Worth clearing something up before getting into this one: when people talk about AI in game development right now, they usually mean generative AI, tools that write code, generate art, or produce dialogue. Capcom's RE Engine is a different, older kind of AI entirely, the systems that decide where an enemy walks, when it attacks, and how a whole pack of monsters coordinates against you. It's worth understanding both, and RE Engine is one of the best examples of the older kind done well.

What RE Engine Actually Is

RE Engine is Capcom's in-house game engine, first built for Resident Evil 7 in 2017 and used across every Capcom title since, Resident Evil Village and Requiem, the Resident Evil 2, 3, and 4 remakes, Monster Hunter Rise and Wilds, Devil May Cry 5, Street Fighter 6, Dragon's Dogma 2. Capcom's own head of development said at the time that a generic third-party engine wouldn't have suited what Resident Evil 7 needed, so the studio rebuilt its tooling from scratch rather than license something off the shelf. That's a genuinely unusual choice in an industry where more and more Japanese studios have shifted toward Unreal Engine over the past decade.

Navigation AI, the Unglamorous Engine That Does the Real Work

Capcom has actually presented on this directly at its own developer conference, walking through how RE Engine's Navigation AI decides where enemies and NPCs go and how they get there. The system builds NavMesh data representing the walkable space in a level, and Capcom's own engineers gave concrete numbers from Monster Hunter Rise as an example: a single map's navigation mesh can run anywhere from 5,000 to 20,000 nodes depending on whether it's built for large monsters or smaller companion characters, totaling roughly 2 to 8 megabytes of data per map. That's a genuinely large amount of spatial data being processed constantly just so a monster can chase you convincingly.

Some enemies are governed by full behavior trees, decision structures that weigh multiple possible actions and pick one based on the current situation, while others use simpler logic depending on what a given creature actually needs to do. Capcom's engineers have also talked about ongoing work automating how much of this navigation data gets reduced and optimized, since a naive approach would eat far more memory and processing time than a game can afford to spend on it.

Where It Gets Genuinely Hard: Monster Hunter Wilds

Monster Hunter Wilds pushed RE Engine's AI systems further than any previous Capcom title, according to director Yuya Tokuda. The game moved away from the series' older loading-screen-separated zones toward fully seamless, significantly larger maps, and introduced groups of the same monster species roaming together rather than single isolated creatures. Tokuda described the actual technical challenge plainly: calculating AI for multiple different monster groups simultaneously, tracking where each group is heading, and still letting the player isolate and target one specific alpha monster out of a moving herd. That's a meaningfully harder problem than one enemy reacting to one player, and it's a good example of what "smarter enemy AI" actually requires under the hood rather than just meaning "the monster is tougher."

Why This Distinction Actually Matters

Most of what gets called "AI in gaming" in coverage right now, including a lot of what we cover here, is about generative tools speeding up production or changing what NPCs can say. RE Engine's navigation and behavior systems are a different, longer-running tradition of game AI, the kind that's been quietly running underneath big-budget games for decades, refined incrementally rather than reinvented by a new model release. Both are legitimately "AI in games." They're just solving completely different problems, and conflating them tends to flatten what's actually a much richer, more varied field than the current generative AI conversation alone suggests.

Sourcing note: technical details here come from Capcom's own 2023 Open Conference presentation on RE Engine's Navigation AI and Capcom's public developer commentary via Creative Bloq's interview with Monster Hunter Wilds director Yuya Tokuda, both primary, on-record sources rather than secondhand summary.

Does traditional game AI like this deserve more coverage alongside the generative AI stories we usually run? Let us know your thoughts in the comments. And if you've got a game you'd like us to review, send it over to editor@gamevibe.io.