Kurt, CEO of Sabre, and Mike Randolfi, CFO of Sabre, today with us. We have a lot of questions to go through, so I'll get started, but let's see if we have some time for Q&As, either in person or on the call. Let's get started. Can you give us some color on what you're seeing with the current travel capacity, supply, and demand, given the ongoing Middle East disruption? I think IATA just recently lowered their forecast from 4.9% growth this year to 2.1%. Any views since results?
Yeah. Well, thanks for the question, Victor. I think it's important to start with what our base planning assumption was when we originally built our plans and our expectations for 2026. As part of that, we had originally guided to air distribution bookings up in the mid-single digits. What's important about that is that assumption has been predicated upon, not that there's a point of view on the industry, but a planning assumption that the GDS industry, from an air distribution bookings perspective, would be relatively flat. Now, in February, that was despite the expectation that airline capacity would be up roughly 5%-6%, and we would expect that the industry, from a GDS booking standpoint, should largely correlate with that capacity growth.
Now, coming through the back part of last year into this year, we saw very strong overall bookings growth to the tune of around 9%. A lot of strength in North America and in the Americas in general, North America, Latin America in the double digits, and solid growth in EMEA and in APAC. Obviously you had the Middle East conflict ensue starting in March, and since then, we had seen bookings at roughly flat with North America and South America largely offsetting declines in the Middle East, EMEA, and APAC, driven by the Middle East. A matter of fact, to quantify that, for us, around 11% of our bookings go to, through, or from the Middle East. With the decline as a result of the Middle East conflict, it's created a roughly seven-point headwind, which is what drives us to the flat bookings.
That continued into April, which is the point we referenced on our earnings call. Now, when we updated our guide for this year, we updated our guide for the year from mid-single-digit bookings growth, to low- to mid-single-digit bookings growth. We had stronger growth in the first quarter than we originally expected. With that, the second quarter, based on how we exited the first quarter, we telegraphed roughly flat bookings growth. Even in the third and fourth quarter, if you look what's implied, we tempered our underlying expectations. Such that if you look at the third quarter and what's implied in our guide, it probably implies something like low single digit growth and maybe a little bit better than that in the fourth quarter.
What I would overall say is, if you look at that, we still have bookings growth up three, I would call it low to mid single digits, all predicated upon our own growth trajectory. Overall, if you look at why we didn't have a bigger impact, it really goes back to our original planning assumption of flat industry growth. Overall, we felt pretty good about our guide, where we see bookings and capacity. We haven't updated anything since our earnings call. Overall, I'd say the environment overall is pretty still consistent with how we created our current projections for air bookings growth and earnings.
Just two quick adds to that. One is that corporate and TMC volumes have been fairly resilient. The question will be the elasticity of the consumer with respect to airlines passing on price increases to offset the impact of fuel, and there's some uncertainty there in terms of what the impact will be going forward. The second thing is, while our core assumption is that the war, and we could be wrong on this, abates in the near term, the impact of fuel from a logistics standpoint and the impact of fuel price and jet fuel, is that that's going to persist through the calendar year. Our hope and expectation is that normalizes by the end of the year, but that doesn't basically turn around overnight. That's going to have a resultant impact on airline pricing yield and perhaps demand impact as well.
Yeah. What's buffering the impact to us is, if you look leading coming out of last year into this year, in terms of our GDS bookings at Sabre, we've been outperforming the industry by about 500- 600 basis points, and that has buffered us in this turbulent time.
Very clear, good color. Maybe then if I jump into AI. First off, on the distribution space, we actually talked about this six months ago in the Fireside Chat about what has happened, what we are doing at the time. Maybe can you give us an update, what has changed since then? What worked and what doesn't? How do you see the distribution space in travel evolving with regards to AI?
Yeah. I think that the broad theme is that t here's going to be an inherent mandate for consumer-grade conversational commerce interfaces, whether you're an OTA or a TMC or a supplier, or you're a new agentic platform. One of the cool things that's happened more recently is we launched a new relationship with PayPal and Mindtrip. We are now live in production with Mindtrip, where they're the LLM player. They're using Sabre for search, book, servicing changes, full live production today. You can do all of that through talking to or typing in freeform spirit into mindtrip.ai. Hotel will go live with that very quickly. That's built basically, we have new MCP server capabilities, new agentic API capabilities. That's all built on the top of Sabre's normal data and infrastructure. The beauty of this is it's simply a distribution extension of what we are doing.
One of the key things we've seen or one of the questions is, I believe that the agentic platforms will emerge as a new distribution channel in the way the online travel agents emerged in the 1990s, and I think that's going to happen rapidly. Now, one of the learnings will be in traditional e-commerce interface today, you enter defined parameters when you try to search for air or hotel travel. In a free-form text, when you say, "I want to go to the South Pacific in the summer," and we have 6 billion airfares stored, plus we go source all the NDC stuff, how do you return relevant content? The great news is we have 17,000 shops per second against our system. We've got the data, we've got the caching to do that intelligently.
What the LLMs are going to have to figure out is how do you take this natural language search? How do you drive better parameters in order to have a more responsive query to the traveler? Because if it's too open-ended, you're not going to give them the response they're looking for. I think that's something that, as we speak to the LLMs, there's no question about. The other is most of the large agentic players are hyper-focused on monetization. They're focused first on e-commerce retail because it's such a large category, and sneakers are presumably easier to sell than air travel. I think they're going to turn their attention to travel starting in the second half of this year. What they all articulate they want, and this includes Google Gemini, is a captive end-to-end user experience where the consumer never leaves their portal.
That's very different, for example, than Google Flight Search. It's more common to an OTA. The way I think about it'll be the next generation of what an OTA looks like with a conversation on the front end of it. Clearly, the OTAs are going to do that. Suppliers are going to try to do that. I think these agentic players are going to get big into that as well.
Right. Then you mentioned Google. If I think about Google, they've continued to push with more AI offerings on travel planning, travel booking. While they're not looking to become an OTA themselves, does it not risk shifting the top of the funnel to direct bookings? How should we think about the funnel and the distribution mix changing in the context of AI?
Google's very interesting. Think about Google Hotel as very discrete from Google Flight Search. Google Hotel is an open marketplace. OTAs, metasearch, and hoteliers compete for that inventory. Highly profitable, high-margin channel for Google, but it's an ad model. Google Flight Search, since they bought ITA Software in 2014, similar in that it's metasearch. If you choose to book, it'll take you to the supplier, but it's not a true open marketplace. They generally restrict the ability of metasearch or OTAs to compete there, and they largely don't charge the airlines. I think that's for antitrust concerns. The airlines are getting a severe amount of what they consider direct traffic from Google Flight Search that they're largely not paying for, and that's transacting in the airline.com environment.
If you think of it a future state where Google says, "I may shift the funnel from Google Flight Search or Google Hotel into Gemini." If they go with what we're hearing from everybody else, which is they want a captive end-to-end user experience. On the hotel world, that means they're taking this incredibly lucrative business and they're converting it into where they're effectively acting as the fulfillment arm. The question is, how would they monetize that versus what they do on Google Hotel? I think that's a big open question.
On the air side, if they were, for example, to say, "I'm going to take all that volume," because they have massive volume coming into Google Flights, and they shifted that to Gemini, what would happen is each airline may get their fair share of bookings, but instead of transacting the airline.com environment, the transaction will now occur in Gemini, and presumably, they're going to look to monetize that from the supplier. Not only does the airline lose the control of the booking, but they would pay for it as well. I don't think that necessarily means a change in the top of the funnel. What it means is a change in where the transaction occurs, and then the question of the economics. The most impacted here, presumably, would be the airline.
Or if the OTA or the hotelier are getting bookings from Google Hotels, that's now transacting in, let's say Google Gemini, that could be very different as well. I do think that there will be some level of channel shift toward the agentic players, whether it's Gemini or Anthropic, whoever wins at this, again, similar to the way the OTA dynamic worked in the 1990s.
Very clear. I'm going to shift gears a bit to TMCs, corporate travel. I see a lot of movement in the corporate space, GBT, Amex GBT going private, partnership with SAP and then Navan, Travelport, very focused on AI efforts. How do you think the TMC will adapt to AI? Does that mean they rely on you less for solutions? If we think about the long fat tail of TMCs as well, can they be competitive in that world? Can you talk a bit about that space?
I think if you look at Navan or you look at GBT going private with one of the sponsors as an AI company, there's three applications for AI that are pretty inherent. One is productivity. TMCs today have a, let's call it a 60% gross margin. Part of that is because it's very labor-intensive, very manual. If you could theoretically AI half of that work, you could drive up the gross margin of the TMC dramatically, where it looks much more like a SaaS company. I would assume that's job number one, for example, in the GBT pending acquisition.
Number two is the TMC model is shifting and looking more like a retailer or an OTA, subject to the confines of their customer contracts to where if you looked historically, maybe 10 or 15 years ago, 60%-70% of the revenues for a TMC were derived from their corporate customers. The inverse is true today, where the majority of revenue is derived from suppliers, and it's about product placement and the ability to move share, et cetera. I would assume you're going to deploy AI there to drive meaningful share, to merchandise better, just as OTAs and retailers do. Number three is, while I don't think TMCs are at risk of disintermediation from AI because the corporation, for example, B of A use TMCs, you're dictated by procurement or HR. You will book in this TMC or that online booking tool.
The consumer, the traveler, the employee is going to mandate that they have a consumer-grade conversational experience like they would have in the next generation of OTA or the next generation of an AI agent. What that's going to do is accelerate the technology arms race and perhaps accelerate consolidation for TMCs. Now we have, as we look at the TMC landscape, we've got dozens of TMCs and agencies using our agentic APIs already, really to help them solve against all three of those challenges. I think the other part of your question, Victor, is will the TMC try to disintermediate what we as the GDS do? I would argue that what we do is highly complex. The sexy way I speak about the business is we're the technical plumbing of the travel industry. That is the data and the underlying complexity that we have is severe.
We have 17,000 airshops against Sabre every single second. When we extend into the agentic world and we provide a relevant search response and caching, we don't have the data just from Mindtrip or the next LLM, but we combine that with everything else, and the relevance of search response, the ability to do caching and scale is fundamental. For all the work we do, our take rate with a hotelier or an airline is about 1.5% of their revenue. That's a very thin revenue for what we do in almost any industry, certainly in travel. For all the complexity we do to go after a 1.5% revenue stream when the profit pools around payments or merchant model are much larger, doesn't seem the best allocation of resource, especially when, by the way, with agencies, we share a portion of our gross revenue with them.
They're capturing, call it, half that margin without having to do any of the work. I think it's unlikely that they're going to go down that path. The other one I would mention is Navan, which spouts its direct connect profile, is a key customer of ours, and we're growing quite well with them. We're going to be supportive of customers, whatever they want to do. We think that in an agentic world where look-to-books are going to explode, the utility provided by Sabre is only going to improve.
Very clear. Maybe, yeah, I got a couple more questions on distribution, maybe last one for now before we come back if we have more time. If I were to put it this way, in the age of AI or agentic AI, who will be the biggest winners, potential winners, and losers in the shifting world of distribution?
Yeah. I think that there's two ways to look at it, which are technology companies that have engineering prowess. For example, the large OTAs are likely to become good as LLM players. Folks that are infrastructure and data players like us and like our largest competitor in Spain are likely to be winners. I think the key question in agentic AI is, as the distribution landscape changes, how will the funnel of eyeballs change? My thesis on this is fairly straightforward. When you look at supplier.com, you have loyal customers of the hotel or the airline. You have credit card customers. They're not going to modify their behavior.
The itinerant shopper who goes to an airline.com once a year and is largely indifferent as to which airline they're flying, if they're getting a better user experience in the next generation of OTA or the next agentic layer, they may shift share away from supplier.com. The airline, the hotel may still get that booking, but it will not occur in their captive environment, and they may have to pay for it tomorrow where they don't pay for it yesterday. When you look at intermediary travel, I'll break it down into leisure and then corporate. On the leisure side, you have metasearch, Google Flights, trivago, KAYAK, Skyscanner, et cetera. They provide a great utility in providing price arbitrage to the consumer. The user experience is a little bit icky in terms of being a link off model, not great servicing if you have a change, et cetera.
Again, to the extent that I used to buy books on Amazon, now I buy groceries. If tomorrow on Anthropic or Perplexity or whomever you choose to buy your travel, it may be that that user experience supersedes and takes some share away from metasearch. When you look further at leisure travel, OTAs, I think the large OTAs at scale are going to continue to do very well. They're going to develop good conversational commerce layers. It's just an extension of what they do already. Scale will be important here to play this game. When you look at sort of offline leisure or brick-and-mortar, the folks that are specialists around complex travel, I don't think are at great risk. If you're selling a point-to-point easy solution, though, that probably is subject to disintermediation.
When you go to corporate travel, again, as I mentioned earlier, I don't think managed corporate travel is subject to disintermediation, but perhaps more of a technology arms race and consolidation. Unmanaged corporate travel, which is the Wild West, which accrues to supplier.com, to the OTA, to the TMC. I think that's going to continue to fragment and certainly there's going to be some channel leakage to the new agentic players. The beauty for Sabre is that when you speak about the first categories of risk, which are supplier.com and metasearch, we derive little or no bookings from those channels. When you look at OTAs, that's about 25% of our turnover. We are the most heavily concentrated with TMCs, it's 45% of our distribution. It's about 25%-30% for the GDSs overall.
There's a slight amount of disintermediation risk, but when we look at, for example, what we've done with Mindtrip and the way we're leaning into potential partnership opportunities with the large agentic players, we think there is a step function growth opportunity for Sabre not reflected in our current projections and numbers, and this could change the revenue trajectory for our company. We're really excited about where this is going.
Very clear. Maybe now let's jump to airline software, airline IT. Maybe I'll try to address the elephant in the room for a lot of investors in the last couple of weeks. I think you made some public comments about Amadeus maybe being a bit more monopolistic in this space. Can you elaborate a bit on that, what you're seeing in that, and how you can maybe break that mold?
Yeah, sure. Quick backdrop. Airline IT, if you think about PSS or Passenger Service System, Amadeus with Altéa and Navitaire has 50%-60% of the global market. Sabre has about 17% of the market. If you go back in time to 2015 to the midst of COVID, Sabre was a net loser, Amadeus was the net winner in that we were selling an older platform against Altéa, and we were losing. What has changed since then is in the middle of 2024, we introduced a new platform for offer and order called Sabre Mosaic. It's a cloud-native modular AI-infused platform. It is agnostic to underlying platform, meaning it can sit on top of SabreSonic, Altéa, it can sit on an airline's proprietary technology. The beauty of it is we can sell individual modules.
We don't require the monolithic binary change of an airline from one PSS to another. That's exactly in line with what the airline's trying to achieve, which is modularity and not being reliant on a single vendor. What we understand Amadeus has developed with Nevio, which is their equivalent, which by the way, they have less in production than we do, on offer and order, is it's modular on top of Amadeus Altéa. What we understand is if you're an airline that's not on Altéa, that is not available for you. You have to first migrate to Altéa, one legacy PSS to another, then upgrade to that. We think what we have is a truly modular solution. We understand that they do not. If you spoke to leading industry consultants, they would back that. My comments about Amadeus's anti-competitive behavior are as follows.
We've had in the last year four instances where Amadeus Altéa customers were ready to adopt one or more of our Sabre Mosaic modules on the Offer side. Ultimately, Amadeus did three things to impede that. Number one is, despite asserting that they are an open platform, they said either we don't have APIs for that piece of technology, which betrays that, or they said we will not make that API available to Sabre. Number two is they engaged in foot-dragging tactics to say things that should take weeks would take nine months to one year. Number three is they went back to the airlines and imposed dramatically high fees that made it commercially untenable for them to adopt the Sabre solutions.
We've engaged with Amadeus, we believe that what's going on here is Amadeus is using its position, its dominant monopoly position with PSS to block the ability of others like Sabre to compete for those airlines business for the new offer and order solutions. They're using one dominant position to block the new technology. What we want is a fair marketplace in which we can compete.
Very clear. If I think about then challenging that status quo, do you think there's a chance that you can win share from that base in the maybe, let's say, medium term? If so, what needs to happen for that to play out?
Right now we're growing in the airline IT business in the mid-single digits. It's turned from a negative to a positive business for us. For the medium to long term, I'll say two years out, we believe that inclusive of the Amadeus market position, this will be a double-digit CAGR business for Sabre. We're very confident about that. I would say what Amadeus is doing is logical, it's simply not legal. We want to compete on the basis of the quality of our technology. Right now, the addressable market there is very challenging for us.
Well understood. If I think about, obviously in the airline IT space, moving to offer and order is a big topic, but if I layer that on top with AI as well, how do you see AI impacting the shift to offer and order? Is it accelerating that trend or are airlines exploring other alternatives? Obviously, we have examples of, for example, Air India working with Anthropic, Ryanair talking about Ryanair Labs, albeit they're not Sabre customers.
There's two different things. One is airlines are looking at distribution, which is how do they sell. I think that's totally discrete from the airline technology or the base hosting conversation. On the hosting side, one is modularity enables a more deliberate shift from the legacy PSS world to the new offer and order world that you don't have to make a binary change and otherwise you basically need like for like on day one for everything. This way you can say we're going to implement pricing first, then revenue management and DCS. Tomorrow we'll do a new inventory solution. Later we'll do a new departure control. That's what modularity enables, and you can choose best-of-breed software, that's where the market goes.
What AI does is With an existing PSS and a network carrier, you'll typically have hundreds of applications or technologies that hang off of the PSS. We could show up tomorrow and say, "Hey, we've got the whole new offer and order breadbasket, take it all." For a large airline, that could be a $500 million+ change effort and introduce a lot of operating risk to their business. Nobody wants that. If they can phase that over three, five, seven years, they can do that in more deliberate fashion. That's what they want to do. Part of the challenge is a lot of their legacy technologies are still mainframe TPF based. They may not have all the code written.
What AI can do is AI can reverse engineer the code, They can basically say, here's what you're moving off of, This is what we're going to lose when we move. They can make that change effort much simpler than it was in the old world. The ability for airlines to make offer and order real and mitigate the operating risk to me is the best benefit that they're going to get from AI.
Very clear. If I think about the opportunities there, obviously there's Sabre Mosaic. There's opportunities to upsell many of those modules as well. How should we think about the potential uplift of revenue per PB? I think some industry consultants put it at 15%, although that number can widely-
Yeah
has a pretty wide range. How should I think about it maybe into next year and then there's longer term potential with it?
I'll be a bit esoteric in terms of how I answer this. The first thing is the PB construct for software is a weird one. Charging per passenger instead of charging a license fee. We're trying to move with Sabre Mosaic more to a gain share or a license fee model. Some airlines are going to be wed to the PB construct, and that therefore is going to persist long term. What we've talked about is overall for our business, Victor, is our average booking fee we expect over this year, next year to stay relatively constant and our margin to be in the range of 57% overall for our combined business. We've not broken out margin or that sort of thing beyond those levels, I'm not going to do that today.
I do think that is there an opportunity for accretion in the unit revenue in Airline IT with new solutions? Absolutely.
Can you explain a bit, sorry, on the gain share model, how does that work?
Yeah. Gain share model would say, let's say we introduce something, our Ancillary IQ product which helps airlines better merchandise and sell ancillaries than they could do in the old world. We think there's a 2%-3% yield uplift, for example, when you do that. You take your baseline, however we in the airline determine the baseline is measured. We say any gains from that, we would take a component of that in pre-agreed methodology based on metrics, where if we don't deliver, we don't get paid. If we do deliver, we probably get paid more than we would in a license fee model. We think we're confident in our technology. All of Sabre Mosaic is built on top of Google's Vertex AI or now Gemini. It is the smartest technology that's out there. That's what a gain share looks like. Again, some airlines love that.
Some airlines want to know exactly what their cost is regardless, and they don't like that structure.
Very clear. Then conscious of time, maybe last question on Air IT. I think you also mentioned on call that you have more AI modules live or maybe rolled out to customers compared to Amadeus. Can you share some examples of where you see most demand in terms of the AI solutions that you have?
The entire offer suite within Sabre Mosaic is all AI infused. Everything we're doing from pricing, revenue management, all our IQ products, those are all built on Gemini, they all have that learning and that smarts in them. I don't know what Amadeus's capabilities are, but these were built cloud-native AI-first starting a couple of years ago, even before the big conversation about agentic. Again, if you speak to the airlines, they love what we've done from a technology standpoint. Now, if you're an Altéa customer, you're frustrated because it's not feasible for you to turn these on commercially yet, but we think that's going to change.
Very clear. If I change gears to talk a bit about maybe Constellation, has anything changed operationally or strategically since Constellation's involvement? Have they contributed so far and do they have more plans down the road to do more maybe?
Constellation now owns 12.7% of Sabre. With our cooperation agreement, they can take that up to 15%. Beyond that, they would need the consent of the company. Damian McKay is who's one of their division leaders on board, very engaged. They've been very supportive. The one thing I would say is this is only the second minority investment they've ever made. I think in their 20-year history, they've only sold one investment they've ever made. They're long-term holders of the business. They believe the business is intrinsically, dramatically undervalued. They do look at the balance sheet and they think that we have more debt and it's more expensive than we would like, and we agree with that. They have a desire to put more capital to work with Sabre.
Finding the intersection point that works for Constellation and its shareholders and Sabre and its shareholders, is something that we're talking about and I'm hopeful that there'll be a solution over time where they can contribute and help. So far so good.
That's good to hear. Maybe quick question on financials. Can you walk us through how you get to positive free cash flow in 2027 and how should we think about that evolution going into outer years as well?
Thank you for the question. First, if you look at as a starting point, 2026. For this year, our last guide, as of our last earnings call was for $585 million of adjusted EBITDA and negative $70 million of free cash flow. Of that negative $70 million of free cash flow, about $60 million is attributable to restructuring costs associated with our inflation offset program. Excluding that, we're near break even free cash flow for 2026. As we move from 2026 to 2027, I would start with the P&L. If you look at our top line, we expect air distribution bookings to be up somewhere in the mid-single digits. We've also indicated that we'd expect the booking fee to be largely in the range of where it is today.
Therefore, we would expect revenue to be up roughly in the mid-single digits as we go to next year. We've also indicated that, as Kurt just mentioned, we would expect the gross margin to be roughly in the same range of where it is now at 57%. I would expect gross income dollars to largely grow in kind with that mid-single-digit growth in revenue. At the same time, with the onset of our inflation offset program, we are intentionally targeting keeping our cost structure relatively flat, except for volume-related hosting costs. With that, as gross income dollars grow, we would expect a large proportion of that to flow down to adjusted EBITDA. Similarly, we would expect a large portion of that adjusted EBITDA to translate to free cash flow, and that will get you to a substantially higher positive free cash flow in 2027.
As we look beyond that, what I'd point you to is, as we look at our top line, and our growth strategies, all of our growth strategies have significant legs to them. We continue to take share. We're continuing to take the long tail of LCC Airlines, where we generate more bookings. Hotel B2B over the long term, we would expect to grow at least at the rate of air distribution bookings. As Kurt has highlighted and mentioned, we see the air IT business translating into a mid-single-digit revenue grower. At the same time, as we go forward beyond 2027, as we have been, you should expect us to be really strong cost managers.
The point is, as we grow revenue, the goal will be to have as much of that revenue translate into gross income, as much of that gross income to translate into higher EBITDA year-over-year, and for that to translate into free cash flow.
Very clear. We don't have a lot of time left, operator, or if anyone have any questions on the web, maybe you can flag the operator, and I will check again shortly before the call ends. Maybe continuing on what you said, when I think about AI, now shifting gears a bit, can you elaborate maybe how AI has been adopted internally? I would assume you're using a lot of Google tools. Do you see scope for maybe further cost optimization with the use of AI as well into this year and next?
Yeah. If you look this year, for example, we enacted our inflation offset program. There's really three components that really have driven that. Some is what I'd call continue to leverage best in class cost locations. The other is, where third parties could do something more efficiently, we intentionally are taking advantage of that. Thirdly, AI is really being pushed through the organization as a productivity driver. In our drive to keep our cost structure, I would say, roughly flat over the next couple of years, AI is a big component of that. What I would say is if there's continued advancements in AI, you should expect that we will take advantage of those to increase the productivity of the team. On things like product and development, we should look for us to increase the throughput of technology advancement.
Very clear. Operator, are there any questions on the web? If you can hear me. While we wait maybe for that, I do have a couple more questions back to distribution. Obviously, you have launched the LCC multi-source product earlier this year. Can you talk a bit about how meaningful it is right now? Are you taking share of Travelfusion, and how does the economics work there?
Preceding the launch of the new platform, we had 150+ low-cost carriers in Sabre, accounted for about 9% of our total distribution bookings. We launched that platform in January, which is the inclusion of a long tail of LCCs that do not participate in the GDS otherwise, and Travelfusion being the proxy for who was winning in that space. The difference versus Travelfusion is we've got the same content or better and full integration into all of our workflows and our shopping. From an agency perspective, it drives much better productivity and user experience than Travelfusion. We're seeing we've added 50+ new low-cost carriers with that. Overall, versus a year ago, LCCs represent about 150 basis points higher of our total air distribution versus what they did, again, at this point a year ago.
It's contributing to the air distribution growth. One of the interesting parts about this is typically an agency has one or potentially more than one GDS, and they'll use one source for that long tail of LCC content, Travelfusion, or there's some others. They're not going to put Sabre in and say, "Well, I'll use Sabre for these carriers and Travelfusion for those." It's more of a binary decision to displace Travelfusion. We are seeing agencies begin to do that, but they've built routines around Travelfusion or again, its proxies. Over time, we think this is going to be a very strong growth vehicle for us over the next couple of years, but the quality of the offering is resonating very well.
Very clear. I guess maybe last question from my end, probably addressing the other elephant in the room that we haven't talked much about, we managed not talk much about throughout the entire session is about NDC. I guess your NDC is scaling a bit on your end, how should we expect revenue per booking to trend over time as you scale NDC? I think there were some debates about maybe lower revenue per booking, but on the EBITDA level is the same. Is it still the same? Has expectation changed there? Can you tell us a bit more?
Two sides of NDC. One is NDC IT, which is on the airline side of the API, where we were late to the game. We're now winning and implementing customers in terms of their NDC IT solution. Victor, where you're asking about is NDC distribution.
Yeah
which is how do you pull from the API? We closed last year with about 4% of air distribution bookings being NDC. That's growing at about a 50% rate year on year. I think you'll see that trend continue. When you look at the economics of NDC, outside of Europe, the revenue per booking is basically a very nominal dilution versus a net-effect booking. The margin is a very nominal dilution versus net-effect booking, but de minimis. European point of sale is quite different. At effect, booking fees here are nearly double what they are in the rest of the world. Perhaps that's because of Amadeus' dominant position or something, I don't know. There wasn't a pricing reset here the way there was in the rest of the market over the last 25 years.
With NDC, there is a more material, both revenue and unit margin dilution for European point of sale. For Sabre, only about 16% of our air bookings derive in Europe, so we have relatively less exposure than our two GDS competitors there. When you look at that overall, you look at that globally, you would say revenue and unit margin are both diluted for NDC versus at effect, not in any sort of catastrophic sense. When you look at that NDC growing at a 50% rate, and we've asserted and we've reaffirmed that we expect average booking fee to stay relatively constant through this year or next year, and we expect margin to stay relatively constant, you have the negative effect of NDC.
On LCC, you have lower unit revenue, you have very high gross margin. You have the impact of non-air products that we're selling, hotels, payments, for example, which are high yield, high margin. Those basically offset the impact of NDC. On the whole, again, we expect that both our average booking fee and our margin are going to stay relatively constant, despite what we think is going to be strong growth in NDC.
Very clear. Unfortunately, we are running out of time. Thanks, Kurt and Mike, for joining us today. Thanks everyone for joining. Hope everyone have a good day.
Thank you.
Thank you.