Booking Holdings Inc

Q2 results: beat guidance across the board; strong profitability and EPS - Room nights +5% YoY (slightly above the high end of guidance). - Gross bookings +9% YoY (≈+8% constant currency) and revenue +8% YoY (≈+7% constant currency), both above the high end of guidance. - Adjusted EBITDA ≈ $2.6B (+9% YoY) with margin expansion of ~40 bps. - Adjusted EPS $2.54 (+15% YoY), helped by ~6% lower average share count. - Surprising/positive: despite an adverse macro environment (Middle East), core travel demand held up well and financial execution remained strong.

Strategic priorities (Connected Trip, loyalty, U.S./Asia expansion, AI) are progressing - Connected Trip: transactions grew in the low double digits and represented a low double-digit share of Booking.com transactions; importantly, growth outpaced total transaction growth by >2x. - Genius loyalty: Level 2 & 3 members are >30% of active customer base and represent a high 50% share of room nights (both up vs. prior year). Management sees additional opportunity to further strengthen Genius. - U.S. growth: unit room nights high single digits, supported by domestic demand and continued direct-channel growth + multi-vertical engagement. - Asia growth: unit room nights mid-single digits, with domestic low double digits; progress driven by localized products/payments/distribution leveraging Booking.com + Agoda footprint.

AI: early customer-facing progress, but top-line impact is still small (and deliberately not “given away” yet) - AI initiatives are being deployed across discovery (Booking.com AI discovery testing), agentic assistance (Priceline’s “Penny”), and UI improvements (Agoda Gallery View). - Customer service impact is measurable: voice AI support scaled across most eligible inbound calls; AI reduces friction/contact rates and helps lower customer service cost per booking by double digits, while customer satisfaction remains high. - Top-line conversion impact: management says Penny users show higher satisfaction and hopes conversion improves, but they won’t quantify CSAT/conversion lifts yet because results are still early. - LLM traffic share is <1% of room nights (based on large-language-model driven traffic patterns); management noted no material change recently—so near-term revenue impact is expected to be limited.