Capcom has unveiled an ambitious, multi‑stage initiative to transform its in‑house RE Engine into what the company calls an "AI‑generation engine." The goal of this transformation is to streamline the game‑development pipeline, improve overall performance, and give developers new AI‑driven tools that can handle routine tasks more efficiently. The RE Engine, which first entered development in 2014, was originally built to power Resident Evil 7: Biohazard, a title that debuted in 2017.
Since then, the engine has grown into a widely adopted platform within Capcom, with estimates from Eurogamer suggesting that over two thousand developers across the company now rely on it for a variety of projects ranging from horror titles to action‑adventure games. At Capcom's most recent annual technology conference, programmer Satoshi Ishida delivered a detailed presentation titled "The Outlook and Future of the REX Project, Further Evolving the RE Engine for the Next Generation." In this talk, Ishida outlined a roadmap for the REX project that emphasizes faster development cycles, more efficient quality‑assurance (QA) processes, and a smoother onboarding experience for both new hires and international team members. Central to this roadmap is the integration of artificial‑intelligence capabilities directly into the engine's core workflow. Several new subsystems are set to be introduced as part of the upgrade.
RE:Dox will serve as a standardized framework for handling and processing data, ensuring that large volumes of asset information can be organized and accessed consistently across teams. RE:UI is designed to speed up the development toolkit, providing a more intuitive user interface that reduces the learning curve for engineers and artists alike.
Meanwhile, RE:Log introduces an ID‑based logging system that offers granular, searchable records of engine events, making debugging and performance analysis far more straightforward. Additional enhancements include RE:Flows, a visual‑scripting environment that allows designers to create gameplay logic without writing code, and a reworked runtime core known as RE:Runtime, which promises greater stability and lower latency during gameplay. The execution framework will also be refined under the name RE:C++, delivering a more robust and flexible C++ foundation that can accommodate the demands of modern, high‑fidelity games. Ishida emphasized that AI integration will not be a gimmick but a practical assistant for developers.
The AI components are expected to generate and interpret code snippets, run automated playtests to uncover bugs, and suggest performance optimizations based on real‑time telemetry. In Ishida's words, the aim is to enable developers to "create games together with AI," turning the engine into a collaborative partner rather than a static toolset.
To foster a broader ecosystem, Capcom plans to open‑source certain improvements, making them available to the community for AI training purposes. By releasing these tools publicly, the company hopes to accelerate the development of AI models that can be leveraged not only within Capcom but also by external studios seeking to adopt similar workflows. This open‑source strategy aligns with the long‑term vision of evolving the RE Engine into a full‑blown "AI‑generation game engine," where AI can assist in everything from asset creation to level design. Earlier in the year, Capcom made it clear that it would not embed generative‑AI‑produced assets directly into its games.
Instead, the focus remains on using AI to boost efficiency and productivity throughout the development process. A company spokesperson explained, "We are currently experimenting with various AI‑driven methods across departments, including graphics, sound, and programming, to see how they can best support our teams without compromising artistic intent." The move toward AI‑enhanced tooling follows a recent controversy involving Resident Evil Requiem, which was featured in a demonstration of Nvidia's DLSS 5 technology. The demo showcased AI‑upscaled character models, including the protagonist Grace Ashcroft, and sparked debate over the role of AI in visual fidelity. While some praised the technical achievement, others warned that over‑reliance on AI could erode the handcrafted feel that fans associate with the series.
Capcom's measured approach—using AI for internal processes rather than as a direct content generator—appears designed to address these concerns while still reaping the benefits of automation. In practical terms, developers can expect AI to handle repetitive coding tasks, such as writing boilerplate functions or refactoring legacy code, freeing them to focus on creative problem‑solving.
Automated playtesting will allow the engine to run thousands of simulated sessions, flagging crashes or performance bottlenecks before human testers even see the game. Data‑driven performance tuning will analyze frame‑time statistics in real time, recommending shader optimizations or asset compression techniques that keep the game running smoothly on a wide range of hardware. Beyond the immediate technical advantages, the AI‑centric strategy could have broader implications for Capcom's global workforce.
By standardizing data pipelines and providing AI‑assisted translation tools, the company hopes to lower language barriers and make collaboration across its studios in Japan, North America, and Europe more seamless. This could accelerate the sharing of assets and code, reducing duplication of effort and fostering a more unified development culture. Overall, Capcom's REX project represents a forward‑looking attempt to blend cutting‑edge AI research with the practical needs of game production.
By opening parts of the engine to the community, the company invites external innovation while retaining control over the core technology that powers its flagship franchises. If successful, the RE Engine could set a new benchmark for how AI can be woven into the fabric of game creation, delivering faster development cycles, higher quality products, and a more collaborative environment for creators worldwide.