MYBIGGAMING
LIVE
Agothic
News 29 October 2026 • 2 min read •

Risks of moving too slowly: Epic Games details genAI use in Unreal Engine and Fortnite development

Executive VP Marcus Wassmer reveals how generative AI is helping Epic Games fix bugs, accelerate coding, and reshape engine architecture.
Author: Флешка •
Risks of moving too slowly: Epic Games details genAI use in Unreal Engine and Fortnite development

Epic Games is moving beyond vague statements about AI toward specific implementation details. Marcus Wassmer, Executive Vice President of Development, shared with This Week in Videogames how generative AI is being deployed across Unreal Engine and Fortnite development.

According to Wassmer, the most prominent tools currently in use are Claude Code and Codex, along with their associated AI models. These tools have already delivered significant technical wins. Most notably, the company has nearly eliminated non-deterministic cooking bugs—an issue that has plagued development for a decade. Additionally, AI has accelerated shader compression work, which has directly resulted in significantly smaller patch sizes.

One of the most striking examples of efficiency was the rewrite of a major portion of the D3D12 rendering layer. Wassmer estimated that a task requiring three to six weeks of work from a single engineer was actually completed in approximately one week. This process not only saved time but also improved CPU performance and reduced bug counts. AI is also being used to maintain and expand the company's internal test suites for better coverage.

For Fortnite, the implementation includes automatic PC stability monitoring, automatic root cause analysis, and automatic fix generation, which has sped up the resolution of long-standing issues. On the engine side, Unreal Engine developers are now attempting larger architectural refactors that were previously difficult due to the massive volume of required file changes.

Wassmer also confirmed the use of what he termed "vibe-coding" within Unreal Engine development. In this workflow, engineers no longer write the bulk of the code manually; instead, they guide AI tools to achieve the desired architectural results. The goal is not to "one-shot" the engine, but to iterate with AI to ship requested features while staying within strict memory, performance, and scalability envelopes.

When addressing the industry's shift toward these technologies, Wassmer suggested that Epic Games is more concerned about the risks of moving too slowly than they are about AI skepticism.

Article author

Флешка

Cookies

We use cookies

The site uses technically necessary cookies for sign-in, security, storing preferences, and correct page operation. Analytics and other non-essential technologies should only be enabled after your consent. You can read more about data categories, purposes, and retention in the privacy policy.

Cookie policy
Quick actions