Royal Caribbean just told Wall Street that AI is core to its business. Not a pilot program. Not a roadmap item. Core.
CEO Jason Liberty used the Q1 2026 earnings call to describe AI as central to "complex, real-time decision making at scale" across pricing, guest personalization, and operations. The numbers back him up. Revenue grew 11% year over year. Adjusted net income hit $1.0 billion. Over half of onboard revenue is now booked before embarkation day through digital systems.
The company manages 15 million price points every day. Ninety percent of those are automated. Their predictive models forecast how many hamburgers to cook every 15 minutes. Not metaphorically. Literally every 15 minutes.
These are real operational results, not demos. And they raise a question that nobody on the earnings call asked.
Where does the AI actually run?
If those pricing models, personalization engines, and demand prediction systems run in a cloud region on shore, they work great in port. They work fine when the Starlink link is clean. They do not work when the satellite connection degrades.
And satellite connections degrade. Regularly. That is what satellite connections do.
Rain fade at Ka band. LEO constellation obscured by the superstructure. Congested anchorage where every yacht and cruise ship in the harbor is fighting for the same spectrum. A single Starlink outage like the one that grounded the Pentagon's drone fleet last month.
When any of those things happen, a cloud-dependent AI stack goes from "real-time decision making at scale" to "real-time nothing." The guest-facing systems that Royal Caribbean just told investors are driving 11% revenue growth go silent.
The revenue math cuts both ways
Here is what makes this interesting. Royal Caribbean's own numbers prove the business case for on-vessel AI better than any white paper could.
If over half of onboard revenue is booked through AI-driven digital systems before embarkation, and the rest is driven by in-voyage upselling, personalization, and dynamic pricing, then the AI is not a nice-to-have. It is revenue infrastructure. It is the thing that keeps the money flowing.
Revenue infrastructure that depends on a satellite link is revenue infrastructure with a single point of failure. We wrote about this exact pattern when Virgin Voyages scaled to 1,500 cloud AI agents. The results were real. The connectivity risk was also real.
A local inference layer that caches guest preferences, runs pricing models, and generates personalization recommendations on the ship itself would keep the revenue engine running regardless of satellite availability. Not instead of the cloud. In addition to it. Sync when the link is up. Serve when the link is down. The guest never notices the difference.
What this means for your vessel
You do not have Royal Caribbean's scale. You probably do not have 15 million daily price points. But the pattern is the same.
If you are deploying AI for guest concierge, crew support, or operational optimization on a yacht or expedition vessel, the same question applies: what happens when the link drops?
The answer for a $4.5-billion-a-quarter cruise line should be the same as the answer for a 60-meter yacht. The AI keeps working. The guest keeps getting answers. The crew keeps getting support. The satellite link is a nice-to-have for syncing, not a requirement for operating.
That is what sovereign AI at sea actually means. Not a marketing phrase. A design principle that separates the vessels that impress people from the vessels that apologize to them.
The real signal in the earnings call
Royal Caribbean just validated the entire thesis. AI is not optional for modern vessel operations. It is the revenue engine. It is the guest experience. It is the operational backbone.
The only missing piece is making sure it works when the satellite link does not. And that is exactly the gap we build for.
Thinking about how to keep your guest AI running when the link drops? Let's talk. We help vessel operators deploy sovereign AI that works in port, at sea, and everywhere in between.