Multifamily starts in the first quarter of this year came in at around 55,000 units. That’s the lowest quarterly total since 2011, and roughly 73 percent off the peak in early 2022. Units under construction are down more than half from their 2023 level. Deliveries were 695,000 in 2024, 531,000 in 2025, and are tracking near 382,000 this year — with another sharp drop expected next year.
Vacancy is projected to peak in early 2027 and tighten after that. Rent growth this year is running around half a percent.
Put those two facts next to each other, and you get the whole picture: the market is soft today, and the pipeline that would supply it in 2028 is being emptied out right now, because almost nobody can make a deal pencil at today’s costs and today’s rates.
That means whoever can deliver units in 2028 will be delivering into a very thin market. That’s the opportunity, and it has a clock on it.
I’d say this even if I didn’t build them, because it lines up with exactly what’s broken.
Schedule is worth more when the window is timed. Modular typically takes 30 to 50 percent off a construction schedule. On a 100-unit project, that can be the difference between roughly 18 months and 10 to 12. In a normal market, that’s margin. Right now it’s the difference between delivering into a tightening market and delivering into the tail end of the glut. Same building, very different rent roll.
A hard number does something specific in this capital environment. Costs have moderated, but contractors are cautious, and they’re building contingency into their pricing rather than cutting margin. That contingency sits on your budget whether you use it or not. We can put a firm number on a third to a half of your project cost 12 to 18 months out and hold it. Right now capital is underwriting risk at least as hard as it’s underwriting cost, and a fixed number on the largest line is worth real basis points.
Labor hasn’t gotten cheaper — and it isn’t going to. Labor costs haven’t come down enough to make development work in most metros, and there’s no reason to expect they will. Factory labor is a different pool with different economics — one building, one trained crew, work that happens indoors year-round.
None of that makes a bad deal good. It makes a marginal deal faster and more certain, and in this market that’s often the whole difference.
This is the part we get asked about least and should get asked about most. Pricing a modular multifamily building is not pricing a site-built one with a different subcontractor.
You’re buying modules, not square feet. The module is the unit of cost. Two floor plans with identical square footage can yield 48 clean modules or 55 awkward ones, and those are very different buildings to us. Get us into the conversation at schematic, and we’ll tell you which one you’ve drawn while it’s still cheap to change.
Repetition is the whole game. A building with four unit types prices differently than a building with eleven. Same layout, same crew, same materials, same sequence — that’s where factory economics actually live. Every additional unit type is its own engineering package, its own approval, and its own learning curve on the floor. If you want one lever to pull on cost, that’s the lever.
Transport sets your dimensions. Module width is bounded by what can legally travel the roads between South Paris and your site. Design to that from the beginning and you give up almost nothing. Draw a 34-foot-wide module, and you’ve drawn a building nobody can ship.
You’re paying for some structure twice. Every module has to be self-supporting for the delivery and set. Modular utilizes independent floor and ceiling assemblies where a site-built building shares one assembly. That’s real material and real cost, and it’s a key reason a module doesn’t come in 40 percent under stick framing. It’s also why schedule and price have to carry the argument — which is the same thing I said in the first post and will keep saying.
General conditions are where the number actually moves. Fewer months of site supervision, trailers, temp power, security, insurance, and interest carry. Developers who set our module price next to a framing bid miss this every single time. Compare all-in delivered cost, or the comparison isn’t telling you anything.
Raise the draw schedule in month one. Lenders release construction funds as work progresses on site, which is a poor fit for a building that gets built indoors months before it arrives. This is the single biggest friction in a modular multifamily deal. It is also solvable — but it’s solvable in month one with a lender who’s done this before, and it’s a fight in month six with one who hasn’t.
Podium buildings with complicated structural transitions get harder. Highly irregular buildings with a lot of one-off geometry give back the repetition advantage. Sites where a crane can’t work are a non-starter no matter how good the rest of it looks.
If your project is one of those, we’ll tell you early.
The supply cliff is coming in 2028, and the people who will be standing on the right side of it are making decisions in the next few quarters, not in 2027.
Modular is a tool for that specific problem: it compresses the schedule, and it converts the largest line on your budget from an estimate into a number. In a market where nobody can get comfortable, certainty is the scarce commodity — and it’s the one thing a factory is genuinely good at producing.
If you’ve got a site and a pro forma that’s close but not there, send it over. We’d rather look at it now than read about it in 2028.
Next week I want to write about something different: the people in South Paris who actually build these buildings, and why a factory job in this industry is a different proposition than a jobsite one.
— Thatcher
