The Unstated Population: Building for a city nobody asked for

In 2017, Singapore's water prices rose 30 per cent in two phases — the first hike that July, the second in July 2018. It was the first increase in seventeen years. MPs asked whether it could be deferred. The answer was no.

The infrastructure, the Minister said, had to be built ahead of demand. 

Nobody asked whose demand.

The demand figure came from a population projection. The population projection came from an immigration pipeline. The pipeline was not up for debate.

Residents received the bill and were told it was an investment in their future.

A year later, the Government announced the Airport Development Levy — a charge on every departing passenger, starting immediately, to help fund Terminal 5 at Changi, a project not expected to open until around 2030. The Second Minister for Transport said collecting the levy early would avoid larger charges later. 

He confirmed the Government had already committed more than $9 billion to the project. Passengers in 2018 were asked to begin funding a terminal for 50 million additional annual travellers — capacity premised on a Singapore substantially larger than the one they were living in. The population assumption was not stated. It did not need to be.

That was the template. In March 2026, three ministers stood up at Committee of Supply and ran it again — at longer range and larger scale than anything that came before.

Taken individually, each announcement had a plausible rationale. Taken together, they raise a question that was not asked in the chamber: is Singapore building infrastructure for a future it has actually thought through — or is it building by institutional momentum, each ministry executing its mandate, nobody responsible for whether the mandates still add up?

The Population Assumption Nobody Stated

Every major infrastructure commitment announced at Committee of Supply 2026 carries an embedded assumption about population size.

Ms Grace Fu described a continuous coastal defence line being built around the entire island, new desalination studies, drainage improvement works across multiple districts, and Singapore's first National Adaptation Plan. Coastal infrastructure of this scale carries a planning horizon measured in decades and a capital expenditure that will be repaid — through taxes, through utility tariffs, through the slow arithmetic of national debt service — by Singaporeans who have not yet been born.

Minister Zaqy Mohamad outlined the Singapore Food Story 2: a $70 million Agri-Food Cluster Transformation Fund, expanded national broodstock centres, new food stockpile categories. It was presented as ambition. The more honest framing is that it was a retreat dressed as a refresh.

In November 2025, Minister Grace Fu quietly announced that Singapore was abandoning its original "30 by 30" goal — the target, set in 2019, to produce 30 per cent of the country's nutritional needs locally by 2030. The new targets push the horizon to 2035, narrow the categories that count as locally produced, and lower the effective bar.

In 2024, 8 per cent of fibre consumed in Singapore was produced domestically. The protein figure, reached by combining eggs and a selective subset of seafood, stood at 26 per cent. Singapore is not approaching food sufficiency. It is recalibrating the definition of the destination to make the distance look shorter.

The reason is not difficult to identify. Food production targets are set for a population. Singapore's population keeps growing. The target was never going to be caught because the city kept importing the gap it was trying to close.

Every new citizen is a new mouth the local farming sector was never resourced to feed. The 30 by 30 goal was chasing a number that the immigration pipeline was moving faster than any agri-tech startup could follow.

The government's response — diversify imports, forge government-to-government supply agreements, stockpile longer-shelf-life goods, study a multi-tenant farming facility — is sensible crisis management. It is not food security. It is permanent structural dependency, managed with greater sophistication than before.

Singapore will always need to import the overwhelming majority of what it eats, because Singapore will always be adding people faster than it can grow food for them. The infrastructure being built to manage that dependency will be paid for, through levies and taxes, by the same residents the dependency was created to serve.

Dr Tan See Leng announced $800 million for a semiconductor R&D flagship, a new $800 million Decarbonisation Grand Challenge, nuclear energy feasibility studies, and — in the same speech — significant investment in what he called Embodied AI: "robots that can sense their surroundings, reason independently, and act with purpose in unfamiliar environments," targeting manufacturing, aviation, and maritime sectors. He also announced energy infrastructure sufficient to power a substantially larger and more energy-intensive economy: solar targets raised by 50 per cent, nuclear feasibility studies accelerated, electricity import corridors from Indonesia and Malaysia under active development.

Singapore's own government statistics project total water demand to nearly double by 2065. The food resilience strategy implies a population that needs feeding through prolonged disruptions. The energy expansion implies an economy consuming power at a rate that current trajectories cannot support without new supply.

None of these projections were stated explicitly in any of the three speeches. They did not need to be. The population growth assumption is not a policy choice being debated. It is the unexamined foundation on which every other policy choice rests.

The Hand That Builds and the Hand That Automates

Dr Tan's speech contains, in close proximity, two things that are rarely placed in the same frame.

He announced Singapore's ambitions in Embodied AI — physical robots targeting advanced manufacturing, aviation, and maritime — and in the same session announced energy infrastructure sufficient to power the substantially larger economy those sectors are assumed to support.

The question these two announcements together invite is simple: who is this economy for?

If the Embodied AI investment succeeds, the sectors targeted for that technology would employ fewer people per unit of output than they do today. The energy infrastructure built to power those sectors may, over time, serve capital considerably more than it serves labour. The population growth being planned to sustain those sectors may find, on arrival, that the jobs the planning assumed have been automated.

The same logic applies to what is being cleared to build this economy. In February 2026, Deputy Prime Minister Gan Kim Yong defended the government's decision to clear 52 hectares of forest in the western corridor for the expansion of the Jurong Innovation District.

The justification was direct: the development would generate approximately 95,000 jobs in advanced manufacturing, robotics, medical technology, and smart logistics. Forest containing at least 56 conservation-significant species — including the Sunda pangolin — sits within an ecological corridor connecting the Western Catchment area with Tengah forest. It will not be replaced. The 14.5 hectares of retained green space is not a corridor. It is a gesture.

What is worth asking is whether the 95,000 jobs that justify this irreversible trade-off will exist on the timeline the planning assumes.

The sectors named — advanced manufacturing, robotics, smart logistics — are precisely the sectors Dr Tan See Leng identified, in the same parliamentary session, as the primary targets for Embodied AI deployment. Singapore is clearing forest that cannot be restored to build jobs for an economy its own technology investment strategy is in the process of automating. The ecological cost is permanent. The employment projection may not be.

The data centre story makes this concrete. Singapore's electricity consumption grew 4.0 per cent in 2024, with the Information and Communications sector up 12.3 per cent — substantially the data centre sector. A data centre consuming tens of megawatts of power and requiring significant water for cooling may employ, in ongoing operations, a few dozen people. The same capital invested in manufacturing or services of an earlier era would have employed hundreds or thousands.

The government's own 2017 projection stated that by 2060, total water demand could double — with the non-domestic sector, not households, accounting for approximately 70 per cent of that growth. Residents who absorbed the 30 per cent water price hike were, in significant part, paying to build infrastructure whose primary projected beneficiary was always industrial and commercial demand.

That demand is now increasingly dominated by data centres growing at 12 per cent annually in electricity consumption — an industry that employs relatively few people, runs AI that is displacing more, and whose resource consumption is paid for by the same workers it is helping to automate out of their jobs.

Neither the data centre trade-off nor the automation-versus-immigration contradiction was visible in the chamber. Nobody appears to be holding both simultaneously and asking what the intersection implies for population planning, infrastructure sizing, and the question of who bears the cost when the assumptions diverge from reality.

Grace Fu's 30-Year Bet

The Coastal Protection Bill tabled in the same week as the Committee of Supply debates is perhaps the most consequential infrastructure commitment Singapore has made in a generation.

A continuous coastal defence line around the entire island is not a five-year project. Its cost will be measured in the tens of billions and its useful life will extend to the end of this century. Minister Fu confirmed that the Government and Statutory Boards own approximately 70 per cent of the coastline subject to coastal protection obligations. The cost flows not through regulatory compliance on private landowners, but directly from the public purse. Taxes and tariffs are the mechanism.

A coastal defence line is also, implicitly, a commitment to a city of a certain size and density. You do not spend what Singapore will spend protecting every metre of coastline for a city you expect to shrink. The engineering parameters, the cost-benefit analysis, the financing structure — all of it embeds assumptions about how many people will be living inside the defended perimeter for the next seventy years.

There is one further detail from the Bill's miscellaneous provisions that deserves attention. The legislation amends Section 7 of the Government Proceedings Act to include coastal management. The Government is simultaneously expanding its infrastructure mandate and updating the statutory framework governing its own liability. The public will fund the infrastructure. Whether they can hold the government accountable if that infrastructure fails — on a timeline of seventy years — is a question the Bill does not invite.

When Fu introduced the legislation, she acknowledged on the record that private landowners along the coast had told the government they had not factored coastal protection obligations into the cost of their leases when they signed them. The obligations materialised after the commitment was made, not before.

The existing Singaporean's relationship to the broader infrastructure programme is structurally identical. Nobody priced in a 70-year coastal defence line when they signed up to pay their water bills. Nobody priced in electricity tariffs that include cross-subsidies for a data centre sector growing at 12 per cent annually. Nobody priced in CPF contributions capitalising infrastructure serving a population size that has never been stated.

In return, they were told — are still being told — that this growth is for them.

Building for a Future Someone Should Explain

Singapore is, right now, committing to infrastructure that will define the physical, demographic, and economic shape of the city for the rest of this century.

The coastal defence line will be there in 2080. The desalination plants will be pumping in 2060. The food stockpile infrastructure will be calibrated to feed a population size determined by decisions being made today. The nuclear feasibility studies, if they proceed, will result in energy capacity serving a demand projection whose underlying population assumptions are being made in 2026.

The people who will live inside that infrastructure — who will pay the tariffs and taxes that service its cost, who will work in the economy it is designed to support — have not been told, in any honest and integrated way, whether the plan that produced it was stress-tested against what is actually coming.

What is actually coming, on the evidence available in public documents, is an automation wave that will reshape the labour market on a timeline measured in years, not decades.

A citizen natural increase approaching zero. A population pipeline sustaining citizen numbers through the conversion of an ageing cohort. A resource base — water, energy, land — being committed to industries whose employment contribution per unit of resource consumed may not justify the allocation, and whose AI output is accelerating the displacement of the workers those resources were notionally being developed to serve.

None of this required a conspiracy. It required only the ordinary operation of institutional momentum — each ministry building what its mandate requires, each planning cycle inheriting the assumptions of the last, nobody responsible for asking whether the whole still adds up.

The existing Singaporean — local-born, naturalised, here by choice or by birth — is entitled to ask that question.

So far, no one in a position to answer it has been asked.

---

 This is the third in a series on AI, automation, and Singapore's economic future. Read the first piece on education and AI displacement [here] and the second on the demographic pipeline and citizen natural increase [here]. The fourth asks the question the series has been building toward: what population is Singapore actually planning for, and who bears the cost when the plan runs out of road.

This series is the foundation of a longer work. If you think these arguments deserve a full book — with the data, the charts, the parliamentary record, and the regional comparisons this format couldn't hold — support the project at https://buymeacoffee.com/theonlinecitizen. Every contribution goes toward making that book happen.