China

By the end of this century, Chinese tourism will most likely have become something that would be barely recognisable to a traveller from the 2020s: a predominantly domestic, senior-led, AI-mediated, and climate-defensive industry serving a population that has shrunk by more than half from its 2021 peak. The single most plausible trajectory is one of contracted scale and deepened quality — fewer travellers but higher revenue per capita, the world's largest national-park system at full build-out, a heritage portfolio digitally twinned for resilience, and a coastline retreating behind hard and soft defences against half a metre to a metre or more of sea-level rise. The binding constraints will not be money. They will be water, heat, and demographic age.

The demographic baseline is the foundation on which everything else rests, and it is halving and greying. China's population peaked in 2021 at roughly 1.426 billion and has fallen every year since. The UN's central projection puts the country at about 1.26 billion in 2050 and 633 million by 2100 — a population last seen in the late 1950s. Total fertility now sits at about one birth per woman, the lowest among major economies, and the median age is already over forty. By mid-century roughly four in ten Chinese will be over sixty; by 2100, more than half will be, with fewer than one in thirteen under fifteen. Tourism will therefore be, structurally, a market for the old, the digitally fluent, and the labour-scarce. The youth-led mass beach travel that defined the 2010s will give way to wellness retreats, accessible heritage paths, slow rail, medical tourism, and multi-generational journeys. Labour scarcity will drive hospitality automation — robotic housekeeping, AI concierges, autonomous shuttles, predictive maintenance — at a pace no Western economy will match, because no Western economy will have to.

The economic scaffolding is already in place. China is currently the world's second-largest travel and tourism economy and will most likely overtake the United States during the 2030s. Sector revenues are projected to roughly double over the next decade, with employment exceeding one hundred million people. Crucially, the domestic market dominates and will continue to dominate. In 2024, more than five and a half billion domestic trips generated upward of eighty per cent of all tourism spending, while inbound and outbound flows together accounted for less than three hundred million journeys. By 2100, with a halved population and a maturing outbound market, the value mix will most likely settle at roughly three-quarters domestic, with inbound and outbound each holding modest shares — a quiet reversal of the late-2010s anxiety that China was exporting tourism dollars at scale.

The climate trajectory, meanwhile, is locked in and heavy on tail risk. Central and East Asia have warmed faster than the global mean since the 1980s, and heavy precipitation is projected to intensify and become more frequent across monsoonal East Asia, with large increases in flood frequency. The empirical anchors of the past five years are stark. The July 2021 Zhengzhou floods killed nearly four hundred people, dropped more than two hundred millimetres of rain in a single hour — the highest hourly figure since records began in 1951 — and caused around nineteen billion US dollars in direct losses. The 2022 Yangtze compound heatwave and drought was the most severe in six decades, affected over nine hundred million people, halved Sichuan's hydropower output, and forced shutdowns at major manufacturers from Toyota to Tesla. Peer-reviewed projections of China's river-flood losses for the 2016–2035 period exceed three hundred and eighty billion US dollars, more than ten times the equivalent United States figure.

Coastal exposure is severe but, by Chinese standards, manageable. Under high-emissions, no-adaptation modelling, much of Shanghai could be inundated by 2100 with mean depths above a metre, with Chongming Island and the Pearl River and Bohai megaregions facing comparable trajectories. Yet Shanghai has already built five hundred and twenty kilometres of seawalls and tidal gates and has, through aquifer reinjection, slowed a century-long pattern of severe subsidence. Hainan, China's premier warm-weather destination, will most likely retreat from low-lying east-coast resorts toward elevated or relocated facilities while consolidating its role as a senior-care, marine-tourism and duty-free hub.

The strategic crux of the entire system, however, is the Tibetan Plateau — the "Third Pole" that feeds the Yangtze, Yellow, Mekong, Brahmaputra, Salween, and Indus, and through them more than a billion downstream residents. Mid-century glacier modelling projects whole-plateau ice-area loss in the range of a quarter under a moderate emissions pathway and over half under a high one by 2100, with some sub-zones losing their debris-free glaciers entirely. This is the silent engine driving every other decision: where to grow food, where to site data centres, which parks to designate, and which water-tower regions to seed with silver iodide in the hope of squeezing more rain from a thinning atmosphere.

Here lies the central paradox of the coming century. The same western and plateau provinces that host the most ambitious protected-area architecture ever attempted are also where the country is concentrating its compute. China inaugurated its national park system in 2021 with five flagship parks and has identified forty-nine candidates covering more than a tenth of the country's land area, with target completion by 2035. The Qinghai-Tibet Plateau alone hosts thirteen candidates and roughly seventy per cent of the candidate land area. At the same time, the "East Data West Computing" project is pushing hyperscale data centres into Inner Mongolia, Ningxia, Gansu, and Guizhou — the very provinces hosting the largest park candidates and the headwaters of China's major rivers. Chinese data centres consumed billions of cubic metres of water in 2022, the great majority of compute capacity sits in severely water-scarce regions, and demand could grow many-fold by 2030. By mid-century, water and electricity for artificial intelligence will be in direct competition with ecological flows, irrigation agriculture, and small-town tourism economies in some of the most ecologically sensitive landscapes on the planet.

Heritage tourism, by contrast, offers the cleanest demonstration of what AI can do well. China now has sixty UNESCO World Heritage sites, near parity with Italy and likely to overtake it within the next decade. The Digital Dunhuang programme — a high-precision digital twin of the Mogao Caves, with AI-assisted mural reconstruction, augmented-reality overlays of spectrally reconstructed pigments, and remote access through everyday messaging apps — is the global template for heritage preservation under climate stress. It is being extended to the Forbidden City, the Mausoleum of the First Qin Emperor, and the Grand Canal corridor. By 2100, a majority of "heritage visits" will most likely be remote, immersive experiences delivered from within China and abroad, with physical visitation rationed as a premium scarcity through AI-managed permit lotteries. The Forbidden City has already moved from a pre-pandemic nominal cap of eighty thousand daily visitors to an operationally enforced forty thousand, made possible by dynamic ticketing and density analytics that the older system could not deliver.

The infrastructure connecting all of this is high-speed rail, now exceeding fifty thousand kilometres of track — more than the rest of the world combined — and targeted to reach seventy thousand by 2035. Short-haul flights have been visibly displaced over the past fifteen years, and a continental low-carbon corridor stitching together China, Laos, Thailand, and Central Asia is taking shape. By mid-century, trans-Asian rail tourism will plausibly compete structurally with long-haul aviation across much of the Eurasian landmass.

What this future will not resolve on its own are the equity fault lines. Coastal megaregions hold the wealth, the inbound visitors, the heritage density, and the sea-level exposure; the interior holds the ethnic-minority cultural product, the national parks, the rural-revitalisation projects, and the data centres. The fifty-five officially recognised ethnic minorities account for under a tenth of the population but a much larger share of the cultural-tourism product, and without enforceable consent and benefit-sharing arrangements, proceeds flow disproportionately to coastal capital. By 2100 the most contested resource will be water, and tourism — Hainan resorts, Yunnan rice-terrace landscapes, Tibetan monastic circuits — will be one of several claimants alongside food security, ecological flows, and digital sovereignty.

The unifying rhetorical frame is "ecological civilization," constitutionally embedded since 2018. Whether it remains a genuine constraint on infrastructure expansion or becomes a slogan attached to extractive projects on the plateau and across the arid northwest is the binding political question of the century. The answer will determine whether the China of 2100 is a country that has used artificial intelligence and high-speed rail to deliver a smaller, slower, deeper kind of tourism to its ageing citizens within ecological limits — or one that has simply paved its water towers with silicon and called it stewardship.