Construction is up 6.5% this year alone. Private finance is sitting at 8.95 to 12.95 percent. Council DA targets just changed to 95 days and most councils are already missing them. Here is the full picture for anyone sitting on a subdivisible site in NSW right now.
We get some version of the same call most weeks: "I've got an 800sqm block in [suburb], been told it can be subdivided, the numbers don't seem to work. What am I missing?"
Usually they're not missing anything. The numbers just don't work right now. Here is why, with the actual 2026 figures.
Sydney residential construction is now running at $2,800 to $3,800 per square metre for mid-range spec and $3,800 to $6,000-plus for premium or architect-designed builds (Buildana, 2026). A duplex on two 150sqm dwellings sits at $600,000 to $840,000 in construction cost alone before civil works, contributions, or professional fees are touched.
The Altus Group Q1 2026 construction price outlook puts Sydney at 6.5% cost escalation for the year, revised up from an earlier forecast of 4.5%. The same 250sqm home quoted at $680,000 in May 2025 is now $710,000 to $725,000 on identical specifications. Material cost movements driving that: concrete up 6.1%, electrical labour up 6.8%, plumbing labour up 5.1%, timber framing up 3.8%. Copper is up 25% year-to-date. Diesel jumped 41% in March alone, and diesel is a cost multiplier across every material that gets trucked to a site.
Budget 10 to 15% contingency on top of whatever number you get. Tree removal, retaining walls, stormwater redesigns, unexpected soil conditions, survey amendments: these show up on nearly every project and they are not small items.
NSW council infrastructure contribution rates were indexed on 29 April 2026. If you ran your feasibility six months ago, run it again.
NSW Contribution Costs Per Lot, 2026 Indexed Rates
"In growth area councils, contributions alone can reach $55,000 to $85,000 per lot. For most sites at current asking prices, that is the entire margin."
Stack those numbers and contributions alone can reach $55,000 to $85,000 per lot in growth area councils. That is not a rounding error. In a lot of outer corridor deals, that figure is larger than the entire land value uplift the subdivision is supposed to generate. The project only works if the site was acquired at a price that leaves room for it. Most sites currently listed are not priced that way.
Private development finance in Australia is sitting at 8.95 to 12.95% per annum on the senior facility, with mezzanine debt running 14 to 22% for borrowers who need to reduce their equity contribution (Innovate Funding, 2026). On top of that: establishment fees of 1.5 to 2.5% of loan value, plus QS inspection fees at $1,500 to $3,500 per draw.
APRA has continued tightening commercial real estate exposure limits for the major banks, which means a growing share of private developers are being pushed toward non-bank and private lenders at the higher end of that rate range. On a $500,000 loan at 10%, each additional month of project time costs approximately $4,200 in interest. A project that runs 18 months instead of the 12 months budgeted loses $25,000 in unplanned finance costs before a single title is registered. That is 5 to 10% of project margin gone purely from a timing miss.
The developers making money right now do not treat finance cost as a fixed line item. They treat it as a function of time. Every week of delay has a number. Every week of acceleration does too.
From 1 July 2026, NSW councils are required to determine development applications within 95 days on average, reduced from the previous 115-day target. The statutory requirement is still 40 to 60 days. The practical reality across most Sydney councils has not changed materially.
In the last published data: Camden averaged 81 days, Penrith 98, Northern Beaches 113, North Sydney 266, Georges River 289. Seven Sydney councils averaged above 215 days. The new 95-day target, even if councils hit it, is still more than double the legal requirement and considerably longer than most feasibilities assume.
For a subdivision, the full timeline from DA lodgement to title registration runs 8 to 16 months in NSW under normal conditions. A request for additional information pauses the statutory clock entirely. A neighbour objection can add months. Each one feeds directly back into Reason 3.
The underlying model for residential subdivision is straightforward: acquire a site, split it, sell the resulting lots for more than you paid. The spread between acquisition cost and end product value is where the return lives. In 2020 and 2021, that spread was wide. In 2026, in most outer Sydney and Hunter Valley corridors, it is not.
Land values in outer corridors ran hard between 2020 and 2022. NSW land values sit at an average of $1,027 per square metre across the state, with growth corridor sites considerably above that. The end product values for finished lots and house-and-land packages in those same corridors have not kept pace with the full cost stack: construction up 6.5% this year, contributions indexed in April, finance north of 9%. The residual land value calculation (what you can afford to pay for a site after stripping out all costs and a real margin) in many of these corridors now barely justifies current asking prices.
B-grade sites in average outer locations are sitting longer. That compresses exit timing and adds holding cost back into a feasibility that was already marginal.
Infrastructure Australia's 2025 Market Capacity Report put Australia's projected construction worker shortfall at 300,000 by 2027. The trades and labourers component alone hits 126,000 by mid-2027. More than 25% of Australia's current construction workforce is over 55. The pipeline for replacing that cohort through training, immigration reform, and modular construction cannot close a gap that size in two years.
"Projects will cost more, take longer, or both."
Infrastructure Australia, 2025 Market Capacity ReportThat quote is from 2025. Nothing in the 2026 data suggests it has become less true.
Between 2020 and 2022, selling lots in greenfield corridors was close to frictionless. Off-the-plan in six weeks, strong prices, low days-on-market. That market exists for a small number of premium pockets. Everywhere else, the picture is different.
Australia is tracking to fall 262,000 homes short of the 1.2 million National Housing Accord target by 2029. NSW is among the worst-performing states for target attainment. Yet private residential development is struggling to get off the ground while publicly funded infrastructure construction is booming. The constraint is not demand. The constraint is the cost stack and the finance environment making private projects unviable at scale.
NSW dwelling approvals ticked up 3.2% in March 2026, but one month of movement does not change the structural picture. A-grade sites in genuine undersupply locations are still trading. B-grade sites in average corridors are not.
Timing risk is back. A subdivision that takes 18 months to complete does not exit into the market you modelled at acquisition. It exits into whatever exists 18 months later. In a corridor where absorption has slowed, that is a different feasibility to the one you started with.
So
Subdivisions are not dead. We are managing subdivision projects right now that are producing real returns. The ones that work share a few things the ones we have walked away from did not.
The site was bought at a price that reflects the actual 2026 cost stack, not the 2022 one. The contributions were modelled precisely for that council's indexed rates, not estimated. The DA pathway was chosen for the specific council's performance history, not defaulted to. The finance structure was built around a realistic timeline, with a buffer. The exit is into a genuine pocket of undersupply, not a corridor where everyone else is also subdividing.
If you are sitting on a site and wondering whether the numbers actually work in this environment, the answer requires a real feasibility study with current inputs. If you want to run that, talk to us. We run feasibility first, every time.
Most feasibilities we see were built on 2022 or 2023 inputs. If yours was too, the gap between what you think the project returns and what it actually returns in 2026 could be significant. We run it properly before anything else.
More from Blackark Insights