Abstract illustration of a glowing, multi-layer data stack representing the layers of travel technology
All Posts

The Travel Data Stack: 11 Layers of Travel Tech in 2026

Ryan Turner
Ryan Turner · Head of Growth

Book a flight in ten seconds and you've touched a dozen systems without knowing it. A search box quietly polled a global distribution system built in the 1970s, a handful of airline-retail APIs, a flight-schedule feed covering almost every commercial route on earth, a hotel bedbank wholesaling 170,000 properties, a payments rail, and a fraud check. In 2025, Travel & Tourism contributed a record $11.6 trillion to global GDP, about 9.8% of the world economy (WTTC, 2026). Underneath that number sits a data stack most people never see. This teardown walks through all 11 layers, names the companies that define each one, and shows where the official feeds stop and live web data takes over.

Key Takeaways
  • Travel & Tourism hit a record $11.6 trillion in global GDP in 2025, roughly 9.8% of the world economy (WTTC, 2026).
  • Three GDS incumbents (Amadeus, Sabre, Travelport) still route an estimated 95%+ of GDS bookings, but NDC transactions grew about 165% year over year into 2024 (Accelya, 2024).
  • Travel is the single most bot-targeted industry online: roughly 48% of traffic to travel sites in 2024 was bad bots scraping fares and inventory (Imperva, 2025).
  • The layers being digitized right now (experiences, eSIM, ancillary checkout) are where the next decade of travel-tech building happens.
The whole pie keeps growing TRAVEL & TOURISM GDP CONTRIBUTION (USD TRILLIONS) $11.6T $16.5T 2025 (actual) 2035 (forecast)

Source: World Travel & Tourism Council, 2026.

1. Global Distribution Systems: the legacy backbone

As of 2026, three companies still sit under most of the world's air travel: Amadeus, Sabre, and Travelport collectively route an estimated 95% or more of all GDS bookings (Future Market Insights / Business Travel News Europe, 2025). A GDS is the switchboard travel agencies and online sellers use to shop and book air, hotel, and car content in one place. These are the oldest systems in the stack, and they're sticky precisely because everyone already connects to them.

The defining players here are Amadeus, Travelport, Sabre, and the fast-growing TBO out of India. The catch for builders: GDS access is gated, contractually heavy, and priced per segment. The interesting tension in 2026 is that the GDS does not actually hold every fare. Airlines increasingly push their richest offers somewhere else, which is the whole reason the next layer exists.

For the full company-by-company view of this ecosystem, see our travel data industry market map.

2. NDC and airline retailing: the modern offer layer

Airlines spent the last few years moving their best fares off the GDS and into New Distribution Capability (NDC) channels. The momentum is real: NDC transactions grew roughly 165% year over year heading into 2024, and Accelya, which says it processes more than half of the global NDC market, generates 30 billion-plus unique offers a day (Accelya, 2024). Industry estimates put NDC at roughly a fifth of bookings in 2024, up from about 10% a year earlier (TWAI, 2025).

This is the busiest, most crowded layer on the map. Duffel, Travelfusion, AirGateway, Verteil, TPConnects, PKFARE, Atlas (ATRIP), Mystifly, Kyte, and a dozen more aggregators all promise the same thing: one clean API that hides the mess of dozens of airline NDC connections behind a single integration. American Airlines made about 40% of its fares NDC-exclusive, so a seller without NDC access simply can't see those prices. The aggregators exist because rolling your own connection to every carrier is a full-time job no startup wants.

3. Flight data: schedules, status, and fares

Booking is one thing; knowing what's actually flying is another. Reference flight data is its own dense layer, and the scale is wild. As of 2025, Cirium processes around 300 terabytes of aviation data a day from roughly 2,000 sources, with schedule coverage spanning 900+ airlines and about 99% of global commercial flights (Cirium via AltexSoft, 2025). That's the feed behind the "your flight is delayed" alert and the route maps in every booking tool.

The names split by purpose. OAG and Cirium own authoritative schedules and connections data. FlightAware and the OpenSky Network track live positions. Aviationstack, Aviation Edge, and AeroDataBox sell developer-friendly REST endpoints for status and routes. Traveldax and Going sit closer to fare intelligence and deal detection. Notice the pattern: the higher-authority the data, the more it lives behind enterprise contracts, while the developer-tier APIs trade completeness for a credit card and a quickstart.

4. Hotel and inventory: the bedbank wholesale layer

Hotels reach the open market through bedbanks, the wholesale distributors that aggregate rooms and resell them to travel sellers. The anchor number here is concrete: HBX Group (Hotelbeds) gives the travel trade access to about 170,000 unique hotels across 180+ countries, distributing to tens of thousands of B2B buyers (HBX Group, 2025). HBX listing on the Madrid exchange in early 2025 told you how much money sits in this quiet middle layer.

The defining set is broad: Hotelbeds (HBX), WebBeds, RateHawk, World2Meet, Expedia Rapid, the Booking.com partner API, plus connectivity-focused players like TravelgateX, Stuba, Yalago, and G2 Travel. The recurring headache for anyone building on this layer is that rates and availability shift constantly and differ by source market. The same room is priced differently depending on where the shopper appears to be, which makes accurate hotel pricing a live-data problem, not a static-feed one.

5. Specialty data: ancillaries, disruption, CO2, and interlining

Beyond the core feeds sits a layer of specialty data that the big systems handle poorly. As travelers and regulators push on emissions, and as airlines chase add-on revenue, this niche data gets valuable fast. Ancillaries alone are now a giant business: airlines worldwide were projected to generate a record $157 billion in ancillary revenue in 2025, up from $148.4 billion in 2024 and now about 15.7% of total airline revenue, versus just 9.1% in 2016 (IdeaWorksCompany and CarTrawler, 2025).

The companies here are sharp and specialized. TripStack handles virtual interlining, stitching separate flights into one itinerary the airlines never sold together. FlightRadar24's API exposes its tracking firehose. IATA CO2 Connect standardizes per-passenger emissions math. Wenrix works disruption and re-accommodation data. If you've ever seen a self-transfer itinerary that no single airline offers, you've used virtual interlining, and the connection logic behind it is some of the hardest data engineering in travel.

6. Travel APIs and connectivity hubs: the glue

Every layer above has to talk to every layer below it, and that integration work is its own product category. Connectivity hubs and broad travel APIs exist to collapse many messy connections into one. The value proposition is boring and enormous: write one integration, reach hundreds of suppliers. The reason this layer keeps attracting investment is that integration debt never stops growing as the stack adds new NDC carriers, new bedbanks, and new ancillary feeds.

The map flags Dida and Hopper as connectivity-and-API players, with Hopper in particular having spun its consumer pricing-prediction smarts into a B2B engine (Hopper Technology Solutions) that other travel brands embed. This layer is thinner on names than the aggregator layer above it, but it's structurally important: it's the connective tissue that lets a small team behave like it integrated with the whole industry. Why rebuild the same supplier connections every competitor already has?

7. Metasearch and partner APIs: the comparison engine

Metasearch is where shopping happens before booking does. These engines scan dozens or hundreds of suppliers and route the traveler to whoever has the best price. Skyscanner alone scans 1,200+ travel providers in a single query (Skyscanner, 2026). Metasearch doesn't usually sell the ticket; it sells the comparison and hands off the click, which makes its entire value depend on having fresh, broad, accurate price data.

The defining players are Skyscanner, Kiwi.com (and its Tequila API), Trip.com, Traveloka, MakeMyTrip, Google Flights, and award-search specialist Point.me. This layer faces the most existential pressure in the stack: Skift devoted a February 2026 analysis to whether AI assistants will replace metasearch entirely as travelers start asking a chatbot to find the trip (Skift, 2026). Whether the front end is a comparison grid or a conversation, the data demand underneath is identical, and it's only getting heavier.

For how teams pull competitor prices at scale, see our 2026 anti-bot playbook for retail data.

8. Consumer platforms: the demand surface

This is the layer travelers actually recognize, and it's where the gross bookings show up. Online travel agencies represented roughly $794 billion in gross bookings, with the OTA segment generating about $94 billion in revenue in 2024 and forecast to reach roughly $107 billion by 2026 (Skift Research, 2025). These are the brands with the marketing budgets, the loyalty programs, and the app installs.

The list is long and familiar: Expedia, Booking brands like Priceline and Agoda, Kayak, Hotels.com, Airbnb, VRBO, Trivago, TripAdvisor, momondo, and a deep bench of specialists from Skiplagged to Alternative Airlines. Every one of these platforms watches the others. Price parity, availability, and ranking all depend on knowing what competitors show, which is why this layer is one of the heaviest consumers of live web data in the entire industry.

9. Travel experience apps: the layer still being built

Flights and hotels are largely digitized; what you do once you arrive is not. The experiences sector reached about $271 billion in 2025 and is projected to hit roughly $342 billion by 2029, yet online penetration was just 33% of gross bookings in 2025, versus 64% for travel overall (Arival and Phocuswright, 2025). Two-thirds of the money still books offline. That gap is the opportunity.

Experiences are still going digital SHARE OF EXPERIENCES BOOKED ONLINE All travel: 64% 17% 33% 42% 2019 2025 2029 (forecast)

Source: Arival and Phocuswright, Global Market Sizing, 2025-2026.

The players chasing it are energetic: GetYourGuide, Viator, Klook, and Headout on the tours-and-activities side; TripIt, Wanderlog, and Stippl on trip planning; and Atlas Obscura for the long tail of unusual things to do. The penetration gap is the single clearest "build here" signal in this whole teardown: a $342 billion market that's only a third online is rare in 2026.

10. Checkout and identity: payments, upsell, and fraud

Once a traveler decides, the transaction has to clear, and travel payments are unusually hard. High ticket prices, cross-border currencies, chargeback risk, and aggressive fraud all collide at checkout. This is also where airlines and sellers capture that $157 billion in ancillary revenue, since the upsell moment (seats, bags, insurance, upgrades) happens right here at the point of sale (IdeaWorksCompany and CarTrawler, 2025).

The map's checkout layer includes Plusgrade (upgrade and upsell monetization), Cell Point Digital (payment orchestration), Hopper Technology Solutions (fintech products like price-freeze and cancel-for-any-reason), and installment players like Flex Pay and Movmo. In our experience, anyone who's built a travel checkout knows the quiet truth here: the payment and fraud stack is often more engineering work than the booking flow it sits on top of, because a single mispriced or fraudulent high-value booking erases the margin on dozens of clean ones.

11. Travel eSIM: connectivity for the trip itself

The newest layer barely existed five years ago. Travel eSIM apps let a traveler land in a new country and get data without swapping a physical SIM, and growth has been steep. Analysts disagree on the exact size, but multiple forecasts put the travel eSIM market's growth above 60% CAGR through the early 2030s (DataM Intelligence, 2025). The direction is unanimous even where the numbers aren't.

The category leaders are Airalo (the largest), Holafly, Saily, aloSIM, and Roamless. What makes eSIM a data-stack story and not just a telecom one is the back end: real-time coverage, pricing, and plan availability vary by country and carrier, so these apps live or die on accurate, geo-specific data about networks they don't own. It's a fitting top to the stack, because it shows the same pattern as every layer below: the product is an app, but the moat is the data feeding it.

The travel data stack at a glance

# Layer What it does Defining players
1 Global Distribution Systems Legacy booking switchboard for air, hotel, car Amadeus, Sabre, Travelport, TBO
2 NDC / Airline Retail Aggregates airlines' modern, GDS-exclusive offers Duffel, Verteil, Mystifly, AirGateway, Kyte
3 Flight Data Schedules, status, positions, fare signals OAG, Cirium, FlightAware, Aviationstack
4 Hotel & Inventory Bedbank wholesale of rooms to sellers Hotelbeds (HBX), WebBeds, RateHawk, Expedia Rapid
5 Specialty Data Ancillaries, disruption, CO2, interlining TripStack, FlightRadar24 API, IATA CO2 Connect
6 Connectivity Hubs One integration to reach many suppliers Dida, Hopper
7 Metasearch Compares suppliers, routes the click Skyscanner, Kiwi.com, Google Flights, Trivago
8 Consumer Platforms The booking brands travelers know Expedia, Booking, Airbnb, Kayak, Agoda
9 Experience Apps Tours, activities, and trip planning GetYourGuide, Viator, Klook, TripIt
10 Checkout & Identity Payments, upsell, and fraud at point of sale Plusgrade, Cell Point Digital, Flex Pay
11 Travel eSIM In-trip connectivity for travelers Airalo, Holafly, Saily, Roamless

How we mapped this stack

This teardown is built from a Q1 2026 market map of the travel data industry, cross-checked against primary and trade sources for every statistic. The 11 layers are ordered from the foundation of the stack (the GDS backbone that predates the internet) up to the newest consumer-facing layer (travel eSIM), following the path a single booking takes from supply to traveler.

A few ground rules shaped the list. Every layer had to represent a distinct job in the stack, not a marketing category, which is why payments and connectivity earn their own entries even though they're less glamorous than booking. Every named company comes from the source market map, and every figure is attributed to its publisher with a retrieval date in the sources below. Where market-size estimates were shaky or came from low-authority aggregators (notably eSIM sizing and the precise GDS share split), the language stays directional rather than asserting a single number as fact.

The data layer running underneath all 11

Here's the thread that connects every layer above. Travel data is valuable, perishable, priced differently in every country, and defended harder than almost anywhere else online. In 2024, travel was the single most bot-targeted industry on the web: roughly 48% of all traffic to travel sites was bad bots, against about 47% human and 5% good bots, as automated traffic crossed 51% of all web traffic for the first time (Imperva, 2025).

Travel is the web's most-targeted industry SHARE OF TRAVEL-SITE TRAFFIC, 2024 48% bad-bot traffic Bad bots: 48% Human: 47% Good bots: 5%

Source: Imperva 2025 Bad Bot Report, 2025.

That statistic explains the whole stack. In our work with travel and commerce teams that collect public pricing data, the same wall shows up at nearly every layer: the data you need is geo-specific, and the site you need it from would rather you didn't have it. Metasearch needs fresh competitor prices. Meanwhile, OTAs police parity across rivals. Bedbanks and eSIM apps serve different rates by source market. All of it depends on collecting accurate, geo-specific public-web data from the right country, and travel sites push back hard against traffic that doesn't look like a real local shopper.

This is the gap the official APIs leave open. A GDS feed won't tell you what your competitor is charging a shopper in Brazil; a bedbank rate won't reveal the public price on a rival's German site. Teams that need that view turn to infrastructure built for it. Massive, for instance, is a device-access network plus rendering stack that returns clean HTML or markdown from any public source in any location. It uses real consumer devices in 195+ countries, so a price check from São Paulo or Berlin reads as a genuine local visit, not blocked data-center traffic. The point isn't any one vendor. It's that live, geo-accurate web data has quietly become the twelfth layer of the stack, the one that fills the gaps the other eleven leave behind.

More on why the origin of a request matters for accurate, geo-specific data: residential vs datacenter proxies.

Conclusion: eleven layers, one data problem

The travel data stack looks like eleven separate businesses, but it rhymes from top to bottom. Every layer, from the 1970s GDS to a 2026 eSIM app, is really a data problem wearing a different interface. The companies that win each layer are the ones with the cleanest, freshest, most geo-accurate view of supply and prices.

Three things are worth watching as 2026 unfolds: NDC eating more of the GDS's share, AI assistants pressuring the metasearch layer, and experiences finally going digital. If you're building in travel, pick the layer with the widest data gap (experiences and eSIM lead that list) and solve the live-data problem underneath it before you polish the interface on top.

If you're building on this layer, start with our guide to giving AI agents live web access.

Sources

Frequently Asked Questions

The travel data stack is the set of connected systems that move travel from supply to traveler: distribution systems, airline-retail and NDC APIs, flight and hotel data feeds, metasearch, consumer booking platforms, payments, and connectivity. Travel & Tourism contributed $11.6 trillion to global GDP in 2025 (WTTC, 2026), and this stack runs underneath all of it.

A GDS (Amadeus, Sabre, Travelport) is the legacy switchboard that aggregates content for travel sellers, and the big three route an estimated 95%+ of GDS bookings. NDC (New Distribution Capability) is the newer standard airlines use to sell richer, often GDS-exclusive offers directly. NDC transactions grew about 165% year over year into 2024 (Accelya, 2024).

Travel prices are high-value, change constantly, and vary by country, so competitors, resellers, and fraud operators all want the data. In 2024, roughly 48% of traffic to travel sites was bad bots, the highest share of any industry (Imperva, 2025). That pressure shapes how every layer of the stack defends and collects data.

Travel experiences stand out. The sector reached about $271 billion in 2025 and is heading toward $342 billion by 2029, yet only 33% of bookings happen online versus 64% for travel overall (Arival and Phocuswright, 2025). A large, fast-growing market that's still mostly offline is rare.