An invitation worth reading twice
This week I received a private invitation to join the Aman Club. The terms, for anyone who has not seen them: a non-refundable initiation fee of US$200,000 and annual dues of US$15,000 per club, for access to private lounges, spas and dining rooms inside Aman’s city hotels in New York, Tokyo, Bangkok and now Amanyangyun outside Shanghai, with Miami Beach, Beverly Hills, Dubai, Seoul and Singapore to follow. Food, drink, treatments and rooms are billed on top (Aman Club founder terms, SCMP).
I am not writing to complain about the price, and I am not on anyone’s list of the ultra-rich. I got this letter because I stay at Aman hotels. That is the point. Aman is the one hotel brand that could plausibly live on the top fraction of one percent alone, and by writing to a frequent guest it has just told me that the top fraction of one percent is not enough to fill a $200,000 club. The money it is reaching for is salaried money: the same money that fills every other luxury hotel.
Now look at what the rest of hospitality industry is building, and who it is building for. The growth plan of luxury hospitality for the next decade rests on a single assumption: that the global middle class keeps growing. That assumption is about to break, and AI is what breaks it. Not AI in the back office. AI in your customers’ jobs. I will say plainly what most of the industry’s AI commentary avoids: AI is going to remove white-collar employment at scale, the customer base for luxury hotels will contract sharply at today’s prices, and you are expanding supply straight into it.
The story we have all been telling for twenty years
The thesis behind almost every luxury hotel signed this decade is the rising global middle class, and the thesis has been right. In September 2018, Homi Kharas and Kristofer Hamel calculated that for the first time in history more than half of humanity, about 3.8 billion people, lived in households with enough discretionary income to count as middle class or rich. Five people a second were joining the middle class. Their projection ran to 5.3 billion by 2030, and their one-line summary of why it matters to business is worth re-reading: the middle class supplies two-thirds of household consumption, while the rich spend more per head but are too few to move the global economy (WEF / Brookings).
The trend has held even through this year’s oil shock. World Data Lab’s baseline still has 113 million people joining the consumer class in 2026, in line with the roughly 110 million a year of recent years, with Asia supplying 79% of them (World Data Lab, May 2026).
Hospitality has built on that number with both hands. The global hotel construction pipeline hit an all-time high of 15,976 projects and 2.43 million rooms at the end of Q2 2026. The luxury chain scale set its own record at 1,385 projects and 257,947 rooms, up 8% year on year, and upper upscale set another at 1,923 projects and 396,483 rooms. Lodging Econometrics expects 2,742 new hotels to open this year, 2,590 next year and 2,499 in 2028 (Lodging Econometrics, Q2 2026). In the United States the total pipeline shrank 4.9% in Q3, yet luxury projects rose 14% to a record 108 (Hotel Dive, Oct 2026). The money is leaving the middle of the market and crowding into the top, on the belief that the top is where demand is safe.
That belief deserves a closer look at who actually pays for a luxury hotel room.
Luxury is a middle-class business
Luxury hotels are not funded by the rich. They are funded, mostly, by salaried people spending a large share of what they earn, and the industry’s own research says so.
McKinsey defines a luxury traveler as someone paying $500 or more a night, sized the market at $239 billion in 2023 and expects $391 billion by 2028, faster growth than any other travel segment. Then comes the detail everyone skips: 35% of that market is travelers with a net worth between $100,000 and $1 million. Spending peaks between ages 40 and 60, and 80% of luxury leisure spend comes from people under 60, which is to say people who are still working. McKinsey’s advice to brands was to build entry-level offers for this tier, and notes that many luxury hotels already sell affordable standard rooms beside their top suites (McKinsey, May 2024; Forbes via Hospitality Net). You did not build the luxury pipeline for billionaires. You built it for the dual-income couple with a bonus.
Luxury goods have just shown what happens when that tier steps back. Bain and Altagamma count the world’s luxury consumers at around 340 million in 2025, down from 400 million in 2022, with new-customer acquisition falling 5% in a year and active shoppers dropping from roughly 60% of the addressable base to 40–45%. The ultra-wealthy kept buying; aspirational consumers pulled back. And the top could not fill the gap: big spenders’ share of the market rose from 30% in 2019 to 45% in 2024 and has now flattened at 46-47% (Bain-Altagamma, Nov 2025). BCG’s numbers are blunter: aspirational buyers, those spending under €5,000 a year, were 74% of luxury value in 2013 and are 61% now, 35% of them cut or stopped buying in the past year, and the brands with more than half their clients in that tier posted the steepest declines (Altagamma / BCG, July 2025).
What broke luxury goods was price increases and a soft macro backdrop, not technology. Treat it as a rehearsal. The next shock is aimed at the same customer, and it is structural.
The guest list, and what AI does to it
I came up as an organizational-design consultant; deciding how many people a business needs was the job. So I will say this without a hedge: AI will reduce employment at scale, not by the few percent in the consensus forecasts.
Look at who is in a luxury hotel tonight. Four kinds of people: the very rich, who live on capital; the powerful; the famous; and a large crowd of knowledge-economy professionals, the bankers, lawyers, consultants, doctors, engineers and executives, who pay for most of the rooms. McKinsey’s numbers say the same thing in its own vocabulary: 35% of luxury travel spend comes from people worth under $1 million, and most of the rest comes from millionaires who made their money from salaries and bonuses, not from capital.
The first two groups are small and will stay small. The fourth is the one AI removes, and it is not an abstraction. Banks are cutting junior analyst classes by as much as two-thirds (Fortune via Resultsense). Goldman has tied more than 1,000 cuts to its AI program and JPMorgan wants fewer junior bankers per senior one (Prospect Rock Partners). Fewer bankers, fewer lawyers, fewer consultants, fewer developers. The professional class that fills five-star hotels at the weekend is being made smaller, on purpose, by the people who run it.
Even the third group is not safe, because AI already makes films and songs. Tilly Norwood, a fully AI-generated “actress”, has been cast in a feature film (Washington Times, July 2026). Billboard counts at least six AI or AI-assisted acts on its charts; one of them, Xania Monet, hit No. 1 on R&B Digital Song Sales, became the first AI artist on a radio airplay chart and signed a record deal reported at up to $3 million (Express Tribune). A star is a person whose face and voice are scarce. That scarcity is ending, and some of the celebrity base goes with it.
So the customer base is not being reshaped. It is being shrunk: hard in the middle, and a little at the top. The only guests AI does not touch are the ones who own the capital, and there are not enough of them. Aman has just demonstrated that: it cannot fill a members’ club with them alone, which is why its invitation came to me.
This is not about AI in your hotel
Every conference panel on AI in hospitality is about cost: revenue management, chatbots, housekeeping schedules, the back office. That is a margin conversation, and it is the wrong one. The strategic question is what AI does to the income of the people who pay for the rooms.
The first hard evidence is already in. Stanford economists tracking ADP payroll records for millions of workers found that employment for 22- to 25-year-olds in the occupations most exposed to AI fell 13% relative to other workers since late 2022. In software engineering and customer service, entry-level employment dropped roughly 20% while employment for older workers in the same jobs grew. Accounting, administration, programming and sales show the same pattern (CBS News on the Stanford study).
The layoff ledger says the same thing in bigger numbers. Challenger, Gray & Christmas reports that AI was the most-cited reason for US job cuts for five consecutive months through July 2026, named in 112,713 announced cuts this year, about 24% of the total. Technology companies alone announced 149,023 cuts through July, up 67% on a year earlier (Challenger, July 2026). On October 26 Microsoft offered voluntary buyouts to about 60,000 people, a quarter of its workforce, in the same year it expects to spend $175 billion on AI infrastructure (CBS News, Oct 26, 2026).
Call it AI-washing if you like; Challenger notes that naming AI in a layoff announcement plays well with investors. It changes nothing for a hotel developer. Money is moving from payroll to data centers, and the payroll it leaves behind is the luxury traveler’s: the 28-year-old developer, the 35-year-old analyst, the 42-year-old marketing director. These are the people McKinsey counts as the bulk of luxury travel spend. The luxury guest of 2035 is the graduate who is not being hired in 2026.
What follows is a distribution problem: whether the gains from AI are taxed and transferred, and how fast. That is a political question with a political timetable. A hotel signed this year opens in 2029 and must perform through the 2030s. It cannot underwrite on a transfer system that does not yet exist.
The digital nomad is the canary
If you want to watch this happen in real time, watch the customer your industry spent the last five years courting hardest. The digital nomad was the purest form of the work-from-anywhere middle class: a developer, designer or analyst billing from a laptop, moving between Lisbon, Bali and Mexico City, filling the long-stay products, coliving brands and “workcation” packages we built for them. Dozens of countries issued visas for them.
MBO Partners counts 18.5 million American digital nomads in 2025, up 153% since 2019. Look underneath the headline. Growth slowed to 2.2%, and the independent nomads, the freelancers who are the archetype, fell 7% in a year, from 7.9 million to 7.3 million. MBO lists a weaker market for remote workers among the causes. The same survey finds that 89% of nomads already use AI in their work (MBO Partners, 2025). They are the earliest adopters of the tool that is replacing them, and the young software developer is the single occupation where Stanford found the sharpest employment decline.
The nomad’s income is the most AI-exposed income there is. When that bed-night disappears, it will not be replaced by a billionaire.
What this means for your pipeline
At today’s prices, the customer base for luxury hotels is going to shrink sharply. That is the base case, not a tail risk. Every pro forma signed in the last three years carries the twenty-year trend line, 110 million new middle-class consumers a year, out into the 2030s, while the people who actually pay $500 a night are being taken off payroll. And a hotel already built cannot shrink its way out. Hermès can cut volume and hold price because its capacity is a choice; a 200-key hotel’s rooms are sunk. That leaves three choices, and all three are revenue declines against the pro forma:
Hold rate and lose occupancy, which a fixed-cost business cannot sustain for long.
Cut rate to hold occupancy, and watch the luxury positioning go. Luxury goods have just shown what raising prices on a shrinking base does; cutting them is no kinder.
· Move up to the capital-owning apex. Aman, with the most apex-heavy guest list in the business, is writing to its frequent guests to fill a club. The apex will not fill 200 rooms for anyone else.
Five things I would do before the next signing:
Re-underwrite with every salaried tier, aspirational and millionaire alike, shrinking rather than growing with the trend line. Test what ADR and occupancy look like if the sub-$1 million-net-worth guest halves. If the deal only works on the old curve, it does not work.
Know your own mix. How much of your revenue is paid from salary and bonus, and how much from capital? If you cannot answer, you are the brand BCG warned about: more than half aspirational, steepest decline.
Stop building “luxury” at upper-upscale economics. A 200-key branded luxury hotel is a volume business dressed as a scarcity business, and most of the record pipeline is exactly that.
Watch three numbers every quarter: entry-level hiring in AI-exposed occupations, Challenger’s AI-attributed cuts, and Bain’s aspirational share of luxury. When all three move the same way, stop signing.
Read Aman’s letter correctly. Turning the top 0.1% into contractual annual income, through clubs, residences and memberships, is the right instinct, and even Aman cannot do it with the top 0.1% alone. If the apex is not enough for the most exclusive brand in the industry, it is not a strategy for yours.
This is not the middle hollowing out while the top grows. It is the middle and the upper-middle disappearing, the famous thinning, and the rich staying exactly as few as they are. The global middle class was never a demographic fact. It was an employment fact. AI is an employment event, and your industry has 16,000 projects in the ground that assume otherwise.
Sources
Aman Club Founder Information 2023 · fees, benefits and billing terms
WEF / Brookings: Half the world is now middle class or wealthier · Sept 2018 tipping point, 5.3bn by 2030, two-thirds of consumption
World Data Lab, World Consumer Outlook, May 2026 · 113m new consumers in 2026
Lodging Econometrics, Q2 2026 global pipeline · 15,976 projects; luxury 1,385
Hotel Dive, Q3 2026 US pipeline · total down 4.9%, luxury record 108
McKinsey: Updating perceptions about today’s luxury traveler · 35% aspiring, 80% under 60
Forbes via Hospitality Net · $239bn 2023 to $391bn 2028
Bain-Altagamma Luxury Study, Nov 2025 · 400m to 340m consumers, big-spender share flattening
Altagamma / BCG Consumer Insight, July 2025 · aspirational share 74% to 61%
CBS News on the Stanford/ADP study · 13% relative decline, 20% in software and customer service
Challenger, Gray & Christmas, July 2026 report · AI top reason five months, 112,713 cuts YTD
MBO Partners, 2025 Digital Nomads · 18.5m nomads, independents down 7%, 89% use AI
Fortune via Resultsense: banks cut junior analyst classes · up to two-thirds; 62% of AI talent from those cohorts
Prospect Rock Partners: Goldman’s 1,000 AI-linked cuts and JPMorgan’s 4:1 proposal
Washington Times, July 2026: AI actress Tilly Norwood cast in feature film
Express Tribune: AI singer Xania Monet breaks Billboard barrier · six AI acts on Billboard charts, up to $3m record deal