The investment paradox: Why travel tech faces decline and how collaboration can save it 

In this fourth installment of my WTM 2026 series, let’s confront the structural crisis underlying our entire ecosystem: travel technology is facing a capital drought at the precise moment it requires generational investment. And the industry has nobody to blame but itself. The numbers are stark. Travel search is now effectively 100% AI-driven. Phocuswright data shows high percentages of all travellers across the globe, for example 92% of Japanese travelers already make trip decisions based on AI. Adoption across Europe doubled in twelve months, and the US sits past 56%. That data doesn’t account for the fact that Google’s Travel Searches are now all AI in part. Consumers have crossed the Rubicon. Yet when travel executives are asked if they are ready for agentic AI, a mere 9% say yes. That is not a readiness gap; it is an existential failure masquerading as an operational backlog.

Venture capital is not gone; it is simply elsewhere. Global institutional funding has been vacuumed up by frontier model labs, - OpenAI, Anthropic, Google DeepMind. A travel tech startup that raised a $10M Series A in 2023 struggles to pull together $2M today. Founders face a brutal, three-front squeeze:

• The Frontier Siphon: Investors believe true alpha lies exclusively in owning general model infrastructure, dismissing domain-specific travel expertise.
• Entrenched Gatekeepers: Legacy distribution incumbents control over 70% of direct and indirect transactions. The commercial rules remain rigged against outside disruption.
• The Tech Debt Mountain: Core airline and hospitality infrastructure still languishes beneath decades of legacy such as EDIFACT, fragile middleware, and incompatible APIs. Innovating within that labyrinth carries prohibitive engineering costs. 

When travel startups do manage to raise capital, they rush to plug into a closed frontier API or an aggregator feed. It delivers immediate speed-to-market, requires minimal in-house engineering, and guarantees rapid commoditization. Relying entirely on rented frontier models carries two fatal vulnerabilities:

• Instant Replication: If your core intelligence is merely a wrapped prompt to an external API, any competitor with a balance sheet can replicate your workflow overnight. You cannot extract margin from an asset anyone can rent by the token.
• Value Transfer: You are paying retail token rates while your domain logic and edge cases train the frontier model’s next iteration.

Decades ago, travel surrendered its autonomy to the GDSs. Today, we are eagerly sleepwalking into the exact same dependency with Big Tech SaaS and model providers. The tragic irony? Open-weight models, DeepSeek, Qwen, Llama etc., have functionally closed the performance gap on core travel workflows. But open models require serious engineering, fine-tuning, and operational discipline. Because capital has dried up, the industry defaults to shortcut convenience over sovereign control. 

The root cause here is only partially technological; increasingly it is commercial and contractual issues that are the challenges. Airlines still operate on EDIFACT, a pre-internet messaging format designed to protect established commercial territory, not facilitate modern commerce. Programs like Offer and Order (modern airline retailing) crawl at an agonizing pace, not due to poor technical architecture, but because incumbent revenue engines are built around extracting transaction toll fees from legacy rails. The GDSs are not really the villains; they are custodians of massive installed bases and billion-dollar revenue obligations. They cannot pivot overnight, yet their commercial models actively restrict innovation outside their own toll booths. As long as distribution frameworks remain frozen, capital will bypass travel tech for greener pastures. The GDS players bear a degree of blame here.  

Unchecked, this trajectory leads directly into a margin trap:

• Skyrocketing Operating Costs: Infinite API token calls create runaway expenses as consumer agentic queries multiply. Many of these surplus transactions are actually being invoked by poorly designed cache engines.
• Structural Paralysis: Stitching frontier AI onto legacy EDIFACT backbones results in brittle operational chaos.
• Diminishing Returns: When everyone licenses the same models and utilizes the same aggregators, product differentiation hits zero. With almost no supplier verifying lowest prices - the searches just keep growing. Are they generating more revenue? Are they generating new customers?

Meanwhile, consumers do not care if back-end operations run on modern event streaming or 1960s mainframe protocols. They care about seamless execution. If travel operators cannot deliver autonomous, conversational execution, players from outside our industry will step in and do it for them. Travel is complex, but we are not immune to outside disruption.

To unfreeze the framework, we need genuine collaboration among operators, vendors, and startups rather than defensive posturing. Incumbents like the public GDSs face lagging equity performance because markets recognize the structural stagnation. Transformation cannot be circumvented; it must be executed together.

• Vendors & GDSs: Must invest directly in open-weight AI infrastructure. Build internal engineering teams, own proprietary vertical intelligence, and offer clients deployment flexibility beyond raw API pass-throughs.
• Operators: Must accept the short-term pain of modernizing off legacy messaging formats. Kicking the migration can down the road is far more expensive than ripping the band-aid off now.
• Innovators: Stop building superficial API wrappers. Focus on deep, vertical open-weight architectures that solve settlement, scheduling, and distribution bottlenecks, licensing back into the ecosystem.
• Investors: Recognize that general frontier models cannot parse the nuances of complex global travel operations. Enterprise value lives in domain execution, agentic workflows, and proprietary distribution.

The industry spent decades choosing the path of least resistance, turning can-kicking into an art form. We are now paying that price. Yet every component required to reverse course is already on the table: mature open-weight models, deep institutional domain expertise, and unprecedented consumer demand. The chasm between 100% consumer adoption and 9% enterprise readiness will close. It will either close through proactive, industry-led reinvention, or through wholesale irrelevance. That is the conversation we need to have across the floor, boardrooms, and corridors at World Travel Market in London. Move past the polished panels. If you want to challenge these dynamics, find me on the floor. Let’s stop waiting for permission and start building.     

Looking for something else?