Amadeus IT Group S.A. (BME: AMS) — Embedded Systems, Exposed Distribution

Published: 2026-09-12 · Verdict: Accumulate · Research confidence: High (82%)

Executive conclusion

Analyst Take

At the September 11, 2026 close of €54.18, I rate Amadeus Accumulate, using a staged entry range of approximately €48–56 and medium investment conviction. This is a price-conscious judgment, not a claim that the stock is exceptionally cheap. With approximately 419.99 million economic shares outstanding and June company-defined net financial debt of €2.58 billion, the price implies equity value of €22.76 billion, enterprise value near €25.33 billion and approximately 10.0× trailing EBITDA. Management’s unchanged €1.35–1.45 billion 2026 free-cash-flow objective equates to a 5.9%–6.4% equity yield [S1][S2][S7][S17][S18].

The central thesis is that investors increasingly value Amadeus as a travel intermediary exposed to AI and direct distribution, while a substantial portion of its value resides in operational infrastructure that is much harder to displace. Airline reservations, inventory, offer creation, order management, servicing, departure control, airport processing and partner interfaces form an interconnected operating stack. Replacing that stack is a multi-year transformation carrying revenue, service and operational risk for the airline. Ryanair’s 25-year Navitaire relationship, TUI’s decision to adopt Stratos after more than two decades as a customer, and the first production Orders at Finnair and Saudia are evidence of relevance and customer tenure [S4][S5]. They are not proof of permanent lock-in: modular architectures and open interfaces also make it possible to procure individual functions from specialists.

The variant view is therefore narrower than a generic moat claim. Air Distribution bookings fell 3.7% in H1 and 7.6% in Q2, yet constant-currency Distribution revenue still grew 1.1%; Air IT and Hospitality revenue increased 8.7% and 9.2%, respectively [S1]. That demonstrates revenue resilience, but not necessarily market-share gains or clean pricing power. Revenue per booking and per passenger includes renewals, inflation, implementation revenue, professional services, airports, product mix and disruption-related transactions. Sabre’s Q2 air direct billable bookings grew 0.7% while Amadeus’s differently defined agency air bookings fell 7.6%, an adverse comparison that cannot be dismissed even though geography and denominators differ [S1][S23].

The strongest counter-case combines operating and allocation concerns. First, management reduced its assumed 2026 origin-and-destination passenger growth from 4.4% to 1.9% and cut revenue and adjusted-EPS guidance in July [S7]. Second, €756 million, or about 52% of 2025 gross R&D, was capitalized; EBITDA therefore does not capture the amortization of current development investment, and even EBIT spreads the expense over future periods [S3]. Third, Amadeus repurchased €1.3 billion of stock during 2025 at an estimated €68.68 per share—about 27% above the current price—before signing an agreement to pay €1.2 billion plus up to €150 million contingent consideration for IDEMIA Public Security [S2][S3][S19]. Fourth, AI could leave Amadeus as essential transaction plumbing while shifting discovery, customer ownership and bargaining power toward consumer platforms.

Evidence quality is high for reported financial statements, debt, share count, transaction consideration and operating KPIs because these reconcile to audited or regulator-filed documents. It is medium for competitive share, product uplift, addressable markets and acquisition synergies because those measures are management or vendor estimates. The factor model snapshot is unavailable, so no quantitative beta, factor loading, alpha or crowding claim is made. Qualitative travel, currency and duration sensitivities are economic hypotheses, not factor-model outputs.

The next decision sequence is concrete. Distribution must first stabilize relative to comparable market and peer measures. Second, the maintained free-cash-flow range must survive weak traffic, higher interest and tax payments and implementation spending. Third, Marriott, Accor and Nevio milestones must convert from engagements and installed properties into recurring contribution and cash payback. Fourth, IDEMIA must close without material financing, regulatory or integration deterioration. I would lower the call if Distribution materially underperformed matched market growth for four consecutive quarters, 2026 free cash flow fell below €1.30 billion without a precisely reversible timing item, or pro-forma leverage exceeded roughly 1.75×. I would raise conviction if booking share stabilized, organic free cash flow exceeded €1.45 billion without working-capital assistance and implementation cohorts demonstrated attractive returns after all development and migration costs.

Stock Price Action — Five-Year Event Map

Amadeus closed at €54.18 on September 11, 2026. The unadjusted series began at €52.00 on September 13, 2021, implying a five-year price return of only 4.2% before dividends despite a much larger earnings recovery. The five-year intraday low was €44.85 on September 28, 2022, and the high was €75.40 on March 3, 2025. The latest 52-week range was €46.21–69.30. The current price is 21.8% below that 52-week high, 17.3% above the low and approximately one-third of the way through the range. It is 27.9% below the five-year high and down 13.6% from the first close of 2026 [S17].

The event map rejects two simple conclusions. First, recovering earnings did not guarantee a durable rerating: IFRS operating income moved from an €83 million loss in 2021 to €1.76 billion in 2025 while the five-year price gain remained minimal [S13]. Second, the 2026 drawdown does not by itself prove structural disintermediation. The company simultaneously faced Middle East disruption, energy-related traffic pressure, Hawaiian’s de-migration and Spirit’s cessation. Only matched booking-share, contribution and cash-flow evidence can separate a cyclical shock from a structural break.

Verdict: The stock has surrendered most of its recovery premium but does not trade at a distressed valuation. The recovery from February’s low occurred before bookings clearly stabilized, leaving meaningful downside if weak Distribution volume proves structural rather than cyclical [S1][S17].

Business Overview

Amadeus is best understood as three connected businesses rather than one homogeneous software company. Air Distribution connects airlines and other travel suppliers with agencies, travel-management companies and online sellers. Air IT Solutions operates reservation, inventory, departure-control, offer, order, digital, revenue-management, airport and related airline systems. Hospitality and Other Solutions includes central reservations, hotel distribution, property and service applications, airport technology, payments, biometrics and adjacent travel applications. Shared infrastructure, data standards and customer relationships permit cross-selling, but the three businesses have different economics and competitive risks [S1][S4][S5].

In 2025, group revenue was €6.517 billion. Air Distribution contributed €3.119 billion, Air IT €2.346 billion and Hospitality and Other Solutions €1.052 billion. Air Distribution therefore remained the largest revenue segment, but Air IT had the highest H1 2026 contribution margin at 70.8%, compared with 51.9% for Distribution and 34.9% for Hospitality. Contribution excludes indirect costs and depreciation, so it is not segment operating profit; it is still useful for comparing the incremental economics of the portfolios [S1][S4].

Distribution is a two-sided transaction and workflow network. Travel sellers value broad content, comparison, itinerary construction, mid- and back-office integration, payments and post-booking servicing. Airlines value access to agencies and corporate travelers, especially for complex international itineraries and higher-yield business demand. Amadeus earns transaction-related fees and associated software and service revenue. The network can reinforce itself because broader useful content attracts sellers while seller reach attracts suppliers. It is not an all-channel monopoly: airline websites, direct APIs, NDC aggregators, competing GDSs and large online agencies all constrain economics [S5][S22].

Air IT is closer to mission-critical enterprise software, although much of its revenue still varies with passengers boarded and transaction activity. Its systems can control reservations, seat and product inventory, ticketing, check-in, departure control, re-accommodation, offer construction, order servicing and airport interfaces. The customer value is not merely software functionality. It includes uptime, correct inventory, reliable recovery during disruption, integration with partners, regulatory compliance and the ability to launch products without rebuilding the core. Failure can interrupt an airline’s principal revenue engine.

Hospitality is broader and less economically mature. Its central-reservation and hotel-distribution products can become deeply embedded, but implementations require substantial professional services and customer-specific development. Payments carry transaction volume and regulatory exposure. Airports and biometric systems may include hardware, integration and long procurement cycles. H1 Hospitality’s 34.9% contribution margin was less than half Air IT’s, showing why contracted properties or revenue growth cannot be equated automatically with group value creation [S1].

Revenue-stability answer: Amadeus has durable contracts and embedded workflows, but its revenue is not a fixed subscription annuity: Distribution and parts of Air IT remain exposed to bookings, passengers, cancellations, airline failures, customer migrations and implementation timing. H1 2026 included €3.100 billion of platform and software revenue and €235 million of professional services and other activities, but platform revenue itself contains transactional usage [S2]. Management describes revenue as highly recurring, yet no current audited percentage cleanly separates subscriptions, transactions, maintenance and implementation services.

The distinction matters in a downturn. A long-term contract may preserve the relationship while billable passengers or bookings decline. Conversely, implementations, renewals, disruption transactions and product adoption can lift revenue while volume is weak. Hawaiian’s de-migration and Spirit’s cessation reduced North American passengers; ANA’s domestic implementation increased them. These examples demonstrate that traffic, customer perimeter and product mix interact [S1].

Business-understandability answer: The business is understandable at the driver level—bookings, passengers boarded, revenue per transaction, implementations, development spending and customer wins—but blended segment ratios prevent a clean separation of price, mix and non-volume services. H1 Distribution bookings fell 3.7% while constant-currency revenue grew 1.1% and revenue per booking rose 5.1%. Air IT passengers increased 1.1%, revenue increased 8.7% and revenue per passenger rose 7.5% [S1]. The transcript explains that these ratios include renewals, inflation, upselling, Nevio, airport and professional services and disruption activity [S11].

Customer contracts create durability but also obligations. Amadeus must maintain functionality, security, reliability and regulatory compatibility over long periods. Customer migrations consume implementation resources before recurring revenue matures. Approximately 30% of H1 R&D was allocated to implementations, including Marriott, Accor, Nevio and NDC customers [S1]. The business therefore resembles a transaction platform funded by continuing product and migration investment, not a maintenance-only installed base.

Security-tax-status answer: The security is an ordinary euro-denominated share listed on the Spanish Stock Exchange under BME:AMS, ISIN ES0109067019; it is not an ADR, MLP, partnership or K-1 issuer. The tax consequences for a non-Spanish holder, including dividend withholding and treaty relief, depend on the investor’s jurisdiction and should not be inferred from the listing form [S2][S13].

The most valuable internally generated assets are not fully recorded. They include customer relationships developed organically, knowledge of airline operations, implementation experience, connectivity with airlines and sellers, operating data, algorithms, trained personnel and a reputation for reliable processing. Yet Amadeus is not economically asset-light in the simplistic sense. At December 2025 it carried €3.912 billion of goodwill and €4.344 billion of other intangible assets. Internally developed software additions were €740 million during 2025 [S3][S14].

Unrecognized-assets answer: The balance sheet understates internally generated relationships, implementation knowledge, network connectivity, operating data and reputation, but it already recognizes substantial goodwill and technology assets; acquisition capital and capitalized development must therefore remain in return calculations. Removing goodwill or internally developed software from invested capital while retaining their earnings would make customer acquisition and product creation appear costless.

Amadeus also serves as coordination infrastructure. NDC changes message formats and gives airlines greater control over offers, but agencies still need multi-airline comparison, itinerary assembly, payment, changes, refunds and disruption handling. AI can change how demand is discovered without eliminating execution or servicing. The economic question is not whether the interface changes; it is whether Amadeus continues to monetize normalization and execution after the interface changes.

Verdict: Amadeus combines embedded operating software with material transaction and implementation exposure. Its revenue quality is better than that of a travel agency but less defensive than a prepaid subscription vendor. The clearest disconfirming evidence is the simultaneous booking decline, customer de-migration and airline failure in H1 2026, proving that long relationships do not eliminate volume and perimeter risk [S1].

Industry Dynamics

The relevant industry comprises the technology and transaction infrastructure used to sell, operate and service travel. Its profit pools include airline distribution, passenger-service and order systems, hotel reservations and property systems, airport processing, payments, identity and government border technology. Labeling Amadeus simply as information-technology services obscures the actual competitive boundaries.

Demand is global. H1 passengers boarded were distributed across Asia-Pacific at 33.7%, Western Europe at 31.2%, North America at 15.5%, Middle East and Africa at 8.3%, Central, Eastern and Southern Europe at 5.8% and Latin America at 5.4% [S1]. Geographic diversification reduces reliance on a single economy but introduces currency, conflict, sanctions, regulatory and carrier-health exposure. It also means aggregate IATA traffic is an imperfect benchmark because Amadeus’s customer and regional mix differs from global traffic.

Market-growth-geography answer: Demand is international and has a credible long-run growth tailwind, but 2026 is a weak and geographically uneven cycle, and IATA revenue-passenger-kilometers are not identical to Amadeus bookings or passengers boarded. IATA’s central long-term scenario projects global RPK growth of 3.1% annually from 2024 to 2050, led by Asia-Pacific at 3.8% and Africa at 3.6% [S8]. Its June 2026 outlook reduced current-year growth to 2.1%, while July RPK was only 0.2% above the prior year; the Middle East declined 10.0%, North America 1.2%, while Latin America grew 6.1% [S9][S10].

The immediate shock is supply-related as well as demand-related. IATA attributes the slowdown to Middle East airspace disruption, jet-fuel constraints, longer routings and weaker real income. It projects only a 2% airline-industry net margin in 2026 [S9]. Weak airline profitability matters because financially pressured carriers cut capacity, delay projects, fail or use procurement leverage more aggressively. Amadeus’s customers’ economics therefore influence its own transaction and implementation outlook.

Management estimates a combined addressable market around €50 billion, with hospitality previously estimated at €13.6 billion [S6][S20]. These are management estimates, not independent forecasts. The total includes categories already served by rivals or internal systems, some categories overlap, and implementation capacity constrains capture. A large nominal market does not establish that incremental capital earns an attractive return.

Industry-profitability-and-barriers answer: The industry supports high contribution margins where a platform combines global content, embedded operations and high transaction density, but implementation labor, development intensity, customer bargaining power and modular competition prevent uniform software-like economics. Amadeus’s H1 contribution margins ranged from 34.9% in Hospitality to 70.8% in Air IT [S1]. Sabre’s Q2 normalized operating margin was about 13% and its free cash flow only $10 million, illustrating how leverage, incentives and portfolio position can absorb gross transaction economics [S23].

Barriers to entry include reliability at very high transaction volumes, airline and agency integrations, content breadth, regulatory and security certification, migration references, servicing depth and the need to support multiple protocols. Amadeus reports three billion flight searches per day, peak capacity of 150,000 transactions per second and at least 99.95% availability across most services [S5]. Those figures are company-reported scale indicators. They estimate replication difficulty; they do not prove monopoly economics because searches can be unproductive, standards are accessible to rivals and a specialist can outperform in one module.

The competitive set changes by profit pool. Sabre and Travelport are the principal global distribution competitors. TravelSky has a large China-centered aviation, airport, distribution and settlement ecosystem, but related-party airline relationships and domestic concentration limit direct comparability [S22][S24]. Airline IT competition includes Sabre, IBS Software, specialist revenue-management and retailing vendors and customer-developed systems. Hospitality faces Oracle and specialist reservation and property-management platforms; Sabre sold its SynXis hospitality business to TPG in July 2025, so Sabre’s current group valuation no longer includes that asset [S27][S34].

Competition-direction answer: Competition is becoming more modular and technologically open, even as operational complexity preserves a role for large aggregators and system-of-record providers. NDC enables richer airline-controlled offers. Cloud interfaces and APIs make specialist modules easier to connect. AI agents may shift discovery toward consumer platforms. Conversely, different NDC versions, servicing rules, payment requirements and high look-to-book ratios create additional normalization and orchestration work [S5][S22][S25].

At year-end 2025, Amadeus reported that 35 airlines’ NDC content was being consumed by 50,000 sellers across 168 markets, while more than 50 airline IT customers used its NDC APIs [S5]. Travelport’s live NDC matrix demonstrates that such capabilities are competitive table stakes rather than an Amadeus-exclusive asset [S25]. The adoption evidence supports continued relevance but does not reveal comparative economics, incentives or channel share.

The supply-side capital cycle is unusual. Writing a software module has limited physical cost, but building a globally credible airline platform requires years of development, migrations, cybersecurity work, infrastructure and reference customers. That slows full-stack entry. At the same time, standardized interfaces reduce the capital needed to contest individual modules. The likely outcome is persistent module-level competition rather than a sudden proliferation of new global GDS networks.

Foreign-low-cost-threat answer: Lower-cost engineering can pressure development and implementation pricing, but it cannot by itself reproduce global integration, operating reliability, migration experience, content relationships and regulatory acceptance. IBS Software is credible evidence that an India-origin vendor can compete in passenger-service, offer, order and departure-control functions [S26]. TravelSky demonstrates regional scale in China [S24]. These are substantive threats, but wage arbitrage alone does not recreate the full network.

Regulation is fragmented across data protection, cybersecurity, payments, consumer protection, competition law, border control, sanctions and sovereign procurement. IDEMIA Public Security would increase exposure to government identity contracts, foreign-investment approvals, biometric fairness, sensitive data and fixed-price project execution. Its reported business mix includes border systems, law enforcement and access control, with several fixed-price or managed-service models [S20]. This is a materially different risk set from booking transactions.

Verdict: Long-run travel growth and difficult migrations support attractive industry economics, particularly in core airline IT, but competition is increasing around interfaces and modules. The disconfirming evidence against a broad moat claim is the combination of credible specialist systems, multiple NDC aggregators, direct airline distribution and low current airline profitability [S9][S22][S26].

Competitive Position

Amadeus’s strongest position is in airline operational technology; its most contestable position is air distribution. Management cites approximately 46% of global passengers boarded and more than 50% of GDS air bookings [S6]. The first statistic indicates broad passenger-service-system scale. The second is explicitly a channel-share measure: GDS bookings are only part of total airline distribution. It cannot be presented as more than half of all airline sales.

The Air IT moat is generated by integration depth, operating risk and accumulated implementation knowledge. Replacing a core platform can touch schedules, inventory, reservations, ticketing, pricing, check-in, departure control, disruption recovery and partner connectivity. The customer must fund testing, training, data migration, contingency capacity and staff change, while accepting potential revenue and service interruption. Long relationships at Ryanair and TUI support this mechanism [S4][S5]. The expected financial outcome is high retention, stable passenger share and contribution margins above less embedded businesses.

Switching-costs answer: Airline switching costs are high because migrations touch inventory, reservations, servicing, departure control and partner interfaces, but modular order architecture lets customers replace or procure individual components and therefore prevents absolute lock-in. Nevio’s open architecture creates the tension directly: it may deepen Amadeus’s position when customers adopt the full stack, but it also legitimizes multi-vendor architectures.

Distribution has a different advantage. Agencies benefit from broad content, comparison, payment and servicing inside established workflows. Airlines gain access to corporate and international demand. Network density is hard to reproduce, but its toll can be pressured by airline-direct strategies, NDC commercial terms, concentrated online agencies and rival aggregators. The return of some low-cost carriers to indirect distribution also shows that direct selling is not a one-way substitution; carriers may value agency reach even while resisting distribution cost [S5][S22].

Nature-of-competition answer: Competition occurs through content access, transaction economics, incentives, implementation capability, uptime, servicing quality and control of the customer interface—not simply through software feature lists. Sabre identifies direct supplier distribution, aggregators, metasearch, specialist technology providers and customers’ internal systems alongside conventional GDS competitors [S22]. This wider boundary is essential for understanding AI.

AI can affect different layers differently. A conversational agent may own discovery and recommendation while still using Amadeus for inventory, price validation, order creation, payment and servicing. In that outcome, Amadeus remains technically relevant but could lose bargaining power to the interface owner. A more severe scenario is direct integration between large agents and airlines that bypasses Amadeus execution. Management says commercial contracts cap unpaid query burdens and technical tools reduce unproductive requests, but this is a management claim that requires transaction and contribution data [S11].

Amadeus’s Google relationship illustrates both opportunity and dependency. The parties are deploying part of Amadeus’s platform on Google Cloud and exploring links involving MetaConnect, Nevio, Google Flights and Google offer management. The infrastructure commitment is concrete; future agent integration remains exploratory [S28]. Google can be a distribution partner, technology supplier and bargaining-power threat simultaneously.

Brand-relevance answer: Amadeus’s brand matters economically to enterprise buyers as a signal of reliability, implementation competence and counterparty durability, but it has little demonstrated consumer pull and cannot protect an inferior product from specialist competition. Airline customer NPS fell five points to +24 in 2025, while airport NPS fell more sharply, disconfirming any assumption that institutional reputation is automatically strengthening [S5].

Current operating evidence is mixed. H1 Air IT revenue rose 8.7% at constant currency despite passengers increasing only 1.1%; ANA’s implementation helped while Hawaiian’s de-migration and Spirit’s cessation hurt. Distribution revenue grew 1.1% while bookings fell 3.7%. Hospitality rose 9.2% [S1]. The figures demonstrate product and mix resilience but do not isolate market share.

The closest adverse cross-check is Sabre. In Q2 2026, Sabre’s air direct billable bookings increased 0.7%, Marketplace revenue increased 6% and the average booking fee increased approximately 4%. Amadeus’s agency air bookings fell 7.6% [S1][S23]. These are not matched statistics: Sabre and Amadeus differ by geography, customer mix, cancellations and booking definition. Nonetheless, the direction gap is large enough to make competitive share an open question rather than a macro conclusion.

NDC is both a threat and a possible extension of the moat. It reduces reliance on legacy messaging but increases versioning, servicing and search-normalization work. Amadeus reports that an Air France–KLM pilot using Advanced Airline Profile reduced unproductive shopping requests by 74% [S5]. That is a company-reported pilot outcome. Economic advantage would require broad adoption, renewal and evidence that efficiency benefits translate into retention, price or contribution.

Hospitality remains less proven. Management describes Amadeus as the leading CRS provider by rooms of contracted customers [S6]. Contracted rooms are not the same as live properties, recurring revenue or profit. Oracle’s deployment across property management, distribution, guest, loyalty and sales functions at Loews demonstrates credible full-suite competition [S27]. The economic test is whether Hospitality contribution rises after major deployments rather than whether the contracted-room count grows.

Verdict: Amadeus has a real but non-uniform moat. Air IT deserves the highest confidence; Distribution retains scale and servicing value but faces stronger channel pressure; Hospitality must prove implementation economics. Falling bookings, weaker customer satisfaction and Sabre’s contrasting booking direction are the strongest current disconfirming evidence [S1][S5][S23].

Growth History and Forward Opportunities

Revenue increased from €2.67 billion in 2021 to €6.52 billion in 2025, while attributable profit moved from a €142 million loss to €1.34 billion [S13]. The early growth reflected post-pandemic normalization, so it is not an appropriate ordinary growth rate. Future growth must come from underlying travel volumes, implementations, new products, pricing, customer wins and adjacent markets.

Product-outlook answer: The product outlook is strongest in airline offer-and-order migration and contracted hospitality implementations, while Distribution depends on retaining an economically relevant role across direct, NDC and AI-assisted channels. Finnair and Saudia processed unified Orders during 2025, Air France–KLM selected Nevio, and management says roughly 25% of Altéa passengers are associated with a Nevio program [S5][S6]. Associated passengers are not the same as fully migrated passengers or realized revenue.

Management has cited potential revenue uplift from modern offer-and-order systems, and Finnair has projected a 4% pricing uplift and 10%–15% ancillary-revenue uplift from selected Nevio capabilities [S4]. These are customer and management expectations, not reported Amadeus cohort economics. The relevant investment return is incremental contribution after development, migration, support and incentives.

Hospitality’s largest programs include Marriott and Accor. On the H1 call, management said approximately 1,700 Marriott properties had been implemented, with Marriott work extending through 2027 and Accor through 2028 [S11]. These programs provide multi-year implementation visibility, but professional-service effort can dilute near-term margins. The decisive milestone is progression from installation expense to accepted recurring platform revenue and cash payback.

Cloud migration can accelerate deployment and product development, but the completion language requires precision. The 2025 report says 100% of applications were migrated from the Erding and Frankfurt data centers while remaining links, middleware and disaster-recovery systems were still in the last mile [S5]. H1 2026 excluded €17.3 million of cloud completion costs, and management expected decommissioning and vendor contracts to run off through year-end [S1][S11]. Savings must therefore be measured after residual costs and ongoing public-cloud expense.

AI offers at least four economic paths: improved developer productivity; lower unproductive search traffic; new agent-originated transactions; or loss of discovery power to large platforms. Management expects early productivity to appear as faster delivery and margin benefit rather than immediate headcount reduction [S11][S12]. The H1 workforce reductions were described as tactical cost containment and contractor flexibility, not an AI-driven structural redesign. Public evidence does not yet quantify output per developer, product-cycle time or AI-attributable revenue.

Long-run air traffic remains a baseline tailwind, not a sufficient thesis. IATA’s 3.1% long-term RPK forecast supports transaction growth, especially in Asia-Pacific, but 2026 shows how conflict, fuel supply and capacity can interrupt it [S8][S9]. Amadeus can outperform traffic through implementations and product adoption or underperform through customer losses and channel migration.

IDEMIA Public Security adds a separate growth vector. The target reported FY2025 revenue of €711 million, adjusted EBITDA of €112 million and adjusted EBIT of €70 million. Management estimates €50 million of annual cost synergies in the medium term and a 9.8× multiple on estimated 2026 EBITDA [S20]. The opportunity is a larger identity and border platform; risks include sovereign procurement, sensitive data, fixed-price implementation, regulatory approvals and paying in advance for synergies.

Other catalysts include Distribution stabilization, cloud run-off, more live Nevio migrations, recurring revenue from Marriott and Accor and evidence that AI-originated transactions still execute through Amadeus. Each needs a measurable KPI: matched booking growth, segment contribution, accepted live properties, lower cash infrastructure cost or paid agent-originated transactions.

Verdict: The opportunity set is credible but uneven. Contracted implementations and core-system migrations are more tangible than addressable-market estimates or unpriced AI potential. Persistent implementation spending without rising Hospitality contribution, or Nevio engagement without incremental cash returns, would disconfirm the growth thesis [S1][S6].

Financial Quality

The financial recovery is substantial. Revenue rose from €2.670 billion in 2021 to €6.517 billion in 2025. IFRS operating income progressed from a loss of €83 million to €1.758 billion, and attributable profit moved from a loss of €142 million to €1.336 billion [S13].

Primary accounts govern where standardized definitions differ. In 2025, standardized EBITDA was €2.493 billion while company EBITDA was €2.507 billion. Standardized CFO less capex was €1.386 billion, compared with company-defined FCF of €1.302 billion [S4][S13][S15]. The company’s FCF definition starts from EBITDA, deducts working capital, capex, taxes and net interest and includes specified derivatives; it is a levered cash measure after interest, not unlevered enterprise cash flow [S1].

Earnings-cycle answer: 2025 earnings were a post-pandemic recovery high in absolute euros, but H1 2026 represents a cyclical slowdown with an unresolved structural Distribution question rather than a proven peak in the entire franchise. Bookings weakened while Air IT and Hospitality continued to grow [S1]. A structural peak would require evidence that direct distribution, new interfaces or competing systems permanently reduce Amadeus’s contribution economics.

H1 2026 separates reported and adjusted performance. Revenue was €3.335 billion, up 2.3% reported and 5.1% at constant currency. IFRS operating income rose 0.6% to €943 million; IFRS profit fell 3.7% to €700 million. Adjusted EBIT was €1.012 billion after excluding purchase-price amortization, impairment, M&A costs and €17.3 million of cloud completion costs. Adjusted profit was €749 million, up 1.4% [S1][S2]. These adjustments reconcile arithmetically, but recurring acquisition amortization and repeated cloud exclusions remain economic costs.

Adjusted diluted EPS rose 6.3% while IFRS profit declined because diluted weighted shares fell from 449.5 million to 428.0 million [S1]. The shares were actually repurchased and cancelled, so this is not fictitious accretion. It nevertheless means per-share growth was materially stronger than organic profit growth.

Business-profitability answer: Reported profitability is high—2025 IFRS operating margin was 27.0% and standardized ROIC was 16.5%—but returns are method-sensitive because of capitalized development and goodwill. A transparent analyst calculation produces approximately 19% goodwill-inclusive 2025 ROIC: after-tax IFRS operating income divided by average equity plus interest-bearing debt less cash. It uses reported tax and debt conventions and should be read beside, not substituted for, the standardized measure [S3][S14][S16].

At year-end 2025, goodwill plus other intangible assets totaled €8.26 billion, versus equity of €4.85 billion [S3][S14]. Tangible equity was therefore negative. Price-to-tangible-book and tangible ROE are not useful valuation anchors. Goodwill-inclusive ROIC and cash flow better reflect the capital committed to acquisitions and software.

Accounting-conservatism answer: Accounting is not unusually conservative because qualifying development is capitalized and amortized over future periods, although the policy is disclosed and audited. Gross 2025 R&D was €1.453 billion: €697 million was expensed and €756 million capitalized, so approximately 52% entered the balance sheet [S3]. Technology capitalization and revenue-system accuracy were key audit matters.

Capitalized development affects all headline measures differently. EBITDA retains expensed R&D but excludes amortization of capitalized software. EBIT recognizes amortization later rather than current cash investment. Cash flow deducts capitalized development when paid. Investors should therefore monitor total R&D, the capitalized share, useful lives, impairments and cash returns together.

Capital-intensity answer: Physical capital intensity is low, but economic reinvestment is high: gross 2025 R&D equaled approximately 22% of revenue and cash capex was about €814 million, dominated by internally developed software. H1 2026 R&D was €682 million and capex €328 million [S1][S3][S4]. Approximately half of H1 R&D supported portfolio and product evolution, 30% customer implementations and 20% infrastructure and cloud partnerships.

Cash conversion has been strong over full years but contains timing items. Company FCF increased from €98 million in 2021 to €805 million in 2022, €1.152 billion in 2023 and €1.335 billion in 2024, before declining 2.4% to €1.302 billion in 2025. The 2024 number included €116.2 million of non-recurring tax collections; excluding them, 2025 FCF grew 6.9% [S4].

Income-cash-divergence answer: Full-year operating cash flow exceeded net income in 2025, but H1 2026 conversion weakened because company-defined working capital absorbed €312.2 million while cash taxes and net interest increased. H1 operating cash flow was €880 million and company FCF €472 million against IFRS profit of €700 million [S1][S2]. Personnel payments accrued in 2025 and ordinary billing seasonality contributed to the working-capital outflow.

The maintained full-year FCF range requires a large second-half conversion. This is plausible because H1 is seasonally weaker, capex was below the prior year and some working capital should normalize. It is not assured. A second-half release produced primarily by stretching payables or temporary collections would be lower quality than EBITDA-driven conversion.

Off-balance-liabilities answer: Recognized lease liabilities are not the central hidden obligation; the more important economic commitments are the signed IDEMIA purchase and earnout, undisclosed break-up fee, implementation obligations and continuing development needed to support capitalized software. These are not all IFRS liabilities at June 2026, but they are claims on future cash and management capacity [S2][S19][S20].

Liquidity is currently sound. June accounting debt was €3.607 billion. Cash was €971 million, including €83 million restricted cash, producing accounting net debt of approximately €2.72 billion. The covenant/APM net-debt measure was €2.578 billion, or 1.0× trailing EBITDA, because it excludes operating leases, accrued interest and certain financing adjustments. Ninety-three percent of interest-bearing debt was fixed. Amadeus issued a €500 million 3.75% bond due May 2031 and had €1.4 billion of undrawn revolving facilities [S1][S2].

Verdict: Financial quality is good but less conservative than the adjusted-EPS trajectory suggests. Strong margins, cash generation and ROIC support the franchise; capitalized development, negative tangible equity, recurring exclusions and the second-half cash-conversion requirement are the principal offsets [S1][S3][S16].

Capital Allocation

Management’s stated order is organic investment, preservation of the dividend, M&A and additional repurchases while retaining balance-sheet flexibility [S12]. Actual behavior became more aggressive in 2025. Cash to shareholders totaled €1.951 billion, compared with company FCF of €1.302 billion, before the subsequent IDEMIA commitment [S4].

FCF-generation-and-use answer: Amadeus generated €1.302 billion of company-defined free cash flow in 2025, while dividends, capital-reduction repurchases and other shareholder cash outflows totaled approximately €1.951 billion; distributions therefore exceeded FCF by about €648 million. The balance sheet could absorb that difference, but 2025 should not be treated as a sustainable payout template [S4].

The principal 2025 capital-reduction program acquired 18.928 million shares for €1.3 billion, an estimated average of €68.68. The shares were cancelled in June 2026 [S2][S3]. The average price was about 27% above the current market price. Cancellation permanently reduced future claims, but the mark-to-market result is evidence of poor timing and opportunity cost.

H1 2026 treasury activity must be separated by purpose. The company acquired 9.816 million shares for €500 million under the capital-reduction program, approximately €50.94 each. It acquired 1.641 million shares for €95.7 million for employee remuneration and another 205,000 shares for €10.5 million for a French employee plan [S2]. Employee-plan purchases primarily offset compensation issuance and should not be credited as discretionary capital return.

Share-repurchase answer: Repurchases materially reduced the share base, but their economics were mixed: the 2025 €1.3 billion capital-reduction tranche averaged about €68.68 and is underwater at the current price, while the 2026 €500 million tranche averaged approximately €50.94. H1 diluted weighted shares fell 4.8%, supporting EPS despite lower IFRS profit [S1][S2][S3].

Insider-share-issuance answer: Employee equity is not large enough to offset recent capital-reduction cancellations, but employee-plan purchases and transfers must not be mislabeled as discretionary buybacks or conviction insider purchases. The reviewed filings did not provide a complete basis to claim either substantial open-market executive buying or no buying across every possible venue [S2][S3][S38].

Dividend-policy answer: Policy targets 40%–50% of adjusted consolidated profit; the FY2025 dividend was €1.54 per share in total, comprising €0.53 interim and €1.01 complementary payment, and was covered approximately two times by 2025 company FCF. At the current price it represents a trailing cash yield near 2.8% [S2][S3]. The annual management review’s wording could be misread as a €1.54 final payment, but the interim accounts establish that €1.54 was the total.

Acquisition-record answer: Recent acquisitions broadened payment, airport and hospitality capabilities, but disclosed early financial contributions do not yet demonstrate attractive returns. Voxel cost approximately €126 million and VisionBox €281 million. In their 2024 partial post-acquisition periods—not cumulatively since acquisition—Voxel contributed €16.7 million revenue and a €2.4 million loss, while VisionBox contributed €51.8 million revenue and a €3.7 million loss [S3]. Those periods are too short to determine lifetime returns.

IDEMIA Public Security is the larger allocation test. The signed agreement requires €1.2 billion cash plus up to €150 million contingent consideration, with closing expected around mid-2027 subject to approvals. A €1.2 billion bridge has an initial one-year term plus two six-month extensions [S19]. Based on reported FY2025 adjusted EBITDA of €112 million, headline consideration is 10.7× before earnout; management’s 9.8× multiple uses estimated 2026 EBITDA and a stated currency assumption [S20].

Management’s official December 2025 pro-forma calculation added €1.2 billion to €2.141 billion net debt and €112 million to €2.507 billion EBITDA, producing 1.3× leverage [S20]. Updating mechanically to June figures gives about €3.78 billion net debt and €2.64 billion EBITDA, or 1.43×, before earnout, fees, acquired cash, closing-date cash generation and definition differences. This is an analyst estimate, not reported leverage.

Compensation-policy answer: Executive incentives emphasize revenue, EBITDA, adjusted EPS, pre-tax operating cash flow and relative total shareholder return, but contain no direct ROIC or acquisition-return gate. The 2026 annual bonus weights revenue at 28%, EBITDA 30%, adjusted EPS 30% and other measures 12%; long-term awards weight adjusted EPS 50%, pre-tax operating cash flow 30% and relative TSR 20% [S21]. The CEO has a two-times-salary ownership requirement, plus malus and clawback provisions.

Management-motivations answer: The incentive design encourages growth and cash generation but can reward debt-funded acquisitions or buyback-supported adjusted EPS unless cash flow and relative TSR provide sufficient discipline. The CEO’s 2025 bonus paid at 124.5% of target and the 2022–2025 performance-share plan vested at 125.1%, while relative TSR was the weakest long-term component [S21].

Management behavior is mixed. Amadeus protected liquidity during the pandemic, restored profitability, reduced shares and retained modest leverage. It also repurchased heavily near the subsequent five-year high and soon committed to another sizeable acquisition. Strategic logic and realized per-share return should be assessed separately.

Verdict: The balance sheet can finance the announced plan, but capital allocation has entered a less forgiving phase. The 2025 repurchase price and the absence of demonstrated returns from recent acquisitions are material disconfirming evidence against describing allocation as consistently disciplined [S2][S3][S20].

Changes and Headwinds — Last Two Years

Environment-change answer: The operating environment changed materially through weaker 2026 traffic, Middle East airspace and fuel disruption, expanding NDC adoption, rising AI-disintermediation concern and the final stages of cloud migration. These forces affect transaction volume, product architecture, infrastructure cost and bargaining power simultaneously [S1][S5][S9][S28].

FY2025 remained strong: revenue grew 6.1%, adjusted EBIT reached €1.894 billion and normalized FCF increased [S4]. In 2026, traffic and bookings weakened. Management reduced constant-currency group revenue guidance from high-single-digit to mid-to-high-single-digit growth and adjusted diluted-EPS guidance from low-double-digit to high-single through low-double-digit growth. The passenger-demand assumption fell from 4.4% to 1.9%; stable adjusted EBIT margin, capex at 10%–12% of revenue and €1.35–1.45 billion FCF were retained [S7].

External-internal-drivers answer: H1 results reflected both external traffic weakness and internal pricing, product and implementation actions: bookings fell 3.7%, yet constant-currency revenue rose 5.1% because Air IT, Hospitality and revenue per unit outgrew volume. ANA implementation helped passengers, while Hawaiian’s de-migration and Spirit’s cessation reduced North American volume [S1]. Neither a purely macro nor purely execution explanation fits the evidence.

Management added important qualifications on the H1 call. Disruption-related activity represented around one-third of Q1 Air IT unit-revenue growth but much less in Q2. Second-half Air IT contribution would face mix pressure from airports and professional services. Cloud decommissioning and vendor-contract run-off were expected through year-end. Capex was expected near the low end of the 10%–12% revenue range [S11]. All are management expectations requiring confirmation.

Cloud language is internally reconcilable only when the milestone is defined. Applications were fully migrated from Erding and Frankfurt, while links, middleware, disaster recovery, physical decommissioning and contract run-off remained [S5][S11]. The economically relevant finish line is lower total cash cost without loss of reliability, not the application-migration date.

AI moved from a productivity opportunity to a valuation risk. Management argues that query caps protect economics and that AI can improve delivery speed [S11][S12]. The Google collaboration provides evidence of cloud deployment and experimentation [S28]. It does not establish that Amadeus will retain customer ownership or capture incremental agent revenue.

The finance leadership changed during the period. Caroline Borg became CFO on May 5, 2025, replacing Till Streichert, who left for another executive role [S29]. The CEO and broad strategy remained stable. The transition matters because acquisition financing and cash-conversion communication are now overseen by a relatively new CFO.

Markets-facilities-management answer: Important changes included weak air markets, the Erding and Frankfurt data-center wind-down, major Marriott and Accor implementations, a CFO transition and the signed IDEMIA Public Security acquisition. Each alters operating leverage, execution demands or balance-sheet risk [S5][S11][S19][S29].

Accounting-policy-change answer: No material recognition or measurement policy change was reported for H1 2026; IFRS 9 and IFRS 7 amendments and annual improvements had no material effect, while IFRS 18 is expected primarily to change presentation from 2027. Development capitalization remains the principal analytical issue rather than a newly introduced accounting policy [S2].

The evidence invalidates several stale shortcuts. February guidance has been superseded. Application migration does not mean all legacy cost has ceased. Highly recurring revenue does not mean fixed revenue. A Nevio-engaged passenger is not necessarily a live migrated passenger.

Verdict: Technology architecture and the product portfolio improved, but 2026 introduced a worse combination of traffic, implementation and allocation risk. Unchanged FCF guidance is supportive; failure to deliver it would damage both earnings quality and management credibility [S1][S7].

Risk Analysis

The principal risks are correlated. Weak travel can reduce transactions while development, implementation and acquisition spending remain fixed. Lower cash generation can reduce flexibility as debt rises. AI and NDC can require higher investment before monetization. A cyber incident can damage both revenue and the trust underlying switching costs.

Stock-decline-factors answer: Another guidance cut, persistent Distribution underperformance, weaker cash conversion, an expensive or delayed IDEMIA closing, implementation setbacks, a material outage or evidence of AI bypass could cause a substantial decline. Multiple contraction could amplify an earnings miss because the stock still trades near 10× trailing EBITDA [S1][S7][S18][S19].

Catastrophic-loss answer: A catastrophic investment loss would most plausibly require a compound event—systemic travel contraction, prolonged platform outage or cyber breach, major customer departures and acquisition-funded leverage—rather than an ordinary recession alone. Pandemic history shows earnings can collapse, while current liquidity and embedded systems reduce but do not eliminate permanent-impairment risk [S2][S5].

Total-loss answer: A literal equity wipeout is remote but conceivable if a severe operating failure coincides with structural distribution bypass, large legal or data liabilities and loss of refinancing access after acquisition leverage rises. Current positive free cash flow, mostly fixed-rate debt and €1.4 billion of undrawn revolvers make that path unlikely; no defensible numerical probability is available [S2][S19].

Gradual value erosion is more plausible than total loss. Distribution revenue could remain stable while incentives, cloud traffic and development cost rise. Hospitality could grow without attaining mature margins. AI could increase search load while consumer platforms capture the surplus. These outcomes would appear first in contribution, capex and cash flow rather than reported revenue.

Measurement complacency is another risk. Revenue per booking can rise through services and disruption. Nevio engagement precedes migration. Contracted hotel rooms precede installation and recurring revenue. The IDEMIA multiple uses estimated 2026 rather than reported 2025 EBITDA [S1][S6][S20]. None of those measures is invalid, but each requires a cash-return cross-check.

Verdict: Permanent-loss probability is low but cyclical and valuation drawdown risk is material. The bear case becomes considerably stronger if booking underperformance, capitalized development and acquisition leverage rise together [S1][S3][S19].

Valuation Discussion

At €54.18 and approximately 419.987 million economic shares, equity value is €22.755 billion. Adding June company-defined net debt of €2.578 billion produces simplified enterprise value of €25.333 billion [S1][S2][S17].

Company Financials reports trailing revenue of €6.592 billion and EBITDA of €2.526 billion at June 2026. Repriced to the September close, Amadeus trades at approximately 10.0× EV/EBITDA, 3.84× EV/revenue and 17.8× trailing earnings [S18]. The €1.35–1.45 billion FCF objective implies a 5.93%–6.37% equity yield. These are analyst calculations using current price and reported denominators, not company valuation claims.

The peer screen does not establish an obvious statistical bargain. Amadeus trades near Expedia and Sabre on EBITDA and below Booking, while possessing a stronger airline operating-system position than the OTAs. Sabre’s apparent P/E is unusable because its 2025 SynXis sale generated a large gain [S34]. TravelSky is operationally relevant but too China-specific for a simple multiple comparison [S24].

Own-history context also requires caution. The stock is 28% below its March 2025 high, but guidance, traffic and capital allocation changed. A consistently reconciled historical series combining IFRS and company APM denominators was unavailable, so no valuation percentile is invented. Price decline alone is not evidence of cheapness.

A reverse cash-flow calculation is more informative. Starting from €1.40 billion FCF, an 8.5% cost of equity and 2.5% terminal growth, the current equity value requires roughly low-single-digit FCF growth over the next five years. The exact result depends on acquisition treatment and working-capital normalization. The price therefore does not require rapid compounding, but it does require that current FCF be durable.

The pending acquisition must be included symmetrically. Adding €1.2 billion of consideration and €112 million of target FY2025 EBITDA produces provisional pro-forma enterprise value near €26.5 billion and EBITDA around €2.64 billion, still close to 10×. Earnout, fees, acquired cash, financing mix, closing-date cash generation and synergies are excluded [S19][S20]. Adding EBITDA without acquisition capital would materially overstate value.

Illustrative 2027 scenarios are assumptions, not forecasts:

The bear case assumes weak revenue quality, high reinvestment and loss of franchise premium. The base case assumes moderate growth rather than a full traffic snapback. The bull case requires successful implementation, stable Distribution economics and operating leverage. All values exclude dividends. A renewed large repurchase, acquisition issuance or different closing date would change the share and debt assumptions.

The market appears correct that 2026 traffic is weak, distribution interfaces are changing and IDEMIA consumes flexibility. It may be too pessimistic if those developments are being applied equally to Air IT’s operating-system switching costs. The fragile bull assumptions are stable matched booking share, implementation conversion and disciplined allocation. The fragile bear assumption is that changed discovery interfaces eliminate the continuing need for normalization, orders and servicing.

The factor model snapshot is unavailable. No quantitative sector beta, value, momentum, quality, rate, currency or residual-alpha estimate can therefore be reported. Qualitatively, the equity should be sensitive to global travel volumes, EUR translation, discount rates, acquisition leverage and the market’s preference for long-duration quality franchises. Those are economic sensitivities, not factor-model measurements.

Verdict: The current valuation is reasonable rather than distressed. It embeds modest cash-flow growth and rewards successful implementation, but peer multiples and the acquisition bridge contradict the idea that the stock is unambiguously cheap [S18][S20][S35][S36][S37].

Variant Perception

A reasonable consensus narrative is that Amadeus is a high-quality travel compounder temporarily slowed by geopolitics, with Nevio, Hospitality and cloud efficiency supporting medium-term growth. The skeptical narrative is that AI, NDC and direct distribution erode intermediary economics while capitalized development and M&A obscure returns. Both contain evidence.

Investor-question answer: The most decision-useful recent investor questions concerned whether AI increases unmonetized query load, whether cloud completion lowers total cash cost, how quickly Marriott and Accor convert and whether order management produces incremental revenue. Management said contracts cap query burdens, legacy-cloud costs should run off during 2026, Marriott and Accor implementations extend through 2027–2028 and offer-and-order can improve airline revenue [S11][S12]. These answers are hypotheses to verify, not independent proof.

The strongest bull case is that investors generalize Distribution disruption to the whole group. Air IT controls deeply integrated operational workflows; Hospitality is scaling from a lower margin base; and Amadeus may normalize and execute NDC and AI-originated transactions even when it does not own discovery. Low required FCF growth creates upside if traffic normalizes and contracted projects convert.

The strongest bear case is that current resilience is less recurring than it appears. Revenue per transaction includes services and disruption; adjusted EPS is supported by a smaller share count; EBITDA is flattered relative to current development cash by capitalization; and management committed acquisition capital after buying back stock at a much higher price. If bookings reflect channel or share loss rather than a cycle, higher R&D and incentives may be necessary merely to defend revenue.

Four load-bearing assumptions determine the outcome:

1. Distribution remains economically relevant. The bull thesis fails if Amadeus materially underperforms matched GDS measures for four quarters while contribution falls or incentives rise. The bear thesis weakens if share and contribution stabilize despite expanding NDC and AI [S1][S23].
2. Nevio and Hospitality generate incremental returns. The bull thesis fails through repeated implementation delay, customer incidents or contribution that does not rise after deployment. The bear thesis weakens if recurring revenue and cash contribution accelerate without disproportionate capex [S1][S11].
3. Current FCF is durable. The bull thesis fails below €1.30 billion in 2026 without a reversible timing item. The bear thesis weakens above €1.45 billion with normalized working capital and lower cloud run-off [S1][S7].
4. IDEMIA does not dilute returns. The bull thesis fails through leverage above roughly 1.75×, material approval remedies, synergy delay or post-deal returns below the cost of capital. The bear thesis weakens through on-time closing, realized synergies and deleveraging [S19][S20].

Positioning evidence is incomplete. The stock sits in the lower half of its 52-week range and the February sell-off shows institutional concern about AI [S17][S32]. Without a factor-model snapshot, complete short-interest history or matched ownership data, the stock cannot responsibly be described as crowded long or short.

Verdict: The differentiated view is not that the market ignores a monopoly. It is that Distribution’s interface risk may be applied too broadly to Air IT’s operational switching costs, while bulls may be underweighting adverse booking evidence and allocation mistakes. Matched multi-quarter booking loss plus lower contribution would be the strongest disconfirmation [S1][S23].

Fact vs. Interpretation

Several analytical rules were revalidated. A material debt-funded acquisition requires a complete pro-forma enterprise-value, leverage and interest bridge. Buybacks must be split between genuine capital reduction and employee-compensation cover. Revenue per booking during falling volume updates resilience before it updates market-share or pure-price confidence [S1][S2][S19].

Several retrieved assumptions were rejected as stale or inapplicable. Biotechnology-specific acquired-research rules do not transfer directly, although the broader rule that development capital remains in ROIC does. A historical recurring-revenue percentage cannot establish the 2026 mix. Fully migrated applications do not mean links, middleware, recovery systems and vendor obligations have ended. February guidance was superseded by July’s lower outlook [S2][S5][S7].

Verdict: Evidence supports a durable and profitable platform, but not every favorable shorthand used around it. Denominator discipline—GDS share versus total share, engaged passengers versus live migrations and estimated EBITDA versus reported EBITDA—is essential [S1][S6][S20].

Open Questions

1. What portion of Distribution’s booking decline came from global demand, geography, cancellations, customer migrations and direct-channel shift? The current disclosure cannot distinguish cycle from share loss [S1][S23].

2. How much of the 5.1% Distribution revenue-per-booking increase was contractual price rather than renewals, services, disruption or mix [S1][S11]?

3. What are Marriott, Accor and Nevio implementation costs, recurring revenue, contribution margins and cash payback periods [S1][S11]?

4. Will residual data-center and vendor costs disappear after 2026, and how will continuing public-cloud expense compare with prior infrastructure cost [S5][S11]?

5. What percentage of AI-agent searches converts into paid transactions, and do query caps preserve contribution as look-to-book ratios increase [S11][S28]?

6. What permanent debt, interest rate, fees and acquired cash will replace the IDEMIA bridge, and what regulatory remedies could change the economics [S19][S20]?

7. How much of the €50 million synergy target is headcount, procurement or infrastructure, and what one-time integration cost is required [S20]?

8. What valuation framework governed the 2025 buyback and the subsequent acquisition commitment [S2][S19]?

9. How should investors reconcile 27% operating margin with 22% gross R&D intensity and roughly half of R&D capitalized [S3][S4]?

10. Is weaker airline NPS connected to booking and renewal outcomes, or is it an independent service measure [S5]?

11. What is normalized working capital after the H1 €312.2 million outflow? A temporary release would be lower-quality guidance delivery than operating profit conversion [S1].

Verdict: The largest evidence gaps concern matched booking share, implementation-cohort returns, AI transaction economics and final acquisition financing. Each is capable of changing valuation materially [S1][S19].

What Must Be True

Bull tests

• Distribution bookings should recover to within approximately two percentage points of matched GDS-market growth by H1 2027. Monitor regional bookings, Sabre air bookings, cancellations and airline-direct mix. Four consecutive quarters of wider underperformance with falling contribution would falsify the test [S1][S23].

• Air IT and Hospitality revenue should continue to outgrow underlying passenger activity without material deterioration in contribution. Monitor revenue per passenger, airport and professional-service mix and segment contribution. Loss of revenue outgrowth after disruption and implementation revenue normalize would falsify it [S1][S11].

• Marriott and Nevio milestones must become recurring cash economics. Monitor accepted properties, live modules, recurring platform revenue, implementation cost and contribution. Repeated schedule slippage or persistently low Hospitality contribution after substantial deployment would falsify the test [S1][S11].

• 2026 company-defined FCF must land within €1.35–1.45 billion without exceptional tax receipts or an unsustainable working-capital release. Less than €1.30 billion absent a specifically reversible timing item would falsify the cash thesis [S1][S7].

• IDEMIA should close on acceptable terms, keep pro-forma leverage around or below 1.5× and deleverage thereafter. Approval remedies, leverage above 1.75× or post-integration returns below the cost of capital would falsify the acquisition thesis [S19][S20].

• AI-originated demand must continue using Amadeus for normalization, orders, payments or servicing. Monitor paid agent-originated transactions, conversion and query cost. Direct large-agent integrations accompanied by falling Amadeus transaction contribution would falsify the test [S11][S28].

Bear tests

• The bear case requires Distribution weakness to be structural. Stable matched share and contribution despite expanding NDC and AI would falsify that premise [S1][S23][S25].

• It requires development spending to rise merely to defend existing economics. Falling R&D intensity alongside continued innovation, reliability and growth would falsify that premise [S1][S3].

• It requires Hospitality implementations to remain service-heavy and low return. A sustained contribution-margin increase after Marriott and Accor deployment without disproportionate capitalized development would falsify it [S1][S11].

• It requires further capital destruction. A pause in expensive repurchases, realized IDEMIA synergies and rapid restoration of leverage would falsify that premise [S2][S19][S20].

• It requires the operational moat to weaken. Major renewals, stable passenger-system share, improving NPS and an absence of core-system losses would falsify that premise [S5][S6].

The decisive monitoring framework is matched booking share, segment contribution, normalized free cash flow, implementation-cohort returns and pro-forma leverage. Revenue growth alone cannot settle the debate because both a healthy mix shift and a costly defense of market position can initially produce rising revenue [S1][S3][S20].

Public source appendix

• S1: Amadeus H1 2026 Management Review — primary company filing; published 2026-07-31; pp. 3–4, 15–18 and 33–39; segment KPIs, contribution, guidance context, leverage and FCF reconciliation
• S2: Amadeus H1 2026 Interim Financial Statements — primary IFRS financial statements; published 2026-07-31; pp. 12–15 and 23–25; revenue categories, debt, liquidity, dividends, treasury shares and accounting policies
• S3: Amadeus FY2025 Audited Consolidated Annual Accounts — audited primary financial statements; published 2026-02-27; Financial statements and notes on intangible assets, R&D, business combinations, remuneration, dividends and treasury shares
• S4: Amadeus FY2025 Management Review — primary company results review; published 2026-02-27; pp. 3–4, 11–15, 23–24 and 31–39; segment results, FCF normalization, capital returns and APMs
• S5: Amadeus Global Report 2025 — primary annual operating report; published 2026-02-27; pp. 27–28, 35–43 and 57–59; customers, NDC, technology scale, NPS and cloud-migration status
• S6: Amadeus Q1 2026 Results Presentation — management presentation; published 2026-05-08; slide 5 and footnotes; passenger-system share, GDS channel share, Nevio engagement and addressable-market estimates
• S7: Amadeus H1 2026 Results Presentation — management guidance presentation; published 2026-07-31; slide 16; revised 2026 revenue, EPS, margin, capex, FCF and passenger assumptions
• S8: IATA Long-Term Passenger Demand Forecast — industry-association forecast; published 2026-03-17; Central forecast and regional 2024–2050 RPK growth scenarios
• S9: IATA Global Outlook for Air Transport — June 2026 — industry-association outlook; publication date unavailable; pp. 3 and 13 onward; energy shock, 2026 RPK forecast and airline profitability
• S10: IATA July 2026 Passenger Demand Release — industry-association operating data; published 2026-08-31; Global and regional July RPK changes
• S11: Company Financials — H1 2026 Earnings-Call Transcript — third-party transcript reconciled to company materials; published 2026-07-31; Prepared remarks and Q&A on bookings, AI queries, cloud run-off, Marriott, Accor, capex and segment mix
• S12: Company Financials — FY2025 Earnings-Call Transcript — third-party transcript reconciled to official transcript; published 2026-02-27; Prepared remarks and Q&A on allocation priorities, offer-and-order economics and AI productivity
• S13: Company Financials — Amadeus Annual Income Statements — third-party standardized financial data; publication date unavailable; FY2021–FY2025 revenue, operating income, profit, EBITDA and EPS; reconciled to filings
• S14: Company Financials — Amadeus Annual Balance Sheets — third-party standardized financial data; publication date unavailable; FY2021–FY2025 cash, debt, equity, goodwill, intangible assets and shares; reconciled to filings
• S15: Company Financials — Amadeus Annual Cash-Flow Statements — third-party standardized financial data; publication date unavailable; FY2021–FY2025 operating cash flow, capex and standardized free cash flow
• S16: Company Financials — Amadeus Profitability Ratios — third-party standardized financial data; publication date unavailable; FY2022–FY2025 standardized margins and ROIC
• S17: Company Financials — Amadeus Daily Price History — third-party market data; published 2026-09-11; 1,279 unadjusted daily observations through September 11, 2026
• S18: Company Financials — Amadeus Trailing Multiples — third-party valuation data; publication date unavailable; June 2026 trailing denominators repriced to the September 11 close
• S19: CNMV Filing — Signed IDEMIA Public Security Acquisition and Financing — regulator-filed transaction notice; published 2026-07-23; Entire filing; signed agreement, consideration, earnout, bridge and expected closing
• S20: Amadeus IDEMIA Public Security Acquisition Presentation — management transaction presentation; published 2026-04-29; slides 5–6 and 9–16; target financials, valuation basis, synergies, business model and pro-forma leverage
• S21: Amadeus Directors’ Remuneration Report 2025 — primary governance filing; published 2026-02-27; Annual bonus, performance-share metrics, ownership requirement, malus/clawback and 2025 outcomes
• S22: Sabre 2025 Form 10-K — competitor primary filing; published 2026-02-18; Item 1 and risk factors; revenue model, direct distribution, NDC, competitors and incentives
• S23: Sabre Q2 2026 Earnings Release — competitor primary results release; published 2026-08-06; Marketplace revenue, air direct billable bookings, fees, margin and cash flow
• S24: TravelSky Technology 2025 Annual Report — competitor primary filing; published 2026-04-22; Aviation, airport, distribution, settlement and related-party customer disclosures
• S25: Travelport NDC Airline Matrix — competitor product documentation; publication date unavailable; Current airline NDC search, sale and servicing capabilities
• S26: IBS Software iFly RES Product Sheet — competitor product documentation; publication date unavailable; Passenger-service, departure-control, offer/order and NDC capabilities
• S27: Oracle Hospitality Loews Implementation Release — competitor company release; published 2026-07-14; Implementation across property, distribution, loyalty, guest and sales systems
• S28: Amadeus and Google Cloud Partnership Expansion — company partnership release; published 2025-05-22; Cloud infrastructure, Vertex AI and exploratory travel-product integrations
• S29: Amadeus CFO Appointment Notice — primary governance notice; published 2025-02-06; Appointment of Caroline Borg effective May 5, 2025
• S30: CNMV Filing — Amadeus Denies Interest in Shift4 — regulator-filed company notice; published 2024-02-29; Entire filing; response to acquisition rumors
• S31: Cinco Días — Amadeus Falls 7.3% Amid Deal Rumors and Downgrades — secondary contemporaneous news; published 2024-02-29; Observed February 29, 2024 price decline and contemporaneous market narrative
• S32: Cinco Días — AI Concerns Push Amadeus to Multi-Year Lows — secondary contemporaneous news; published 2026-02-22; February 2026 market narrative, analyst concerns and peer travel-technology sell-off
• S33: Amadeus FY2024 Results Release — primary company release; published 2025-02-28; FY2024 results and announcement of €1.3 billion repurchase
• S34: Sabre Completion of SynXis Sale to TPG — competitor regulator-filed release; published 2025-07-01; Completion and $1.1 billion consideration for hospitality sale
• S35: Company Financials — Sabre Trailing Multiples — third-party valuation data; publication date unavailable; June 2026 trailing denominator repriced to September 11, 2026
• S36: Company Financials — Booking Holdings Trailing Multiples — third-party valuation data; publication date unavailable; June 2026 trailing denominator repriced to September 11, 2026
• S37: Company Financials — Expedia Trailing Multiples — third-party valuation data; publication date unavailable; June 2026 trailing denominator repriced to September 11, 2026
• S38: CNMV Amadeus Current Filings Register — regulatory filing register; publication date unavailable; 2024–2026 results, treasury-share, governance and transaction filings