A Sentosa Cove apartment rents for more than a prime mainland one of the same size, and costs less to buy.

A Sentosa Cove apartment rents for more than a prime mainland one of the same size, and costs less to buy. Both are true in every size band we can measure, and together they make the enclave yield about 40% more.

That is an odd thing to be able to say about the worst-performing address in Singapore for capital. It is also the reason: fifteen years of falling prices, and rents that never fell with them.

For anyone who bought at the launches it is still not enough. Sixteen years of rent recovers about 60% of what they are down on price, and no group of launch buyers ends level.

Key takeaways

• Size for size, Sentosa rents for 2.5% to 21% more than prime mainland. The pooled figures look level only because Sentosa apartments are twice the size. Rent, size for size
• Size for size, Sentosa costs 4% to 29% less to buy. Price, size for size
• The result is a yield of 3.07% to 3.96% against prime mainland's 2.33% to 2.84%. Five size bands, and Sentosa leads in all five. The yield in every band
• Rent per square foot rose 36% since 2009 while prices rose 12%. Rent held, prices did not
• Sixteen years of net rent covers about 60% of a 2007 buyer's capital loss. Does rent make it back
• An owner from 2007 has held nineteen years and needs about nine more before rent alone makes them level. If they are still holding

Rent, size for size

Comparing the two markets straight across, rent per square foot against rent per square foot, gives the wrong answer.

A Sentosa Cove apartment has a median floor area of 2,099 sqft. A prime mainland one is 1,023 sqft, less than half. Rent per square foot falls as apartments get larger, so a straight comparison sets Sentosa's large homes against the mainland's medium ones and makes Sentosa look cheaper to rent than it is. Done that way the two look level. Matched on size they are not.

Matched on size, leases from 2024 to 2026:

Size	Sentosa Cove	Prime mainland	Sentosa rents more by
969 to 1,499 sqft	$5.33	$5.20	2.5%
1,500 to 1,999	$5.02	$4.78	5.0%
2,000 to 2,499	$4.97	$4.76	4.4%
2,500 to 3,199	$5.53	$4.56	21.3%
3,200 and over	$4.57	$4.01	14.0%

Five bands, and Sentosa is ahead in all five.

The 2,500 to 3,199 band is where the enclave's actual appeal shows up. A tenant wanting a large waterfront apartment has few alternatives on the mainland, and pays 21% more per square foot for it here. That is the only band where the premium is large, and it is the band Sentosa was built for.

Price, size for size

The sale market runs the other way, in every band.

Size	Sentosa Cove	Prime mainland	Sentosa costs less by
969 to 1,499 sqft	$1,614	$2,196	26.5%
1,500 to 1,999	$1,625	$2,163	24.9%
2,000 to 2,499	$1,571	$2,227	29.5%
2,500 to 3,199	$2,164	$2,246	3.7%
3,200 and over	$1,768	$2,061	14.2%

Resales from 2024 to 2026. Sentosa transaction counts here are small, from 12 to 45 in each band over three years, so read the direction rather than the precise gap.

The 2,500 to 3,199 band is the exception again, and in the same direction: it is the one place where Sentosa prices come close to the mainland's, and the one place where its rent premium is largest. That band is the enclave working as intended.

The yield in every band

Put the two together.

Size	Sentosa yield	Prime mainland	Sentosa ahead
969 to 1,499 sqft	3.96%	2.84%	1.39×
1,500 to 1,999	3.71%	2.65%	1.40×
2,000 to 2,499	3.80%	2.56%	1.46×
2,500 to 3,199	3.07%	2.44%	1.26×
3,200 and over	3.10%	2.33%	1.32×

Five bands, five wins, and the ratio sits between 1.26 and 1.46 throughout. Nothing here depends on a pooled average.

And the cause splits in two. Sentosa rents for more, size for size, by a little in most bands and a lot in one. And it costs substantially less. Yield is rent divided by price, and both halves point the same way.

The larger half is price. A 2,000 to 2,499 sqft apartment rents for 4.4% more here and costs 29.5% less, so most of that 1.46 times yield advantage is the discount, not the rent.

Rent held, prices did not

The reason the discount exists is fifteen years of falling prices against rents that held.

Rent per square foot went from $3.78 in 2009 to $5.15 in 2026, up 36%. Prices for apartments of the same size went from $1,400 to $1,571, up 12%.

So the rental market outran the sale market three to one, in the same buildings, over the same years. An owner watching only the rent would have thought the enclave was doing reasonably well. We covered what was happening to the prices in an earlier note: the launches priced far above the level the resale market has ever sustained, and it never caught up.

Does rent make it back

Now the question an owner actually asks. Someone who bought at a launch, rented the apartment out, and sold in 2025: does the rent close the gap?

The capital side of this table is the finding from that earlier note. Every buyer from 2007 to 2011 sold below what they paid, without exception.

Bought new in	Paid	Sold 2025	Capital	Net rent	Where they end
2007	$2,563	$1,626	-$937	+$561	-$376
2008	$2,734	$1,626	-$1,108	+$561	-$547
2009	$3,252	$1,626	-$1,626	+$561	-$1,065
2010	$2,672	$1,626	-$1,046	+$532	-$514
2011	$2,719	$1,626	-$1,093	+$497	-$596

Gross rent from 2009 to 2025 comes to $863 per square foot. Net of costs it is less, and the table assumes 65% of gross survives property tax, maintenance, agent commission and vacancy. At 55% the 2007 buyer ends at -$462; at 75%, -$290. Behind on any assumption.

Buyers from 2007 recover about 60% of their capital loss through rent and are still $376 per square foot down. On a 2,000 square foot apartment that is a shortfall of roughly $750,000, after sixteen years of collecting rent.

The yield on what they paid explains it. A 2007 buyer at $2,563 collecting today's rent is earning about 2.4% gross. The rent is fine. The price they are earning it against is not.

If they are still holding

Most owners have not sold. For them the question is how many more years of rent, at today's level, closes what is left.

Counting from purchase to August 2026, at a 65% net margin:

Bought new in	Held so far	Paper loss	Rent so far	Still short	More years of rent
2007	19 years	-$926	+$561	-$365	9.1
2008	18 years	-$1,097	+$561	-$536	13.3
2009	17 years	-$1,615	+$561	-$1,054	26.2
2010	16 years	-$1,035	+$532	-$503	12.5
2011	15 years	-$1,082	+$497	-$585	14.6

A buyer from 2007 has held nineteen years and needs about nine more to be level on rent alone. That is 2035, twenty-eight years after they bought.

A buyer from 2009, who paid the highest launch price the enclave ever saw, needs about twenty-six more years. That is 2052, forty-three years after purchase, on a lease that runs to 2108.

At a 55% net margin those figures run from thirteen years to thirty-four. At 75% they run from six to twenty-one.

What this calculation assumes. It holds the sale price still and the rent still, and asks what rent alone would have to do. That is not a forecast. If Sentosa prices rise the gap closes faster and rent is not what closed it; if rents fall back to the $3.70 they sat at from 2016 to 2020, every figure above stretches by roughly a third.

What this means

Two markets, same buildings, different verdicts. As the project-level note showed, outcomes inside the enclave varied enormously. Sentosa's rental market says the enclave is worth a small premium over prime mainland, and a large one for big waterfront homes. Its sale market says it is worth a quarter less. Both have been saying it for a decade.

Yield is mostly a statement about price. Sentosa's rent premium is real but modest outside the largest band. The 1.4 times yield advantage is roughly a quarter rent and three quarters discount.

Rent is a consolation, not a rescue. For the launch buyers it recovered most of the loss and closed none of it. Sixteen years is a long time to be made whole and not be.

And the buyer today is in a different position from the buyer in 2009. Same building, same tenant, same rent. One is earning 3.8% on what they paid and the other 2.4%.

What this does not tell you

• What any specific owner earned. A rental record lists the development, a size range and the rent. It does not say which unit. So we can say what an apartment of that size in that building rented for, not what one owner earned on the home they bought.
• How firm the size-banded prices are. Sentosa resales run from 12 to 45 transactions per band over three years. The direction is consistent across all five bands; the individual gaps are not precise.
• The true net margin. We show 55%, 65% and 75% because costs are not in the record. Property tax on a home that is not owner-occupied, maintenance on a waterfront development, agent commission and vacancy all vary, and a foreign owner's tax position differs again.
• The first years for 2007 and 2008 buyers. Our lease records begin in 2009, so those buyers collected rent we cannot count. Their true recovery is higher than the table shows, by roughly one to two years of rent.
• Whether today's yield holds. It assumes rents stay near their current level and prices near theirs. Rent stepped up sharply in 2022 and 2023 and has eased since.

Method

Rental figures are from URA rental contract records for the Sentosa Cove non-landed projects, excluding records with no floor area or no rent. Sale figures are transacted prices from URA caveat data for the same projects.

The Residences at W Singapore is excluded throughout. Its sales history is concentrated in a way that does not describe a market: twenty sales in 2010, a handful between 2012 and 2014, then eighty-three in a single year in 2024. Those recent buyers have not had time to sell again, so the project contributes almost nothing to any resale measure, and including its leases would blend a one-off distribution into a market series. Our earlier Sentosa Cove notes exclude it on the same basis.

Rent per square foot uses the midpoint of the recorded floor-area band, which URA publishes as a range rather than an exact size.

The prime mainland comparison is Core Central Region non-landed rentals and resales excluding Sentosa Cove.

Sentosa Cove is part of the Core Central Region under URA's definition, so prime mainland here means the rest of that region rather than a different one.

Size bands are applied to both the rent and the price side of every comparison, so each yield divides like by like. Bands are set at 969, 1,500, 2,000, 2,500 and 3,200 sqft.

Gross yield is twelve months of rent per square foot divided by price per square foot, before all costs.

The recovery and still-holding tables use whole-enclave prices rather than size-banded ones, matching the launch prices our earlier note establishes. The launch prices here are the pre-2015 caveat figures from that note, recorded before any discount or rebate, so read them and the losses built on them as an upper bound; if they overstate what buyers paid, rent recovers more of the loss, not less. Applying a size band to the 2009 launch figure would rest it on three sales.

Holding periods are counted from the year of purchase to August 2026. The forward calculation holds both the sale price and the rent at their current level and asks how many further years of net rent close the remaining gap. It is arithmetic, not a projection.

Rental data to June 2026. Sale data to 4 August 2026.