Engineering disciplines speed and cost; creative accelerates storytelling — bound by a human-centric principle and a distributed org

■ Key Points

  • Strategic shift: from a marquee external deal (OpenAI/Sora, dead in March) to in-house tools, usage discipline and internal capability.

  • Engineering front: raise development velocity while curbing “tokenmaxxing.” An “AI Adoption Dashboard” checks usage (tokens = cost). Claude, Cursor, DisneyGPT, and Jarvis in the works.

  • Creative front: Imagineering will design theme-park attractions with Adobe Firefly Foundry — a “responsible” model trained on Disney IP (sketch→2D→3D) — even as Disney sues Midjourney and others over copyrights.

  • Robot front: a 10-year, $60B parks investment — aquatic manta ray (→ Moana’s Gramma Tala) and dolphin robots, a hovering-droid food truck, and next-gen projection aimed at Disneyland Abu Dhabi.

  • Through-line: a “responsible, human-centric” approach — technology as the means of storytelling. Some staff still worry about job displacement.

  • Distributed org: no single chief AI officer — policy in OTE (under Bergman), engineering with Smith/Rohe, creative in Imagineering R&D (Laughlin).

  • Business & Korea: with streaming at its first double-digit margins, cost discipline is pivotal; for Korea, lessons in cost design, IP-specific models, and strategic autonomy.

Disney’s AI strategy has turned from a marquee external deal toward in-house capability. After its $1 billion OpenAI deal collapsed in March, Disney disciplines engineering — raising development velocity while curbing output-agnostic “tokenmaxxing” — and, in creative, has brought a “commercially safe” Adobe Firefly Foundry, trained on its own IP, into theme-park design.