Epic’s reported ‘AI first’ pivot puts its Disney shooter on an indefinite delay, Ars reports

Ars Technica, citing a Bloomberg report, says Epic Games is shifting its collaboration with Disney to an ‘AI first’ development model. Internal messages describe human developers managing and guiding assets produced by models such as Anthropic’s Claude, with ‘no more hand-written code’ and art generated through full AI workflows. According to Ars, the Marvel- and Star Wars-themed shooter from the $1.5 billion Epic–Disney partnership was supposed to arrive in March 2027 and is now delayed indefinitely amid the pivot and slower, understaffed development. The Verge described the project as slipping from a March window (Verge report).

Epic’s public position is acceleration, not headcount reduction. Tim Sweeney says the goal is to build more and better work and carry the benefits into Unreal Engine for other teams. Sweeney has argued that AI will be part of nearly all future production and has criticized Steam’s AI disclosure tags. Epic laid off more than 1,000 people earlier this year, citing Fortnite engagement.

If accurate, the reporting shows Sweeney’s views becoming organizational structure, with developers directing model output rather than authoring every line and asset. Game production is expensive and full of repetitive work where coding assistants and art pipelines already help, and better Unreal tools would benefit studios across the industry. Other studios report gains from AI prototyping, such as Playco (Playco and GPT-6 Astra).

The timing raises concerns. The project was already behind, and an AI-voiced Darth Vader in Fortnite showed how quickly a branded character can go off-script. Marvel and Star Wars characters are closely watched by fans, and guiding a model is a different craft from creating the work directly. A developer survey cited by Ars found generative AI used mostly for research and brainstorming, far less for code and prototyping, so Epic is moving well ahead of industry practice. An Epic executive’s earlier statement that the goal is efficiency, not job replacement, will be tested by whether teams grow their output or creative roles shrink.