AI in Travel: Threat or opportunity? By Investing.com

Bernstein says AI could materially disrupt online travel agencies by enabling suppliers like Accor, Hyatt, and Wyndham to acquire customers directly through platforms such as ChatGPT, potentially lowering commissions and disintermediating OTAs. The key risk hinges on the AI monetization model: ad-supported systems could preserve current take-rate economics, while subscription or user-funded models could push the industry toward a 0% take-rate environment. OTAs Booking and Expedia are leaning on a trust gap argument, but analysts warn AI adoption may gradually erode that defense.

Analysis

This is less about a near-term rerating of BKNG/EXPE and more about a potential structural margin reset in digital distribution. The key second-order effect is that AI intermediaries can compress the difference between brand demand and channel demand: if the consumer starts the session in an agent, the supplier’s bargaining power rises and the OTA becomes a payment rail rather than the owner of demand. That would pressure gross bookings growth first, then marketing efficiency, then multiple durability, with the sharpest risk showing up over 12-24 months rather than immediately.

The market is probably underestimating how quickly hotel direct-connect infrastructure can scale once a few large chains normalize agent-native booking. If even a modest share of inventory moves to lower-cost AI referrals, OTAs may have to defend share with higher paid acquisition, rebates, or exclusive inventory, which is structurally negative for EBIT margins. The more important competitive threat is not a full replacement of search, but a gradual shift in the mix of high-intent traffic away from OTA-led funnel economics.

The contrarian point is that the current bear case may be too linear: in an ad-funded AI regime, OTAs can become the highest-bidding monetizers and preserve take rates. That argues for a more selective rather than blanket short. The real asymmetry is that the market is treating trust as a gating issue, but trust often lags utility by only a few product cycles; once users successfully book a handful of trips through agents, conversion friction can collapse faster than consensus expects.

Key Decisions for Investors

• Short BKNG vs long GOOGL/GOOG over a 6-12 month horizon: if AI distribution becomes ad-funded, search incumbents capture the monetization layer while OTAs face margin pressure; target 1.5-2.0x downside on BKNG relative to market if take-rate compression becomes visible.
• Buy put spreads on EXPE 6-9 months out to express a slower-burn disintermediation view: limited premium outlay versus meaningful downside if AI referrals start to divert high-intent traffic and force higher CAC.
• Long supplier beneficiaries with direct distribution leverage, funded by short OTAs: prefer hotel/platform operators with strong loyalty ecosystems over intermediaries; this is a relative-value trade with better risk/reward if AI booking flows improve conversion for direct channels.
• Avoid chasing the panic in OTAs intraday; wait for evidence in app traffic, marketing expense, and channel mix over the next 1-2 earnings cycles before sizing the short, since the first-order trust barrier can support a tactical squeeze.
• Set a catalyst watch on AI platform monetization announcements over the next 3-6 months: a shift toward subscriptions would be the cleanest bearish catalyst for BKNG/EXPE and would justify adding to shorts on any relief rally.