AI Catering Is Finally Leaving the Lab

For decades, inflight catering has been managed with a mix of historical averages, route knowledge, cabin-class rules, and operational caution. Airlines would rather carry too much than disappoint passengers. But that buffer has a cost: wasted food, extra weight, unnecessary fuel burn, crew complexity, and a sustainability story that is getting harder to defend.

That is why Airbus’s recent Smart Catering trial with Virgin Atlantic is worth watching. This April, Airbus said it had tested an AI- and data-driven catering system in live conditions on Virgin Atlantic flights in 2025, including its A330 service between London and New York and its A350 service between London and Orlando. The system automatically captured passenger meal consumption and traced unused food and drink, moving the idea of AI-enabled catering from a cool demo into real flight operations.

IATA and the Aviation Sustainability Forum have estimated that the aviation sector generates more than 3.6 million metric tonnes of cabin and catering waste annually, with food and beverage waste making up 65% of that total. Untouched meals account for 18% of all waste. Airbus frames the problem similarly, saying that 18% to 20% of cabin and catering waste is untouched food and drink, much of which is incinerated or sent to landfill due to international rules.

Framing the results as simply “less waste” doesn’t do justice to the opportunity for airlines. It’s a working example of how better demand intelligence can lead directly to reduced costs and better margins.

Airlines know who is flying on the aircraft. They know route history, cabin mix, loyalty status, booking channel, special-meal requests, pre-order behavior, time of day, and seasonality. But those signals have not always translated cleanly into catering decisions. The galley still often reflects static assumptions made long before the passenger sits down. AI changes that by creating a feedback loop between what was loaded, what was actually consumed, what was untouched, and what should be adjusted on the next flight.

Airbus’s Smart Catering concept is especially interesting because it doesn’t appear to require a radical change in crew workflow. The system uses AI software on existing crew tablets or mobile devices. During normal service, a device camera recognizes meals and beverages as they are removed from the trolley, updating onboard stock. The same interface can show live inventory, trolley, and galley location, dietary and nutrition information, and route-level insights once data is uploaded to a ground cloud for analysis.

That matters because the cabin is not a laboratory. Any technology that adds friction for crew is unlikely to scale. Airbus says the system was designed to capture consumption data without extra crew activity, while also reducing manual forms and reporting. During the trial, the tool provided interactive galley search, live inventory, and dietary information, giving crew a clearer picture of what was available and where it was located.

For airline leaders, this creates four business cases at once.

The first is cost. Catering waste is purchased, handled, loaded, flown, removed, and disposed of. Even before considering the price of the food itself, every surplus item carries labor, logistics, fuel, and waste-management costs. Airbus says Smart Catering has the potential to deliver double-digit reductions in preventable food and beverage waste carried onboard.

The second is operational reliability. A smarter galley can help crew answer basic service questions faster: what is left, where is it, and whether a passenger’s dietary need can be met. That has implications for service recovery, premium-cabin consistency, and crew workload.

The third is personalization. Gategroup’s North America president Jens Kuhlen recently described a similar direction for the industry: using data to understand passenger preferences by cabin class and support more curated experiences, including plant-based, allergen-sensitive, culturally relevant, and wellness-oriented options. He also said AI can capture real-time data on consumption, packaging weight, and onboard waste, helping airlines rethink portioning, streamline packaging, support ESG commitments, and improve load efficiency and cost.

The fourth is customer choice. United’s January 2026 expansion of online meal pre-ordering shows the same logic from a passenger-facing angle. United said passengers on eligible North America and Caribbean flights can pre-select meals through its app or website, giving catering partners clearer production signals and helping reduce waste. The airline claimed the system would save more than 100,000 pounds of food waste.

There is also a regulatory wrinkle. IATA notes that many countries require international catering waste to be incinerated, sterilized, or sent to deep landfill to protect agriculture, making reuse and recycling difficult. That means prevention is more valuable than recovery. Once excess food is on an international aircraft, the airline may have few attractive options. The smarter move is not to load it in the first place.

The larger story is that catering is becoming part of the airline data stack. It touches customer experience, procurement, sustainability, labor efficiency, loyalty, fuel burn, and brand perception. For executives, the question is no longer whether AI can make menus smarter. It is whether airlines can turn catering from a cost center governed by averages into a responsive system governed by evidence.

The winners will not be the airlines that simply cut meals. They will be the ones that know, with increasing precision, what passengers actually want, what they actually consume, and what never needed to be loaded at all.