rather than a narrative breakdown.
There's a design principle here that extends well beyond quantum technology: don't wire a new technology whose output is hard to fully control directly into core progression; instead, place its variability where the story's context can absorb it. The team carefully bounded the scope for character relationships, and turned unpredictability into a storytelling device for the puzzle. In other words, the game was designed around the technology's inherent characteristics.
That said, the quantum circuits aren't generating text or images out of thin air. According to Skult's dissertation, in-game content is authored by developers, and no generative AI is used. What Qiskit and the game systems handle is the selection and combination of pre-built modules. Any marketing claim that "quantum computing generated the characters or graphics" should be read within these bounds.
No Real Hardware, No Quantum Advantage—Yet
Quantum simulators are useful for designing and verifying quantum circuits. But that doesn't mean they harness the computational power of quantum computers. In typical simulations that faithfully track the state vector, both the memory and computation required on the classical side grow exponentially as the number of qubits increases. IBM itself has indicated that even with high-performance computing, simulator usefulness tends to plateau at around 50 qubits.
For C.L.A.Y., neither the type of circuits nor the number of qubits used has been disclosed. The number of measurements and execution times are also unknown. There's no evaluation of how much production time or quality differed compared to ordinary random-number generation or classical procedural generation. It's more reasonable to view this as a case study exploring whether the approach is usable for game design, rather than as a benchmark demonstrating quantum advantage.
The development team itself decoupled this dependency. C.L.A.Y. lets players disable quantum simulation in the settings menu and still run on classical methods instead. This avoids a scenario where a failure in third-party software takes down the entire commercial game, and it also allows for comparing the experience between the two modes. The quantum feature was built in as a swappable subsystem, not as an essential engine.
This decision also has implications for future connections to real hardware. In a design where the core game can't proceed without waiting on a response from the quantum side, any problem with that connection immediately becomes a product-breaking failure. By keeping the classical path intact, C.L.A.Y. separates experimentation from product operations.
MiTale has mentioned the future possibility that actual quantum hardware could accelerate complex processing, much like a GPU does. However, C.L.A.Y. does not demonstrate that future. What can be confirmed at this point is that the team tested the software's connection points and game design ahead of real hardware, rather than waiting for it.
The "First Commercial Game" Claim Awaits Verification After Launch
MiTale's press kit describes C.L.A.Y. as "the world's first commercial game to use quantum computing for immersive storytelling." But as of August 28, 2026, its Steam page still reads "Coming soon," with no reviews yet. The press kit's current product listing puts the release for 2027. In the sense of a finished product actually on the market, the claim doesn't hold up yet.
The scope of "first" also deserves careful handling. Games themed around quantum technology, educational quantum games, and small-scale games that run on actual quantum hardware have existed before. A 2024 IEEE paper also reports that hobbyists and companies have built numerous quantum-related games. A more accurate framing of C.L.A.Y.'s claim might be: an early example of a narrative-driven RPG aiming for commercial distribution that incorporates quantum circuit simulation into both its production pipeline and player experience.
What sets C.L.A.Y. apart isn't the "first" label so much as the length of its development history. The concept was first explored in 2019, experimented with from 2020 to 2021, and documented in a peer-reviewed book chapter in 2022. Following a peer-reviewed journal paper in 2024, the 2026 dissertation went on to record the connection architecture and the design choices made to avoid failure. The gamescom showcase wasn't a sudden demo—it marks the stage where years of research and development are being tested against the demands of an upcoming commercial RPG.
Still, the effect on players hasn't been proven. The dissertation acknowledges that too few people participated in the demos to gather sufficient data on immersion or replayability. The Steam page makes no mention of the quantum feature either, so it's unclear which features will make it into the final release.
What will ultimately settle the matter is the scope of implementation and comparative results the development team presents after launch. Do player choices and experiences actually differ between quantum mode and classical mode? When the same design is connected to real hardware, does the difference justify the added wait time and operational cost? For C.L.A.Y. to become a precedent for quantum-enabled games, its 2027 release will need to answer these two questions with reproducible data.