A Record S$42.4 Billion. A Quarter of It Was the Government Selling Land.

Singapore's big-ticket real estate investment sales reached S$42.4 billion in the first nine months of 2026, a record — above all of 2025, and above the S$35.5 billion set in 2017. Knight Frank, whose quarterly bulletin produces the figure, now expects the year to close at about S$50 billion, having forecast about S$30 billion in April and "on par with last year" in July.

The headline is real. The composition is the part worth reading, and it is checkable against primary documents.

Residential became the single largest sector in the third quarter at S$4.3 billion. URA awarded four private residential land tenders in July, August and September. Add them up from the award notices: S$4,338,265,554. The quarter in which residential led the market was a quarter in which residential investment was, to within rounding, the state selling land.

What "investment sales" counts

Knight Frank's definition is broad by design: transactions of at least S$10 million, bulk or land sales of S$10 million, and institutional transfers where a REIT or fund exchanges an asset. That means a government land tender, a REIT selling a mall to its own sponsor's fund, and a collective sale all land in the same number.

It is a measure of capital changing hands in large blocks. It is not a measure of what homes cost, and the two can move in opposite directions — which is roughly what is happening.

The quarterly path runs the other way to the headline

Each quarter has been smaller than the one before it, and the third is 34% below the first. A nine-month record built from a declining sequence is a statement about the first quarter, not about current conditions.

The full-year guidance makes the same point arithmetically. S$50 billion less S$42.4 billion leaves about S$7.6 billion for the fourth quarter — which would be the weakest quarter of 2026 by a wide margin, and weaker than any quarter since the middle of 2025. The "about" in "about S$50 billion" is doing real work here, but no reading of it produces a strong Q4.

One caveat on the series itself: these quarterly numbers have been restated upward as the year went on. First quarter was published in April as S$15.4 billion and now reads S$16.2 billion; second quarter was published in July as S$15.1 billion and now reads S$15.5 billion. Full-year 2025 was published in January as S$40 billion and now reads S$40.7 billion. The restatements are small individually and all in the same direction, which matters when the story is "a record was broken."

The residential number is a land number

This is the finding. Knight Frank's 3Q2026 residential figure is S$4.3 billion and is described as comprising four private residential GLS sites. URA's own award notices give those four:

The match to the reported S$4.3 billion is to within rounding, and no residential collective sale completed in the third quarter. So private, non-state residential investment sales in 3Q2026 were close to nil.

Extend it across the year and the same shape holds. Residential ran S$4.4 billion in the first quarter, S$5.3 billion in the second and S$4.3 billion in the third — about S$14.0 billion, or 33% of the S$42.4 billion total. Government land sales inside that were roughly S$3.2 billion, S$3.2 billion and S$4.34 billion: about S$10.7 billion, which is 76% of residential investment sales and 25% of all big-ticket investment sales this year. The quarterly GLS components for the first two quarters are as reported at the time rather than computed from award notices, so treat the 76% and 25% as close estimates, not published figures.

The residual — roughly S$3.3 billion of non-state residential across nine months — includes exactly one collective sale: Loyang Valley, S$880 million, April 2026. In 2025 there was also exactly one, Chiku Mansions at a little over S$22 million.

Where the private money is not going

If a record investment-sales year meant broad private appetite, the smaller tickets would show it. They do not.

Shophouses, the clearest small-ticket private market in Singapore, are down by a quarter. Fifty-one shophouses sold in the first nine months of 2026 for S$441.3 million, against 68 for S$644.5 million in the same period of 2025 — −25% by count and −31.5% by value. The third quarter alone was S$157.7 million, 18.6% below the second quarter's S$193.7 million. Average deal size fell from S$12.1 million in 2Q to S$7.5 million in 3Q.

Developers are bidding more and competing less. Across the twelve private and mixed residential GLS tenders awarded in 2026, 51 bids were submitted — 4.25 bids per site. Two tenders drew a sole bid, Holland Plain in May and Berlayar Drive in August. There were no sole-bid residential tenders in 2025. The prices kept setting benchmarks anyway: Berlayar Drive's single bid came in at S$1,515 psf per plot ratio, above the S$1,100 to S$1,450 range analysts had forecast.

That combination — record aggregate value, record psf, four bids a site — is what a market looks like when the supply of buyers thins faster than the supply of capital.

The quarter is concentrated. The four largest confirmed third-quarter deals (Bayshore Drive, New Upper Changi Road, Wheelock Place at about S$1.1 billion to a Hongkong Land–managed fund, and Berlayar Drive) total S$5.23 billion, or 48.9% of the S$10.7 billion quarter.

The cost of money is the actual story underneath

Three-month compounded SORA fixed at about 1.23% on 7 October, with overnight SORA at about 1.25%. That is roughly a third of comparable US rates. Cheap domestic funding is a better explanation for institutional blocks trading at record aggregate value than any thesis about Singapore residential demand — and it is the one variable that can reverse without any policy change at all.

What to do with this

If you are reading this as a signal for home prices, don't. URA's 3Q2026 flash estimate has overall private residential prices up 1.4% for the quarter, but CCR down 0.1% while OCR is up 2.2%. A record investment-sales total driven by state land sales in the Rest of Central and Outside Central regions and by institutional transfers in the CBD tells you where capital went, not where prices are going. Full third-quarter data lands on 23 October.

If you are a prospective buyer in a new launch on one of these sites, the land price is the floor under your eventual asking price. New Upper Changi Road at S$1,537 psf per plot ratio is a record for an Outside Central Region site; Bayshore Drive went at S$1,323 psf ppr. Those numbers reach you in two to three years as launch prices, and they were set by developers bidding against three others, not thirteen.

If you own in a development hoping for a collective sale, the count is one, in April, in nine months. Developers with land bank are buying it from the government, on 99-year leases, at prices they set in thin tenders. That is a more orderly supply of sites than a collective sale, and it is the competition your committee is up against.

If you are watching for the turn, watch bids per tender and SORA, not the aggregate. The aggregate is a cumulative total that cannot fall until the year resets.

Method

The S$42.4 billion nine-month figure, the quarterly series, the S$4.3 billion third-quarter residential figure, the S$50 billion full-year expectation, the comparison years and the definition of big-ticket investment sales are from Knight Frank's Q3 2026 Singapore Investment Market Update of 6 October 2026, as reported by Stacked Homes on 9 October; the report itself is listed by Knight Frank but the document is gated, and at the time of writing no other outlet had published the third-quarter figures, so they rest on a single secondary source. The earlier quarterly figures and forecasts, and the first- and second-quarter sector and GLS splits, are from EdgeProp's April and July reports on the same bulletin. The four third-quarter land awards and their prices are from URA's award notices: Bayshore Drive, Berlayar Drive, New Upper Changi Road and Lorong Puntong; the S$4,338,265,554 total is ours and excludes the Chitty Road / Veerasamy Road site awarded on 7 August at S$35,288,000, which is not residential. Whether Knight Frank books the mixed-use Bayshore Drive site wholly under residential or splits it is not disclosed; if it splits it, the residential figure would include some non-GLS component we cannot see. The 76% and 25% GLS shares combine our URA sum for the third quarter with reported GLS figures for the first two, and are estimates. Collective sale counts and the Loyang Valley and Chiku Mansions figures are from EdgeProp's en bloc tracker, which is built on lodged caveats and which EdgeProp notes may understate activity. Shophouse figures are Huttons Data Analytics on URA caveats as at 8 October, via EdgeProp; the average deal sizes are ours. Bid counts and the sole-bid comparison are from our own GLS tender review. The flash-estimate figures are from URA. SORA fixings for 7 October 2026 are as published by rate trackers citing MAS daily data; MAS's own statistics pages were unavailable when we checked, so treat the exact decimals as indicative. The S$7.6 billion implied fourth quarter is simple subtraction from an "about" figure and is not a forecast. This is general market information and not investment advice.