New Graves to Dance On
By Shlomo Chopp
Originally published in Real Estate Issues, a publication of The Counselors of Real Estate
In 1975, Sam Zell introduced us to “The Grave Dancer,” and then like Rip van Winkle, his Grave Dancer reawakened in 1984 on the pages of this very publication.
Sam taught us to dance on the graves of bad capital encumbering good assets. Today’s graves are different. They are dying business models underwritten as eternal.
It’s time, I suggest, to update the tune.
For decades, commercial real estate has comforted itself with the same bedtime story: we own hard assets. Bricks. Steel. Dirt. Things that don’t disappear when markets get volatile. And for a long time, that story worked well enough to make a lot of people rich.
But here’s the inconvenient truth: commercial real estate is only as durable as the business models that occupy it. And those business models are changing faster than our leases expire and far faster than our debt amortizes.
This isn’t a cyclical observation. It’s structural. And for those underwriting yesterday’s stability in tomorrow’s economy, it’s creating a new generation of graves, some already filled, others freshly dug.
The first mistake: confusing occupancy with reality.
Commercial real estate underwriting has become a triumph of process over thinking. We call them “fundamentals”: occupancy, rent, WALT, trailing NOI. They’re easy to measure. They fit nicely into spreadsheets. They make investment committees feel disciplined.
But they are merely snapshots, not the drivers. Occupancy tells you nothing about why a tenant is there, how profitable they are, or whether the space still makes economic sense for them. Rent rolls don’t explain whether the tenant’s business model is expanding, shrinking, or being quietly eaten alive by technology.
Yet, we continue to underwrite as if a fully leased building is a safe building, even when the tenants inside it are operating businesses that no longer require the same space, the same configuration, or sometimes any space at all. That worked when change was slow but it fails when change compounds.
This shows up in almost every loan we touch. Lenders are fixated on making each modification tie up neatly with a bow, often off appraisals that minimize lease‑up risk by capitalizing future leases while giving little weight to the fragility of in‑place cash flow. I dare not ask why, but if I did, the likely answer is some version of: “If that happens, we’ve got bigger problems than this loan.”
Technology isn’t a trend you adapt to. It’s a wrecking ball.
Technology doesn’t politely ask real estate for permission to disrupt it. It rewires how companies operate, and space demand follows — reluctantly and often violently.
Retail didn’t lose relevance because landlords got lazy. It lost relevance because e-commerce broke the store-centric distribution model. The mall wasn’t overbuilt because developers were irrational; it was overbuilt because the old model worked until it didn’t.
Office didn’t stumble because people forgot how to collaborate. It stumbled because technology made physical co-location optional for a meaningful portion of work. COVID didn’t invent that reality, it simply removed the last social friction preventing its adoption.
Industrial isn’t immune either. E-commerce logistics has been treated as a one-way escalator. Build more. Build faster. Build closer. And yet here we are, staring at rising vacancy and a reminder that demand growth without durable unit economics eventually meets supply.
And now data centers, the new golden child, are being underwritten as if megawatts are a moat, not a constraint. They treat the sector as if power density, cooling requirements, and chip evolution won’t make today’s “state of the art,” tomorrow’s awkward retrofit.
Does this sound familiar? It should. Every cycle has its version of certainty.
Credit is not permanent and it never was.
There was a time when department stores were the best credit in retail. Long leases. Name recognition. Corporate guarantees. Those boxes anchored entire portfolios. Most of those anchors are gone, weakened, or quietly renegotiating their relevance.
The mistake wasn’t that investors failed to predict which retailers would struggle. The mistake was assuming that credit quality was a permanent attribute rather than a moment in time. Today’s “great tenant” is yesterday’s business model with a fresh haircut. Survival does not equal creditworthiness. Existing does not equal durable.
If your underwriting assumes that tenants will continue to pay rising rents simply because they have in the past, you’re not underwriting real estate — you’re underwriting nostalgia.
Duration is leverage (and most people don’t see it).
Low cap rates and long-duration debt feel conservative. They are anything but. When yields compress, time becomes your enemy. The longer your capital is locked in, the more you are betting that the tenant landscape, and the technology shaping it, stays cooperative.
By the time a loan workout is needed and we’re called in, the bet has usually gone bad. The reason these cycles are even tolerated by lenders, is that decay doesn’t happen all at once. Borrowers, however are handcuffed to 10‑year paper underwritten to a world that no longer exists, and lenders are torn between enforcing documents and admitting they mispriced duration risk.
All this is fine if business models decay slowly. It’s disastrous if they don’t. The industry loves to say, “Real estate is a long-term asset.” But is cash flow? What they really mean is: we hope nothing important changes before we refinance. But technology doesn’t wait for maturity dates. It doesn’t care about your DSCR. It doesn’t respect appraisals.
And when assets require continuous reinvestment just to remain relevant: new layouts, new systems, new power, new amenities, the yield you thought you bought was never real. It was gross of reality.
Capex isn’t optional — it’s rent you haven’t paid yet.
Here’s another industry fiction: capital expenditures are treated as episodic inconveniences instead of the cost of fighting obsolescence. If a building must be reconfigured every cycle to attract tenants whose businesses keep changing, that’s not upside. That’s maintenance of relevance.
A property that produces a 5% cap rate before CapEx but a 2% yield after true required reinvestment is not a stable asset. It’s a wasting one – particularly if you don’t reinvest.
So how do you get around it? You ignore it. Property condition assessments commissioned at loan origination are often laughable in retrospect. Almost every distressed loan we work out has decaying systems and pent‑up capital requirements that everyone pretended wouldn’t be there. And that’s before technological deprecation.
The only honest question is who’s paying for the depreciation. Typically, it’s a lender – even if not the current one.
Data centers: be careful what you crown.
Data centers may well be critical infrastructure. They are also highly specialized assets with narrow re-use options and fast-moving technical requirements.
Power availability, cooling technology, chip architecture — these are not static inputs. They are evolving constraints. The risk isn’t that data demand disappears. The risk is that today’s configuration becomes tomorrow’s bottleneck.
How many real estate experts chasing this space even consider Moore’s law—the idea that the number of transistors on a chip roughly doubles every two years? While it’s true that this pace has run into limits over the past decade, those limits are technological, and the next breakthrough could do to today’s data centers what the iPhone did to personal computing. We’ll applaud this development and the stock market will boom – but how about its real estate?
We’re not pricing any of that in. We’re treating today’s rack layout, power draw, and cooling spec as if they are permanent infrastructure rather than a rapidly depreciating guess about how compute will want to live five years from now.
We’ve seen this movie before. It just had different actors.
The latest craze: AI. Just another grave in waiting?
Every cycle needs a hero, and right now, it’s AI. The industry is underwriting artificial intelligence as if it’s a structural savior — a limitless driver of demand, productivity, and data center absorption.
But for real estate, AI is not a new paradigm. It’s a very old pattern: irrational overconfidence dressed in new code. Like every technological revolution before it, the AI buildout will discover the limits of energy, economics, and scalability.
We’re already seeing it. Power constraints are real. Model training is capital-intensive. Most crucially, nobody yet knows the stable business model behind AI demand — who pays, who profits, and who is just subsidizing the experiment
So yes, AI is today’s engine of growth. And soon enough, it may likely be tomorrow’s reckoning. A momentary boost that didn’t last. Leasing office space to AI companies is not a fundamental. It’s great for the owner who wants to sell and terrible for the lender underwriting a 10‑year loan term.
Will they admit it? I very much doubt it.
But can anyone possibly imagine their unit economics, or even corporate viability in 10 years from now?
The graves aren’t hypothetical. We’re physically digging them now. Can we stay out?
The real problem: we underwrite buildings, not businesses.
Commercial real estate has become a finance game. As such, it is not a collection of independent physical objects. It is a derivative of operating companies. Yet we almost never underwrite it that way.
In borrower–lender negotiations, I rarely hear anyone talk through a tenant P&L or customer acquisition cost; I hear a lot about “rent bumps,” “credit quality,” and “rollover,” as if those exist independently of the business model paying the rent.
We don’t amortize lease value based on the depreciation of tenant business models. Most market participants don’t even understand those business models. We’re too busy deploying capital and assume the building will always have demand.
We don’t shorten duration when technology shortens relevance, partially because we can’t measure it so far in advance. We also don’t demand higher returns when reinvestment risk rises, because, frankly, we can’t coherently protect ourselves without blowing up the system.
Instead, we institutionalize yesterday’s assumptions and call it prudence.
A better way (and why it creates new graves).
Here’s the uncomfortable solution: Commercial real estate should be underwritten more like an operating business and less like a perpetual bond.
That means higher cap rates to reflect tenant volatility, shorter effective lives in valuation, faster amortization of debt, explicit pricing of reinvestment risk, and a clear-eyed view that occupancy is a lagging indicator, not a cause. Essentially, this means lower values and less proceeds, and a fear of what we don’t know.
Institutional capital would hate this framework because it forces recognition of decay, and opportunistic capital will feast on it. That tension is where opportunity would come from.
But just because it’s bad, doesn’t mean we can afford to ignore it. My day job is living in that gap between paper assumptions and the actual results of those assumptions. We represent borrowers who can’t make yesterday’s math work and lenders who are paid to defend it. When it doesn’t add up, there’s always somebody to blame or a clause to litigate.
That vantage point makes one thing painfully clear: the real risk isn’t vacancy; it’s outdated thinking about the businesses inside the walls.
New graves to dance on.
Every cycle creates its own mythology. This one insists that technology is a tailwind without consequence, that credit tenants are durable, and that hard assets protect soft assumptions. They don’t.
What technology giveth in demand, it taketh away in permanence. And commercial real estate, particularly when financed with long-duration, low-yield capital — is brutally exposed to that tradeoff.
The next generation of opportunity won’t come from guessing which asset class is “hot.” It will come from recognizing which assumptions are dead. And for those willing to acknowledge that reality early, there will be no shortage of new graves to dance on.