Workforce housing near data centers pencils when three things line up: a real demand signal with a multi-year lag, a cost structure you actually control, and a refinance assumption that survives contact with the rate market. The announcement itself is not a deal. It tells you construction crews and permanent operations staff are coming, and that the housing to hold them does not exist yet.
Mark Seruya is a partner in Windsor Properties, a roughly 60-person vertically integrated developer based in Columbus, and runs capital markets for KSR in New York. He’s building into the Columbus–Dayton corridor, where he estimates $50 to $60 billion of data center and manufacturing facilities are going up. He’s also carrying two deals underwritten five or six years ago that are hitting refinance against rates that have since doubled.
What follows is how to read the announcement, what build cost discipline rent-capped product demands, how to stress-test the refi, how to match investors to deals, and how one third walkthrough killed a deal that looked good on paper.
Key takeaways
- Megaproject announcements create a lagged demand signal — cranes and construction labor arrive years before permanent housing supply catches up, which is the window developers are buying into.
- Rent-capped workforce housing forces cost control rather than revenue optimism. Windsor builds everything in-house, down to cabinets and roofs, and owns a tree farm in the Carolinas after COVID lumber shocks.
- Underwriting off 2–3% money required assuming those rates would persist. Seruya says that was never a good assumption, and two of his developments now need fresh equity or creative financing at refinance.
- Know the investor before the property. Risk tolerance and asset-class preference vary, so a balanced book might pair low-basis New York office for total return with East Coast shopping centers for income.
- Walk a retail deal at least three times and talk to anchor store management directly. KSR passed on a Richmond-area Walmart center after a third walkthrough surfaced shrink and site issues the first two missed.
From the Real Estate Pros Show
This article draws on an interview with Mark Seruya of KSR (Kassin Sabbagh Realty) and Windsor Properties on the Real Estate Pros Show, hosted by Issa Hanna.
Reading a Megaproject Announcement as a Housing Signal
A data center or plant announcement is a demand signal with a lag, not a buy trigger. The construction jobs land first — hundreds of cranes and thousands of trades workers — then the permanent operations staff, then the service economy that supports both. Housing supply does not respond on that schedule, which is the gap a developer is buying into.
Seruya’s thesis centers on the corridor between Columbus and Dayton, Ohio, where Meta, Google, Amazon and Intel have committed to data centers and manufacturing facilities. He puts the total at roughly $50 to $60 billion, and describes a severe shortage of workforce housing across those submarkets. Ohio’s draw is straightforward: business-friendly posture, low cost of living relative to either coast, and a national push to bring manufacturing back domestically.
Dayton is the quieter half of the story. It was a technology hub in the early 1900s and still hosts Wright-Patterson Air Force Base. Seruya points to a reported $300 million bid to bring a military academy to the area as a second, independent demand driver on top of the corridor build-out.
The pattern is not unique to Ohio. Host Issa Hanna, who lives in Austin, describes Taylor, Texas — one Samsung plant announcement, property values up sharply, population up roughly 30% in anticipation of the plant opening, because there simply wasn’t enough housing.
Practical read for an investor: the announcement tells you where to start underwriting, not what to pay. The lag is your entitlement and construction window. If you can’t get a shovel in the ground inside that window, you’re buying at the top of a pricing move you didn’t create.
Why Build Cost Control Decides Whether Workforce Housing Pencils
Workforce housing has a rent ceiling by definition. That kills the usual escape hatch — you cannot underwrite your way out of a cost overrun by assuming higher rents. The only variable you genuinely control is what it costs to put the building up.
Windsor’s answer is full vertical integration. Roughly 60 people covering acquisitions, underwriting, leasing, management and all construction in-house. They make their own cabinets and their own roofs. After COVID-era lumber shortages, they bought a tree farm in the Carolinas rather than keep bidding for wood supply. Other materials are sourced directly from China and Turkey. Seruya’s claim is that this makes them among the cheapest builders in the state, and they currently have four shovel-ready projects.
Most investors chasing the same corridor will never be vertical, and building a construction arm to do one deal is a worse idea than paying a premium. Realistic alternatives, in rough order of how much control they give you:
- Partner with a builder on the GP side. Give up promote in exchange for someone whose margin depends on hitting the budget, not on change orders.
- Pre-negotiate scopes before you close on land. Price the hard costs against the actual entitlement, not a per-foot rule of thumb from a different submarket.
- Build a standing GC relationship across multiple deals. Repeat volume buys you better pricing and earlier honesty about cost movement than one-off bids.
The discipline point matters more than the mechanism. In a rent-capped product, cost overrun and yield compression are the same event. If your pro forma only works at the low end of a contractor’s range, you do not have a deal — you have an option on a contractor’s good mood.
Don’t make extreme assumptions that rates are going to drop in half or rates are going to go up 50%. You really need to stress test what could happen.
— Mark Seruya, KSR and Windsor Properties
Stress-Testing Rates Before the Refinance, Not After
Underwrite the refinance rate, not the acquisition rate. Seruya is living this in real time: developments he underwrote five or six years ago assumed a rate environment that has since roughly doubled. Those properties are now coming up on refinance, and on two of the deals his team is raising additional equity or arranging creative financing to bridge the gap.
His rule is narrow and useful. Do not assume rates drop by half. Do not assume they spike 50%. Instead, stress test the range of outcomes and confirm the deal survives the bad end of it. That means running the refi at a rate you’d hate, checking the debt service coverage that results, and knowing in advance how much fresh equity that scenario demands and where it comes from.
He makes a second, less obvious argument about entry timing. Buying with rates elevated is a better assumption set than buying at 2–3% money, because underwriting off historically low rates forces you to assume those rates return. That was the flawed premise five years ago. At current levels, his view is that rates are unlikely to double again from here, and a refinance at lower rates over the next year to two years is a plausible upside rather than a required one.
The distinction is whether rate relief is in your base case or your upside case. If a deal only works when rates fall, it is a rate bet with a building attached. If it works flat and gets better if rates fall, you have a development.
Matching Deals to Investors When You Raise LP and GP Capital
Know the investor before you know the property. Seruya runs capital markets at KSR, a 150-person New York firm: he evaluates deal flow, approves or declines with his two partners, then raises both LP and GP capital on what clears. He’s blunt that understanding the investor matters more than understanding the investment, because every investor carries a different risk tolerance and a different asset-class preference — multifamily, retail, industrial, office. Showing every deal to everyone is not a strategy.
What that produces in practice is a deliberately balanced book. On the higher-risk side, KSR has been buying New York City office in a window when most capital was running from it, accumulating several million square feet at prices Seruya says haven’t been seen in 15 to 20 years. The buildings were fine; the previous owners bought at the wrong moment. That’s a low-basis, total-return position.
Against that, they hold East Coast shopping centers with their partners at the Wharton Group — roughly 7 million square feet, leased to top national retailers. That’s the income side, and it’s what makes the office risk tolerable for an investor who can’t stomach a pure total-return bet.
The application to a development raise is direct. An investor who wants quarterly distributions does not belong in a ground-up workforce housing deal with no cash flow until stabilization. An investor who wants total return and can wait does. Sorting that out before you present the deal is the difference between a fast close and six weeks of explaining why there’s no distribution in year one.
The Third Walkthrough That Killed a Deal Worth Buying
KSR was three weeks from buying a Walmart-anchored neighborhood center outside Richmond, Virginia. On the third walkthrough, a partner noticed people selling drugs in front of a small office building behind the center. The team then went in and talked to Walmart’s store management directly.
What they learned split two ways. The store ranked among the best sales per square foot in Walmart’s entire portfolio — genuinely strong performance. But management also described an extreme amount of theft, and the shrink problem ran across multiple tenants in the center, not just the anchor.
That’s anchor-departure risk that no rent roll displays. KSR retraded, cutting the bid by roughly $3 million to price in the possibility that Walmart eventually leaves. The seller countered a million dollars below their original price. KSR passed.
Seruya’s reasoning was that they didn’t want to carry that risk on behalf of investors, and the payoff came immediately — every investor who had committed to the deal thanked him. That’s the reputational trade: one passed deal buys years of credibility on the next raise.
The diligence takeaway is procedural, not philosophical:
- Walk the site more than twice, at different times of day, and include someone who wasn’t on the earlier visits.
- Talk to anchor store management in person. Sales performance and operating pain are different data sets, and only one shows up in the offering memorandum.
- Check whether a tenant problem is isolated or center-wide. Shrink across multiple tenants is a location signal, not a tenant signal.
- Inspect what’s behind and adjacent to the property, not just the parcel you’re buying.
Adjacent Supply-Demand Gaps He’s Watching Next
These are theses Seruya is exploring, not deals he has closed. He says so plainly. Treat them as demand-imbalance hypotheses worth your own diligence.
Sober living real estate, owned not operated. From conversations with Ohio state agencies, he describes a supply-demand imbalance on the order of 30 or 40 to one. His interest is specifically in owning the buildings and leasing them to operators who run the programs, on the view that sober living is hard to manage well. He flags old motels as natural conversion candidates — the unit layout is already close.
Senior housing against the retirement wave. He estimates more than 10,000 people a day are retiring, with insufficient senior housing supply to absorb it. Both assisted and independent senior housing, where strong service levels support top-of-market rent.
Land for robo-taxi fleet infrastructure. He’s seen test fleets operating in Scottsdale, Las Vegas and the New York boroughs. His read is that owning a fleet is interesting but owning the infrastructure is better — fleets need parking, charging and washing, ideally in one location. He’s applied for a fleet license in New Jersey but is primarily looking to buy land ahead of the demand.
The common structure across all three is the same one driving the Ohio housing play: identify a demand curve that is moving faster than the supply response, and take the real estate position rather than the operating position. Operating businesses carry management risk. Land and buildings leased to operators carry credit risk, which is easier to underwrite.
Frequently asked questions
How long after a data center or plant announcement does workforce housing demand actually show up?
Construction labor demand arrives first, well before permanent operations staff, and housing supply lags both. In the Columbus–Dayton corridor, Seruya describes hundreds of cranes already working against a severe existing shortage of workforce housing — meaning the demand preceded any meaningful supply response by years.
The practical implication is that your entitlement and construction timeline has to fit inside that lag. If you can’t break ground before local supply catches up, you’re competing with completed product instead of filling a gap.
What rate assumption should I use when underwriting a development I’ll refinance in three to five years?
Use a range you stress test rather than a single number, and make sure the deal survives the unfavorable end. Seruya’s rule is to avoid extreme assumptions in either direction — don’t model rates halving, don’t model a 50% spike — and instead test what happens across plausible outcomes.
He also argues that buying when rates are already elevated is a safer assumption set than underwriting off 2–3% money, because the latter required betting those historically low rates would still be available at refinance. Two of his developments underwritten five or six years ago are now needing fresh equity or creative financing because rates roughly doubled. This is not financing advice; run your own numbers with your lender.
Is it better to own sober living real estate or operate the program?
Seruya’s position is to own the real estate and lease it to operators. His reasoning is that sober living is difficult to manage well, and the operating risk is a different business from the real estate return.
He points to old motels as strong conversion candidates and cites conversations with Ohio state agencies suggesting a demand imbalance around 30 or 40 to one. He has not executed on this yet, so treat it as a thesis requiring your own market and regulatory diligence.
What should a third-round walkthrough on a retail center look for that the first two missed?
Operational reality that never reaches the offering memorandum. On KSR’s Richmond-area deal, the third walkthrough surfaced open drug dealing behind an adjacent office building, and a direct conversation with Walmart store management revealed extreme shrink across multiple tenants despite the store ranking among the best in the chain on sales per square foot.
Send someone who wasn’t on the earlier visits, walk the perimeter and adjacent parcels, vary the time of day, and talk to anchor management face to face. Strong sales and a strong location are not the same finding.
How do I decide whether to offer investors an LP position or a GP position in a development deal?
Start with the investor’s risk tolerance and asset-class preference, not the deal structure. Seruya raises both LP and GP capital at KSR and says knowing the investor matters more than knowing the investment, because tolerance for illiquidity, timing of distributions, and preferred asset class vary widely across the same investor list.
GP capital generally means more risk, more upside, and closer involvement — appropriate for someone seeking total return who can wait through a no-distribution development period. LP capital suits investors who want defined terms and lower exposure. Sorting this before you present avoids a raise that stalls on the wrong pitch.
The bottom line
If you’re underwriting anything in an announced megaproject corridor, run the refinance scenario first and the acquisition math second — that single reordering is what separates the deals that survive to stabilization from the ones that need a rescue equity round five years in.
