S16.10 · Travel, Hospitality & Leisure
Ticketed destinations and experience venues, the sector's most AI-insulated segment given its physical-asset moat.
There is no single reliable global revenue figure for theme parks and attractions — vendor estimates range $50-130B+ depending on scope — but the anchors are solid: Disney's Experiences segment alone generated roughly $34B in FY2024, and the top 25 global parks draw approximately 500M visits a year combined (TEA/AECOM Theme Index). Of every segment in the sector, this is the one AI touches least. Land, ride capital expenditure and exclusive intellectual property make a physical-asset moat that no booking agent can route around.
Growth is forecast at mid-single digits, driven by experiential-spending trends among consumers and new-build expansion in Asia and the Middle East. Revenue concentrates in North America, where Disney, Universal and Six Flags are the dominant operators; attendance growth concentrates in Asia-Pacific, led by parks such as Chimelong and Universal Beijing, and in Saudi Arabia's large-scale new destination developments. Concentration is high: Disney, Universal (owned by Comcast), Merlin Entertainments and Six Flags account for the large majority of global attendance among branded operators. Upstream, parks depend on ride manufacturers and, where relevant, licensed intellectual property; downstream, tickets sell through both direct channels and online travel agencies.
The core economics are extremely capital-intensive — long payback periods and high fixed costs are structural — but a built park generates strong recurring visitation, which is what justifies the capital burden. On top of ownership sits a much higher-margin layer: licensing intellectual property to build branded lands and attractions with international partners, an asset-light business closer to a royalty stream than to real-estate ownership.
The exposure is shallow, and shallow by design. Guest-service and wayfinding labor is the cost line contracting, as in-park AI trip-planning tools reduce the staff needed for basic guest questions and itinerary logistics. Beyond that, the moat holds: land ownership, ride capital expenditure and exclusive intellectual-property rights are not things an AI agent can substitute for or disintermediate, in contrast to the search-and-booking funnels that define AI exposure elsewhere in travel. The horizon is long, 5-10 years, and even at the end of it the effect stays shallow.
The one strategic move worth watching — intellectual-property owners building their own parks rather than only licensing content to third-party operators — is expansionary, monetizing existing content assets beyond streaming and merchandising revenue. It is driven by content economics, not by anything AI-related.