RLJ Lodging Wants You to Buy AI Capex Through Hotel Rooms, Not Data Centers
A 9.7% RevPAR quarter and a 200bp guidance raise turn a deeply discounted hotel REIT into a second-derivative AI trade — just as the tape abandons the first derivative.
RLJ · Earnings Call · 2026-11-05
The quarter that broke the pattern
RLJ Lodging Trust spent most of the last three years apologizing for urban hotels. This quarter it stopped. Comparable RevPAR rose 9.7% to $153 on a 6.4% ADR gain and a 230bp occupancy gain; hotel EBITDA climbed 17.5% to $93.9 million — nearly twice the pace of revenue — and margins expanded 190bp to 26.5%. Adjusted FFO per share of $0.34 was up 25.9%. Full-year RevPAR guidance went from roughly 4% to 5.75–6.25%, “an increase of 200 basis points at the midpoint” — Nikhil Bhalla, Chief Financial Officer · 2026-11-05, and the corporate EBITDA range moved up $9 million. CEO Leslie Hale's framing is that this is a regime change, not a good print: “first time in many years that lodging fundamentals converge” — Leslie Hale, President and Chief Executive Officer · 2026-11-05 — group demand providing base occupancy, transient accelerating on top, and limited new supply. CFO Nikhil Bhalla supplied the historian's version, arguing that 2015–2019 masked the true GDP-to-RevPAR relationship with supply growth, and that “demand converge with GDP growth, but no supply” — Nikhil Bhalla, Chief Financial Officer · 2026-11-05 is the better setup. That is the GDP growth thesis in one sentence. So far, so much REIT boilerplate. The interesting part is what is actually driving it.The AI trade the tape refuses to price
RLJ's largest market is Northern California: RevPAR up more than 13% in the quarter, 16% year to date, with EBITDA up 63% over two years. Management attributes this explicitly to artificial intelligence. From the prepared remarks:That is a hotel REIT volunteering itself as a levered play on AI capital expenditure. And the timing is exquisitely awkward, because the public tape is doing the opposite: over the last 30 days AI data centers as a theme has fallen roughly 31% across 66 tickers, with data center adjacent themes — capacity, power, interconnects — all in the steepest decliner bucket. Analyst Chris Woronka pressed exactly this, asking whether the demand pool in AI-linked markets is deep enough to absorb a pullback in AI spending. Hale's reply is the most quotable line of the call: “You have to decouple what is happening in the stock market from the reality of using it on the ground.” — Leslie Hale, President and Chief Executive Officer · 2026-11-05 COO Tom Bardenett added the practical version — data centers next to a hotel do not generate room nights, but the payroll, contractor traffic, vendor meetings and software implementation work surrounding them do. It is a real distinction, and it is the kind of claim that will be tested over the next four quarters. Note also that the AI demand is showing up not just in Northern California but in Washington, D.C. (business transient revenue +53%), Austin (+24%) and New York (+19%).Performance in this market continues to be underpinned by the expansion of the AI industry, which is fueling corporate investment, citywide activity, and business travel... We continue to believe in the multi-year growth trajectory for Northern California as the epicenter of the broader AI boom that is driving the next phase of economic growth.