At its in-house Capcom Open Conference RE: 2026 on Oct. 2, programmer Satoshi Ishida outlined the Capcom RE Engine REX project: a step-by-step upgrade of the company’s in-house RE Engine toward what he called an “AI game engine,” focused on code, development tools and automated test play, not AI-made art for shipped games.
Part of our Tech, AI and Business coverage. Browse more in Artificial Intelligence.
Why Capcom started REX
According to GameBiz, Capcom began developing RE Engine in 2014, and it first shipped with Resident Evil 7 in 2017. Video Games Chronicle reports Capcom has used it for nearly 30 titles and that the shared technology is widely credited for the publisher’s recent run of hits. Ishida said eight-plus years of use created three problems, per GameBiz: games have grown so large that even small edit-and-check cycles take long waits; the user base has widened to more than 2,000 Capcom developers, including overseas staff and people from other engines; and AI tools now have to fit into everyday workflows. REX is the umbrella name for the improvements aimed at performance, tools, efficiency and QA.
What Capcom RE Engine REX includes
Capcom is not rebuilding the engine from scratch. GameBiz says it will add new technologies named “RE:” something in stages. The subsystems covered in the talk were:
- RE:Dox translates different data formats into one common set of rules for faster loading and editing.
- RE:UI is a lighter framework for building development tools, designed to avoid freezes and memory pauses and to run without a window being displayed, which makes automated testing easier.
- RE:Log records events as small ID numbers instead of text, gathers logs on a central server, and unifies logging and communication so failures are easier to trace.
- RE:Flows lets designers build game behavior visually while the system converts it into standard code behind the scenes.
- RE:Runtime manages large numbers of on-screen characters and objects in grouped blocks and translates easy-to-write code into Capcom’s own RE:C++ for speed.
AI for code and test play, not game art
The “AI game engine” label comes from unifying the whole system around a widely used programming language. Ishida said, as reported by both outlets, that this makes the engine easier for AI to read, so in the future it could write programs and run automated test plays to check for bugs. The stance matches earlier comments from Capcom VP Shinichi Inoue, quoted by VGC, that the company sees AI “not for creating art, but for unleashing the potential of creators.” Inoue described the goal as a medium that clarifies intent between people. Ishida’s stated aim is a future where Capcom can “create games together with AI.”
GameBiz adds that RE:UI uses a one-way data flow and redraws only when something changes, which matters because developers spend long stretches looking at tools without touching them. Capcom says it can be mixed into existing tools gradually, so teams do not have to rebuild everything at once.
Open technology for outside developers
Capcom also said it will publish some of its technologies online so outside developers can use them, including for AI training. GameBiz reports it has already released pieces such as RE:Dox and RE:Log as open source.
Why it matters for players and the industry
If REX shortens QA loops without flooding storefronts with generative art, it could ease the cost pressures that show up in other gaming business stories we have covered, from EU scrutiny of Minecraft and Candy Crush currencies to Live Nation’s Creator Nation live-events bet. Capcom has not announced a consumer ship date for REX tools. The near-term signals to watch are whether the published engine components gain outside users and whether automated playtesting shows up in postmortems for the next RE Engine titles.
Sources
- GameBiz (Japanese): Capcom RE Engine report
- Video Games Chronicle: Capcom details plans to evolve RE Engine into an ‘AI game engine’
Further Reading
Read next in Technology
Artificial IntelligenceOpenAI Expands Free ChatGPT Access to as Many as 200,000 Researchers, Doubling Its July Goal
Artificial IntelligenceOpenAI Signs Four Independent Labs for Embedded Safety Assessments, Starting With a Six-Person, Four-Month Apollo Pilot
Artificial IntelligenceMicrosoft Unveils Seven AI Models at Ignite and Says Its Coding Model Hits 98% of GPT-6.1 Sol’s Quality at 40% of the Price
Most read this week
- Finland Confirms Strike on Russian Warship in Gulf of Finland; Ukraine’s SBU Claims Drone Attack
- U.S. Envoys Witkoff and Kushner Hold Talks With Putin in Moscow, Then Head to Dublin to Brief Zelensky
- Israeli Strike on Gaza’s Nasser Hospital Kills Six, Including a Surgeon, Health Ministry Says
- Nvidia Board Hires Wachtell’s Joon Kim to Lead Huang Probe as Shares Fall 5.47% and an Investor Sues
- Drones Kill Workers at Zaporizhzhia Nuclear and Thermal Plants, IAEA Says, as Russian Strikes Hit 5 Substations
Latest
- Kazakhstan Replaces Its Security Chief After Corruption Arrests at Samruk-Kazyna and the National Railway
- Malaysia Lets France Buy Raw Rare Earths in a “Unique” Deal as Macron and Anwar Seal a Strategic Partnership
- Papua New Guinea’s Marape Tells Xi Closing Taiwan’s Office “Wasn’t Easy”
- Denmark’s Parliament Sets Thursday Vote on Greenland Defense Deal That Opens New U.S. Bases
- Guatemala’s Congress Ratifies State of Siege in Zacapa and Izabal and Lets Troops Take Over Prisons
Get The Boston Report Briefing
Our free weekly email with the most important stories from Boston and the world. Unsubscribe anytime.
Free. We never sell your address. Privacy policy