For most of the last century, an airline seat had a price you could look up. A fare class, a booking code, a number on a chart. That model is disappearing, and not gradually. Offer and Order, IATA’s framework for airline retailing, is replacing static fare filing with real time offers built per shopper, per search, sometimes per second. The industry has spent a decade building the infrastructure to price a seat almost any way it wants. In 2026, regulators and antitrust enforcers started asking whether that is entirely legal.
This piece traces three threads that are usually covered separately and treats them as one story. How pricing mechanics are changing. Where the money actually moves once ancillaries are unbundled. And why the same technology enabling personalization is now generating the evidence regulators need to act on it.
NDC ends the 26-price-point ceiling
Under the old EDIFACT distribution standard, an airline could only publish 26 fixed price points per route. NDC removes that ceiling. An airline can generate a functionally unlimited number of price points and update them continuously, a shift the industry calls continuous pricing.
Adoption is real but uneven. Industry surveys put airline respondents with live NDC channels at roughly 81 percent, with some carriers already exceeding half of their indirect bookings through NDC. IATA’s own framing for 2026 treats order management, not the fare filing system, as the new center of gravity for retailing, with carriers targeting concrete milestones this year: full Order pilots in selected channels, retirement of legacy flows, and live continuous pricing in indirect markets.
The more consequential shift is who is doing the pricing. A cluster of carriers, including Delta, Virgin Atlantic, WestJet, Azul and Viva Aerobus, have deployed AI based pricing engines from vendors such as Fetcherr, which builds what it calls a Large Market Model to generate real time fare and inventory decisions and publish them directly. Delta has described its early results from the rollout as encouraging, while pointedly emphasizing a deliberate pace rather than a rushed one. Scale context matters here: roughly 258 carriers now run some form of dynamic pricing, up from 220 just a few years ago, according to distribution data group ATPCO.
What this is not, at least according to the airlines using it, is personalized pricing based on an individual traveler’s identity or data. Delta has said publicly that dynamic pricing has existed in some form for three decades, driven by demand, fuel costs and competition rather than any single customer’s information. Whether that distinction holds as these systems mature is, as the next section shows, no longer a purely internal question for the airlines.
Ancillaries, not fares, are where the margin moved
Offer and Order matters commercially because it is what makes granular ancillary pricing possible at scale. Instead of a bare fare, an airline can present a full merchandised product: economy with no bag, economy with one bag, a flexible fare with free changes and seat selection, business with lounge access, each bundled and priced independently. Global ancillary revenue already exceeds 100 billion dollars, and NDC is positioned by the industry as the infrastructure for the next stage of that growth.
For carriers in TDN’s core markets, where base fares are compressed by low cost competition, ancillaries are increasingly the actual margin, not a supplement to it. The pricing sophistication described above is, in large part, in service of this: the more granular the offer, the more precisely an airline can price bags, seats and flexibility separately from the base fare, and the harder it becomes for a buyer to compare one airline’s true all in cost against another’s.
Washington reversed course twice in five months
That difficulty is exactly what United States regulators spent 2026 fighting over, in a sequence worth laying out precisely because it reversed twice within a single year.
A 2024 rule from the Department of Transportation had required airlines to disclose baggage and change fees upfront, the first time a fare and schedule search returned results. On February 3, 2026, a federal appeals court vacated that rule, though on procedural grounds, faulting how the department had built its economic case rather than the underlying policy. The court noted DOT had signaled it intended to draft a new proposal. Instead, on July 2, 2026, DOT went the other direction: a final rule took effect rolling disclosure standards back to the pre-2024 framework set in 2011. Separately, DOT has been weighing whether to go further still and allow airlines to advertise a base fare that excludes taxes and fees entirely, a full repeal of the so called Full Fare Rule.
The money at stake is not abstract. US airlines collected roughly 7.3 billion dollars in checked bag fees in 2024 alone. DOT’s own estimate was that the 2024 disclosure rule, had it survived, would have saved travelers more than 500 million dollars a year.
Running alongside the fee disclosure fight is a separate and arguably higher stakes question: whether algorithmic pricing itself can constitute illegal coordination between competitors. This is no longer treated as theoretical. In a May 14, 2026 speech titled Old Crime, New Code, Daniel Glad, the Justice Department’s Acting Deputy Assistant Attorney General for Criminal Enforcement in the Antitrust Division, said companies using shared algorithmic pricing tools face criminal exposure, not just civil liability, though the department has yet to bring an actual criminal case built around an algorithmic pricing tool. Its closest precedent so far is a civil consent judgment against RealPage, a property management pricing vendor accused of enabling landlords to coordinate rents through a shared algorithm, which required the company to limit the granularity of its pricing outputs and submit to a court appointed monitor. Glad has called it the Division’s paradigm case for how these arrangements can raise antitrust concerns even without a traditional agreement, though the matter was resolved civilly under a rule of reason analysis, not as a criminal prosecution.
Airlines are already named in this conversation, not hypothetically. The reported use of Fetcherr’s AI based pricing software by US carriers has drawn direct attention from lawmakers. Democratic legislators Greg Casar and Rashida Tlaib introduced legislation aimed at barring companies from using AI to set prices based on Americans’ personal data, with airline pricing specifically cited as a target. Even within the industry the position is not unified. American Airlines’ chief executive has said publicly that AI driven ticket pricing risks damaging consumer trust, a notably different posture from carriers actively rolling the technology out.
The throughline across both the fee disclosure fight and the algorithmic pricing question is the same. The infrastructure that makes granular, personalized, continuously updated pricing commercially possible is the same infrastructure generating a detailed, contestable record of how prices were actually set, which is precisely what regulators need to build a case.
Buyers still can’t see the full market
None of the above matters to a travel buyer unless the offer actually reaches them intact, and right now it frequently does not.
The structural reason is that GDS platforms were built for a world of fixed fare classes and standardized display, not for dynamic, ancillary rich offers generated per shopper. Most carriers now run NDC and GDS content in parallel rather than choosing one, and GDS providers have built NDC layers of their own to keep pace, Amadeus’s Altea platform carrying Air Canada’s NDC content being one concrete example. But parallel tracks mean two versions of the same seat can exist at once, and there is no guarantee a given booking channel is showing the richer one.
The cost of that gap is measurable from the buyer side. A recent GBTA survey found 46 percent of travel managers cited limited access to NDC fares as an active pain point with their booking tool. A separate industry data point suggests the problem is not only about which price a buyer sees but whether they can even account for what they paid: according to ARC’s 2025 Corporate Distribution Report, 47 percent of US travel management companies had not fully reconciled NDC order data with their back office reporting platforms as of December 2025, creating gaps in expense reconciliation and savings tracking.
Travel management companies are positioning themselves as the fix, promising to aggregate and normalize NDC content across carriers that each implement the standard differently. Reed and Mackay’s Richard Lindsay, the company’s director of air partnerships, has described the value proposition directly: richer content, avoided surcharges, and access to fares that would otherwise be invisible outside NDC, all surfaced inside the TMC’s own booking platform rather than requiring the traveler to shop the airline directly. Reed and Mackay’s own client data shows roughly 27 percent of their British Airways bookings and 13 percent of their Lufthansa bookings went through NDC in the final quarter of 2025, though this figure is specific to that one TMC’s client base and should not be read as an industry wide adoption number.
Even where content parity exists, continuous and algorithmic pricing adds a second layer of instability. A fare shown to a GDS based corporate booking tool in the morning may no longer match what an AI pricing engine is publishing direct by the afternoon. What a buyer sees, then, can depend less on the airline’s actual price than on which channel they are looking through and at what moment they happen to look.
A pricing system moving faster than its guardrails
Airline pricing has become genuinely more sophisticated over the past two years, and that sophistication is not evenly distributed. Airlines have it. Distribution infrastructure is still catching up to it. And regulators, after a year of reversing themselves twice on fee disclosure alone, are still deciding how much of it they are willing to allow. The carriers betting hardest on algorithmic, continuous pricing are making a wager that the commercial upside outweighs a regulatory and antitrust environment that is visibly still being written in real time, one court decision and one DOJ speech at a time. For buyers, the practical consequence is not a single clear price to compare, but a moving target that depends on which channel they happen to be looking through, and increasingly, at what moment.
