Analysis
Booking Holdings spends somewhere between $8 billion and $9 billion a year just to get travelers onto its platforms -- search ads, social, metasearch -- and CFO Ewout Steenbergen told Fortune that AI's bigger near-term opportunity isn't replacing that spend, it's making the trip itself more automated once a traveler is already booked.
The 'Connected Trip' Bet
Steenbergen's example: AI that notices rain in the forecast and swaps an outdoor Paris itinerary for a Louvre visit, or automatically rebooks a dinner reservation after a flight delay. Booking calls this the 'connected trip' -- tying flights, hotels, restaurants, rental cars and activities into one system that can react to a traveler's circumstances in real time rather than just helping them search once and forget it. It's a retention play as much as a product one: more touchpoints inside Booking's own apps means fewer reasons to go back to a search engine for the next leg of the trip.
“The company logged $186 billion in total bookings volume in 2025 and trades around a $119 billion market cap as of this month.”
Booking Holdings -- the parent of Booking.com, Priceline, Agoda, KAYAK, OpenTable and Rentalcars.com -- posted $26.9 billion in revenue for full-year 2025, up 13.5%, with $5.4 billion in net income. Priceline.com was founded in 1997, and the parent company renamed itself Booking Holdings in 2018 as its Amsterdam-headquartered Booking.com brand became the larger revenue driver of the group.
The company logged $186 billion in total bookings volume in 2025 and trades around a $119 billion market cap as of this month. That scale is exactly why an $8-9 billion acquisition-cost line matters: even a few points of efficiency from AI-driven retention would move Booking's margins more than most standalone travel startups' entire revenue.
Expedia Group is running a related playbook from a different angle: after largely shelving its all-in-one AI concierge Romie, Expedia acquired Berlin-based conversational trip-planner Layla in July and is shifting toward a 'multi-agentic' architecture of specialized AI agents rather than one assistant trying to do everything. Google has entered the same fight more directly, building flight-price tracking and hotel booking straight into AI Mode -- a move that threatens to intercept travelers before they ever reach Booking or Expedia's own apps. Airbnb, meanwhile, has leaned on its own AI-enhanced search and an active services-and-experiences marketplace rather than copying Booking's bundling directly -- a reminder that 'AI travel' is being fought on at least three different product theories at once, not one obvious winner. Booking's own research found 89% of consumers say they want to use AI in future trip planning, which is as much a warning as an opportunity: that demand will go to whoever's AI travelers trust first, not necessarily whoever has the biggest acquisition budget.
What the headline misses: Steenbergen himself has said publicly that even the AI hyperscalers don't yet know the ROI on their own AI spend, and his main concrete uses so far are internal -- an 'AI coach,' a 'strategic thought partner' for board materials, and a 'critical equity research agent' that pressure-tests his own earnings-call prep. The traveler-facing 'connected trip' is still a vision Steenbergen is narrating, not a shipped product with disclosed usage or retention numbers -- the $8-9 billion acquisition-spend line hasn't moved yet.
The number to watch in Booking's next few earnings calls is that acquisition-spend line as a percentage of bookings volume -- if AI-driven retention is real, it should start showing up there before it shows up in any press release about a new feature.
