全球航空科技：AI分销——航空分销战争的新前沿

Global Air Tech: AI distribution - the new front in airlinedistribution wars AI is everywhere, even the slow-to-change world of air travel distribution. A fourthdistribution channel looks ready to be born. What does it mean for Global DistributionSystems? In our view, if AI becomes a significant distribution channel for airlines, then wewould expect airlines to refuse to cede control of this channel to GDS intermediaries. Havingalready broken free of full content agreements, we expect them to resist content parity, andtherefore set the incentives for LLMs to connect directly to airline APIs. Alex Irving, CFA+44 20 7676 7044alex.irving@bernsteinsg.com Antoine Madre+33 1 58 98 74 52antoine.madre@bernsteinsg.com A new distribution channel for airlines.Airlines basically have three distribution channelstoday: direct (website, app etc); classical indirect (offer creation outside airline) and modernindirect (offer creation inside airline). AI is a potential fourth channel. Agentic APIs couldallow customers to book tickets on airlines without leaving their favorite AI tool. Airlineshave to decide how to engage with this, urgently. Specialist Sales Kiran Shah+44 20 3547 1533kiran.shah@bernsteinsg.com Airlines have the power to own this channel.This is not the advent of OTAs, or evenmetasearch. Airlines have broken free of full content agreements, and have far morecontrol over what content they offer in what channel. We expect GDSs to be unable to forcecontent parity. Airlines that take control of their data and distribution should be able to offera better product, without paying commissions to legacy intermediaries. The need to treadcarefully is not the same: going dark on a GDS created revenue risk, but there currently isno revenue in this channel to disrupt. Airline CCOs, in our view, can afford to take more riskthan in the move to NDC. AI tools should have both the scale and technology to connect direct.Say an airlinebuilds an agentic API to connect into. Can the AI distribution channels use it? We havelong argued that going direct to airlines requires the buyer to have both technologicalcompetence (to build the connection) and scale (to make the cost economics work). Withmeaningful content differentiation, we expect the AI tools to opt for the superior content,which airlines should restrict to their own API. Yet expect change slowly.We remain to be convinced that large booking flows willcome through AI to airlines in the next couple of years. That looks especially difficult in thecorporate channel, where servicing capabilities and TMC processes will likely mean gradualadoption at best. Furthermore, it would be wrong to think of this as only replacing indirectbookings: airline.com and airline app bookings may also move into AI channels. Net net, wecontinue to expect GDSs to grow, but lose share of total air bookings (c. 2% GDS bookinggrowth vs air travel growing 3-4% p.a.), but not to see a full-on collapse of the GDS channel. Remain Outperform on both Amadeus and Sabre.We continue to like Amadeus for themedium term. We see the company as able to deliver a 10% EPS CAGR with a 3% dividendon top, with a deep and widening competitive moat around its main businesses. We expectfurther significant wins for its Nevio product and associated Order Management System todrive upside, and remain Outperform with a new price target of €87.00. Sabre meanwhilelooks to us oversold, with new agency wins driving EBITDA growth into 2026, leading tofurther deleveraging. We remain Outperform on a target price of $3.00. BERNSTEIN TICKER TABLE INVESTMENT IMPLICATIONS Amadeus (Outperform, €87.00)is the largest player in Air distribution and Air IT. It is skewed to better businesses, with >60%of EBITDA contribution (and rising) coming from Air IT and Hospitality, leaving 40% - and falling - coming from the ex-growthGDS business. Amadeus continues to take share among GDSs in distribution, and keeps picking up new contracts in Air IT. Thebalance sheet is healthy, and the group is returning cash. We see a c. 14% annual TSR in steady state at a constant multiple:7% annual growth in net income, with a 3% dividend yield and 4% buyback (which keeps leverage constant, but we do notrule out the possibility of accretive M&A). The products are well-invested and well-positioned for the future. Nevio also lookslikely to lead to further contract wins in Air IT, and likely to pick up steam following the announcement of deals with both BritishAirways and Air France-KLM. Further eventual share gains for Amadeus, either from competitors or through newly-outsourcedcontracts, continue to look likely. We remain Outperform. Our price target of €87 is set at 23.27 our 2026 EPS estimate of €3.7,increased from €84 in line with our estimate of EPS. Sabre (Outperform, $3.00)is far more skewed to the ex-growth GDS market, at c. 80% of revenues following the hospitalitysale. Geographically, it over-indexes to North America, where we expect volumes and u