The first is a railroad map of the United States from 1850. I bought a reproduction from the Library of Congress, $18 plus shipping. Hand-drawn lines radiate outward from the Eastern Seaboard: thick clusters around New York, Philadelphia, and Boston, thinning as they push west. A few tentative lines reach across Pennsylvania. One thin strand crosses into Ohio. Beyond that, nothing. Blank space, the western two-thirds of the continent waiting for the lines to arrive.

The second is a data center construction map from a CBRE market report published in early 2026. Red dots for operational facilities. Blue dots for under construction. Gray dots for planned. The dots cluster in a few specific places: Northern Virginia. Central Texas. Ohio. Georgia. A handful of smaller spots in Wisconsin, Tennessee, Oregon. Some familiar names. Some places you’ve probably never thought about.

I put these two maps next to each other about six weeks ago, and I haven’t been able to stop thinking about what they share: the logic.

In 1833, Chicago Had No Reason to Exist

The place that would become Chicago was, in 1833, a flat, marshy stretch of land at the southwestern tip of Lake Michigan. Fort Dearborn, a small military outpost, sat near the mouth of the Chicago River. The river was a sluggish, shallow thing that barely qualified as navigable. The land around it flooded on a regular cycle. Mosquitoes were a significant problem.

The population was about 200 people.

Nothing about this location suggested it would become important. It had no deep-water harbor like New York. No confluence of major rivers like St. Louis, which sat at the junction of the Mississippi and Missouri and was already a thriving commercial hub. No established merchant community like Cincinnati, which by the 1830s had earned the nickname “Porkopolis” for its dominant position in meatpacking, population 30,000 and growing. No strategic mountain pass, no mineral deposits, no obvious reason for capital to flow there rather than to a hundred other places.

In 1833, the smartest investor in America would have said Cincinnati or St. Louis. Both had natural advantages: river access, established trade networks, growing populations. Both were already attracting capital.

No one would have said Chicago.

1848: the Galena and Chicago Union Railroad completed its first line running west from Chicago. 1852: the Michigan Southern and Northern Indiana Railroad connected Chicago to the East Coast rail network. 1856: Chicago became the terminus of ten trunk lines and eleven branch railroads, more than any other city in America.

The population numbers are more efficient than any explanation I can give.

1840: 4,470.

1850: 29,963.

1860: 112,172.

1870: 298,977.

1880: 503,185.

1890: 1,099,850.

From 200 to over a million in fifty-seven years. No city in recorded history had grown that fast. None of it happened because of the location itself. It happened because of where the lines went.

Subscribe now

How a Swamp Got a 500,000x Multiple (Network Effects Waaaay Before We Called Them That)

William Cronon’s Nature’s Metropolis is about how Chicago was built, and it’s one of the most important books I’ve read for understanding how money moves. Cronon argues that Chicago’s rise was topological inevitability, not geographic luck.

Railroads follow freight. Freight flows from where goods are produced (the western prairies, full of grain and livestock) to where goods are consumed (the eastern cities, full of people with money). The question for any shipper in the 1840s was: where does the grain come off one rail line and go onto another? Where does the livestock get processed before heading east as packed meat?

The shape of the network determines the answer. As builders laid more rail in the 1840s and 1850s, running east-west from the prairies to the coast and north-south along the Mississippi corridor, the lines converged at specific points. The geography of the Midwest, flat and mostly unobstructed with the Great Lakes creating a natural northern barrier, funneled them toward a small number of junction points. Chicago was one. It turned out to be the most important one.

And being a junction compounds.

Two rail lines meet at your city. Freight has to be unloaded from one line and loaded onto another. That transfer creates demand for warehouses. Warehouses create demand for workers. Workers need housing, food, services. Services attract businesses. Businesses attract more rail lines, because rail lines follow freight, and freight goes wherever businesses are. More rail lines make the junction more central. More centrality makes the junction more valuable.

Once this loop starts, stopping it is close to impossible. Chicago became dominant because it was the first to reach criticality as a junction. After that, the loop locked in advantage faster than any competitor could close the gap.

The Chicago Board of Trade was founded in 1848, the same year the first railroad arrived. Think about that timing. The Board’s founders didn’t create it because Chicago had a long tradition of commodity trading. Chicago didn’t have a long tradition of anything. They created it because so much grain was flowing through the city that the old merchant-to-merchant system couldn’t handle volume. They needed standardized contracts. They needed a central marketplace. They needed what we now call commodity futures.

Grain futures, one of the foundational instruments of modern commodities trading, were invented in Chicago. The inventors were no smarter about finance than people in New York or London. Chicago was where the grain was. The grain was where the junction was. The junction was where the lines converged.

Financial innovation followed infrastructure topology.

Cincinnati: A Case Study in Holding the Wrong Asset

For every Chicago, there’s a Cincinnati.

Cincinnati’s story is the one that should make you uncomfortable, if you own commercial real estate in a major knowledge-economy city.

In the 1840s, Cincinnati was the largest and most prosperous city in the American interior. Population 46,000 in 1840, growing fast. The meatpacking industry was enormous. Farmers drove hogs overland from Kentucky and Ohio, slaughterhouses processed them, packers salted the meat, and shippers sent it down the Ohio River to New Orleans and from there to the world. Cincinnati’s nickname, Porkopolis, was a market position.

Cincinnati’s advantage was geographic: it sat on the Ohio River, which connected to the Mississippi, which connected to the Gulf of Mexico and the global shipping lanes. In the river-commerce era, this was the equivalent of sitting on a fiber optic trunk line.

Then the railroads arrived. The railroads didn’t need the river.

Rail moved freight faster, in any direction, regardless of water routes. The relevant question shifted from “are you on the river?” to “are you at the junction?” Cincinnati wasn’t a junction. It was a river city, and the river was becoming irrelevant.

The meatpacking industry illustrates the shift with painful clarity. Chicago’s Union Stock Yards opened in 1865. Rail lines brought livestock to Chicago from a vast catchment area across the Great Plains, far larger than the overland driving range that fed Cincinnati. Within a generation, Chicago had absorbed the bulk of the meatpacking trade. Cincinnati’s facilities were fine. The infrastructure topology had changed, and in the new topology, Cincinnati’s location was a siding.

Cincinnati didn’t die. In 1870, it still ranked 8th among American cities by population, with 216,000 people. It’s a functioning city today. But the premium, the extra economic value that its location had captured, evaporated. Cincinnati went from the interior’s dominant commercial hub to a mid-tier regional city. The businesses that had located there because of its river advantages relocated to wherever the new infrastructure logic favored.

St. Louis tells a similar story, with a darker twist. St. Louis was the gateway to the West via the Mississippi, and its civic leaders knew the railroads were coming. They had time to respond. They made a choice that, in retrospect, was catastrophic: they invested in river infrastructure. They doubled down on the system that had made them rich. They could not imagine that railroads would make the river irrelevant for overland freight.

By the time St. Louis pivoted to railroads, Chicago had already captured the junction position. St. Louis in 1870 still ranked 4th in the country, population 310,000. But the trend was set. The city with every natural advantage, with the established relationships and existing investment, lost to a swamp. The swamp was where the new lines crossed.

St. Louis’s mistake was path dependency. The infrastructure they’d built, the commercial relationships they maintained, the capital they’d invested, all of it said: the river is the thing. The river was the thing. Until the railroads made it a secondary channel.

The Underwriting Model That Keeps Being Right

Infrastructure creates nodes. The places where new infrastructure converges, junctions, hubs, transfer points, capture disproportionate value. The reason is position in the network, not inherent quality.

Investors, workers, and commercial activity flow toward the junction points. This creates a feedback loop: more capital makes the node more valuable, which attracts more capital. Chicago’s grain futures market, its meatpacking industry, its manufacturing base, all of these grew downstream of the junction.

Old nodes lose their premium. The places that were valuable under the old infrastructure don’t disappear. But the scarcity premium they captured, the extra value from being the place to do business, erodes. Cincinnati in 1890 was still a functioning city. It was no longer Porkopolis.

This pattern has repeated for every major infrastructure transition in American history. Canals made New York dominant (the Erie Canal, 1825). Railroads made Chicago dominant (1850s-1860s). The highway system made suburbs possible (1950s-1960s) and redistributed retail and residential value away from urban cores. Fiber optics made Northern Virginia the internet’s physical hub (1990s-2000s), because that’s where undersea cables landed and early government networks interconnected.

The infrastructure creates the topology. The topology creates the nodes. The nodes capture the value. And the people who owned assets in the old nodes, confident the premium would last, discovered that premiums expire.

1.21 Gigawatts in Abilene, Texas

I keep staring at this data center map from 2026. The one with the colored dots.

Virginia leads the country with over 600 operational data centers and another 136 under construction. Most of these cluster in Northern Virginia’s “Data Center Alley,” centered on Ashburn and Loudoun County. This region handles something like 70% of the world’s internet traffic. It became the most important digital infrastructure hub on earth for the same reason Chicago became a railroad junction: early fiber optic networks converged there, and once the junction formed, the feedback loop took over.

Texas has 413 data centers operating today, with 140 more under construction. Texas is projected to surpass Virginia as the largest data center market by 2030. Cheap electricity, abundant land, fast permitting, no state income tax, and a power grid (ERCOT) that, whatever its other problems, allows rapid connection of new load. The $100 billion Stargate campus in Abilene is scaling past 1 gigawatt of power capacity across eight buildings. Abilene. Population 125,000. A city most Americans couldn’t place on a map.

Then there’s the layer that caught my attention: 64% of all data center capacity under construction sits in what the industry calls “frontier markets.” Places like Ohio, Wisconsin, Tennessee, Georgia, and Indiana. Places that, in the data center world five years ago, were nowhere. Small towns with cheap power and available land are becoming the next generation of infrastructure nodes.

I see Chicago.

A swamp no one cared about, becoming the most important place on the map because the lines converged there. The physics of the infrastructure, where the power is, where the water is, where the land is, funneled the investment toward specific points.

Abilene, Texas. Prineville, Oregon. New Albany, Ohio. These are the 1840s Chicagos. They have no inherent importance. They have no tradition of being technology hubs. They have cheap electricity, water, available land, and permissive zoning. In the data center topology, these are the junction points. Capital is flowing toward them at a velocity that the local populations are beginning to process.

Subscribe now

Your Trophy Asset Might Be a River City

The data doesn’t say what I expected it to say, and the honest version is more interesting than the simple one.

The simple version would be: “AI enables remote work, so expensive office cities lose value.” That story is clean, and it’s incomplete in a way that matters.

San Francisco’s office vacancy rate hit 37% in early 2024. A ghost town. The pandemic had emptied the offices, and tech companies weren’t coming back. The obituaries were written.

Then AI happened. OpenAI leased a million square feet across Mission Bay. Anthropic secured over 600,000 square feet on Howard Street. AI companies accounted for three-quarters of the city’s net absorption. By Q2 2026, San Francisco’s vacancy rate had dropped to about 30%, the fastest improvement of any major city in the country. Mission Bay ran out of space.

AI didn’t kill San Francisco. It resurrected it. At least, it resurrected the parts of San Francisco where AI companies want to be.

But it resurrected them selectively. The submarkets near AI company headquarters (Mission Bay, SoMa, the Financial District) tightened. Others barely moved. Some are still over 30% vacant. A specific piece of San Francisco recovered, because that piece sits at a node in the AI topology, the place where the talent, the companies, and the network effects compound.

The rest of San Francisco is still on the wrong side of the map.

Manhattan’s version of the same story: vacancy dropped to 13.1% by Q1 2026, well below the national average of 17.8%. AI companies are leasing aggressively. Trophy buildings in Midtown and the Plaza District are full, commanding rents approaching $200 per square foot. Class B buildings in less desirable locations? Vacancy above 20%. The market isn’t recovering. It’s splitting. Trophy assets at network nodes are tightening. The rest is drifting.

The railroad parallel translated into commercial real estate: Chicago didn’t make every city more valuable. It made the junction more valuable, and made the non-junctions less so. AI is redrawing the topology, creating new nodes where AI talent and compute converge, draining value from locations that don’t sit on the new lines.

The Cap Rate Nobody Is Modeling (and They Should)

The data center map and the office recovery data are the surface. There’s a deeper layer that I think most real estate investors aren’t seeing.

Every previous infrastructure transition rewired the social geography alongside the economic geography. Railroads moved people, not just freight. The cities that became junctions didn’t just get richer; they became new kinds of places. Chicago in 1890 was not a bigger version of the trading post at Fort Dearborn. It was a dense, industrialized, ethnically diverse, politically volatile, culturally generative urban organism. The junction attracted everything: labor, culture, conflict, innovation. The junction became the center of gravity for a new form of civilization.

I think we’re in the early stages of a similar reorganization, and the data center map is the first indicator.

AI compute requires specific physical conditions: power, cooling, connectivity. Those conditions exist in specific places. Capital flows to those places. Capital brings jobs, not factory jobs, but data center construction, operations, maintenance, security, and the local services that support them. Jobs bring people. People bring demand for housing, retail, education.

Meanwhile, the output of those data centers, the AI itself, enables a different kind of geographic redistribution. If AI coordination tools let a team of five do the work of fifty, and if those five don’t need to be in the same room, then the economic logic that justified the extreme density of Manhattan, San Francisco, and London weakens. The shift won’t happen overnight. But the direction is set.

The premium that these cities command, $78 per square foot for Class A office space in Manhattan, $200+ for trophy buildings, rests on a specific assumption: you need to concentrate knowledge workers in physical proximity to generate outsized value. That assumption drove commercial real estate pricing for the knowledge-economy era. It held true when the critical resource was human-to-human interaction: brainstorming, mentoring, deal-making, the ambient information transfer that happens when smart people share physical space.

If AI handles an increasing share of the information processing, coordination, and analysis that justified the density premium, the assumption weakens. I don’t know how fast. The data is too early, the transition too incomplete. But the railroad precedent suggests that geographic premiums, once they start eroding, erode faster than anyone expects. Cincinnati’s civic leaders in 1850 were confident the river advantage was permanent. They had data to support that confidence. The data described the past accurately and the future incorrectly.

Three Trades From a $18 Map

If you had looked at the railroad map of 1850 and understood the junction logic, three bets would have outperformed most other investments available in America over the following fifty years.

Buy land at the junctions before capital arrives. Chicago land that sold for close to nothing in the 1830s was worth fortunes by the 1870s. The early investors in Chicago real estate, the ones who understood that the junction effect would compound, earned returns that dwarfed anything available in Eastern financial markets. The same logic applied to smaller junctions: Indianapolis, Kansas City, Omaha. Each place where multiple rail lines converged experienced a version of the Chicago story at smaller scale.

Build the commercial infrastructure the junction needs. The Chicago Board of Trade. The Union Stock Yards. The grain elevators, warehouses, hotels. These bets weren’t speculative wagers on an uncertain technology. They were bets on certain demand, because once a junction exists, the commercial activity it generates is locked in. The junction needs processing capacity. Processing capacity needs physical infrastructure. The infrastructure earns returns as long as the junction remains active.

Reduce exposure to old-node premiums. Cincinnati riverfront property. St. Louis warehouse districts built for river commerce. These assets didn’t go to zero. But their premiums eroded as the new infrastructure topology redirected commercial activity to new nodes. The investors who held on, confident that the river would always matter, were right about the river (it still exists) and wrong about the premium (it evaporated).

Apply these bets to the 2026 map.

The new junctions are forming where cheap power, water, fiber, and permissive regulation converge. Some of these are existing tech hubs that happen to sit on the right topology (Northern Virginia). Some are places no one in commercial real estate has on their radar (Abilene, New Albany, rural Wisconsin). The early investors in real estate and infrastructure at these new nodes, the people buying land and building power capacity before the next wave of data center construction arrives, are making the 1840s Chicago bet.

The commercial infrastructure the new junctions need, power generation, cooling systems, fiber networks, worker housing, logistics facilities, is the 2026 equivalent of grain elevators and stockyards. These are infrastructure bets with demand locked in by lease commitments and power purchase agreements.

And the premiums of old-node assets, the $200-per-square-foot trophy offices in cities whose value proposition was “this is where the smart people sit close together,” these premiums may or may not erode. San Francisco’s AI-driven recovery suggests that some knowledge-economy cities will find a new role in the AI topology (the way New York, which was a canal city, managed to also become a railroad hub and then a financial center). Others may not.

The question I’d be asking if I had significant commercial real estate exposure in a traditional knowledge-economy city: is my asset on the new map, or on the old one? Is the premium I’m paying for this location based on a topology that’s strengthening, or one that’s being redrawn?

St. Louis’s leaders asked themselves a version of this question in the 1850s. They answered it with confidence. They answered it incorrectly. They were smart people who had made a fortune on the old map and could not see the new one.

I keep going back to the two maps on my wall.

The 1850 map is gorgeous. Hand-drawn, precise, radiating confidence about a system still being built. Each line represents someone’s capital at risk. Someone looked at the terrain, looked at the freight routes, looked at the population centers, and decided: the line goes here. Some of those decisions were brilliant. Some were catastrophic. The lines that converged on Chicago minted fortunes. The lines that dead-ended at river towns consumed capital and returned nothing.

The 2026 map is ugly. A corporate slide deck, colored dots, no elegance. But it encodes the same type of information. Each dot is someone’s capital at risk. Each cluster of dots is a bet about where the new infrastructure will converge. The clusters in Virginia and Texas are the equivalent of the thick lines radiating from New York and Philadelphia in 1850: established, obvious, already attracting the bulk of the capital. The scattered dots in Ohio, Wisconsin, Georgia are the early, tentative lines pushing west into unknown territory. Some of them will converge into the next Chicago. Some will dead-end.

The 1850 map, read correctly, would have told you: buy Chicago, sell Cincinnati. The reason had nothing to do with either city’s character or ambition. It had to do with where each sat in the network.

I think the 2026 map is saying the same thing. The network is different, compute infrastructure instead of railroads. The logic is the same. Capital follows infrastructure. Infrastructure redraws geography. Geography reprices the assets sitting on it.

This has been true for two hundred years. I see no reason it stops now.

The question is whether you’re reading the map. Most of the investors I talk to are reading the earnings reports, the model benchmarks, the API pricing comparisons. They’re asking which AI company will win. They’re not asking which geography will win.

The railroad investors who asked “which railroad company will win?” lost money, on balance. The ones who asked “which city becomes the junction?” made it.

I think the same split is coming for AI. And the answer, like it was in 1850, is sitting right there on a map.

You just have to tape it to your wall and look at it long enough.

Not a hundred thousand. A hundred. One bill, Ben Franklin’s face on it, traded for 1.1 million square feet of American retail outside Pittsburgh, a building that once held a food court and a carousel and a mall Santa. The lender was owed 142 million on it. They handed over the keys, took the hundred bucks, and felt good about the deal.

When I first read that, I assumed it was a typo, or some tax trick, or a story that fell apart on the second click. It did not. And the more I pulled on it, the more I realized this one absurd transaction is a tiny, sped-up version of something enormous and slow that is happening right now to office towers all over the country. Buildings that sold for hundreds of millions are changing hands for a third of that, a fifth, sometimes less than the parking garage underneath them.

You already know the story everyone tells about this. Greedy investors paid stupid prices, ignored the obvious fact that the internet would let everyone work from home, and got exactly what they deserved.

I am going to try to talk you out of that story, because it is comforting and it is wrong, and what is actually underneath it is a lot stranger and a lot more interesting than greed.

Here is the part that should bother you. The people who bought these buildings were not dumb. Most of them were pension funds and insurance companies, which is to say the most cautious money on the planet, the money whose entire job is to never, ever blow up. They ran the models. They knew about working from home. And they walked into one of the biggest asset wipeouts in modern history anyway, calmly, doing the responsible thing.

So forget greed. The real mystery is sharper and weirder: how did the most careful investors alive lose this much money, this predictably, while doing nothing a sensible person in their chair would have done differently?

The answer is one equation. It is almost insultingly simple. It is also the most important thing in this essay, and we are going to spend the whole piece staring at it, because every time you look at it a little harder it tells you something a little bigger, until at some point you look up and realize it stopped being about offices a while ago.

Let me introduce the person who is going to lose all the money.

Meet Carol

Let’s give the money a face. Call her Carol.

Carol runs the pension fund for a few hundred thousand schoolteachers. This means Carol has roughly forty billion dollars and exactly one rule: do not, under any circumstances, blow up the schoolteachers. Carol does not chase moonshots. Carol does not buy magic beans. Carol buys boring things that pay steady money for thirty years, because in thirty years a kindergarten teacher in Ohio is going to want her pension to exist.

In 2015, Carol buys a gleaming office tower full of bankers in a major city. Solid building, blue-chip tenants, ten-year leases, the works.

Every sober person in finance nods. This is the most Carol thing Carol could possibly do. Office towers are the oatmeal of institutional investing. Nobody gets fired for buying the oatmeal.

Hold onto Carol. We are going to follow her all the way down, and at the bottom we are going to find out that her safest, most oatmeal decision was secretly the riskiest bet in the building. Not because she was careless. Because of what she was actually buying, which was not what it looked like.

What is a building, actually

Quick question, and it sounds dumb on purpose. When a company pays rent, what is it buying?

Not bricks. Not the elevators. Not the marble lobby that smells like a fancy hotel. A company pays rent to have somewhere to put its people. Rent is the fee for desks that have humans sitting at them. A law firm with 4,000 lawyers fills a tower. The same firm with 400 lawyers keeps two floors and waves goodbye to the rest.

Which means Carol did not buy a building. Carol bought a bet on a number. The number is how many white-collar workers need a place to sit. The granite and the elevators are a costume. Underneath, she is holding a financial claim on the size of the knowledge workforce.

This is the first turn of the screw, and it already moves the furniture around. Carol believed she was buying a piece of the growing economy. The equation says she bought a quantity of occupied desks. For seventy years nobody noticed the difference, because for seventy years there was no difference. We are about to find the difference. It is the whole game.

So let’s write the dumb little equation down. How much office space does the world need?

office demand = (number of office workers) × (space each one uses)

Two numbers, multiplied. Carol’s entire forty-billion-dollar life is downstream of these two numbers. Watch what happens when we poke the first one.

The number that everyone takes for granted

“Number of office workers” sits there pretending to be a simple fact, like the population of Canada. It is not a simple fact. A company’s headcount is something the company calculates, out of two deeper things: how much work it has to get done, and how much one worker can get through. You hire enough people to cover the work, then you stop.

So let’s crack open that first number and see what is inside. The equation grows a layer:

office demand = ( work to be done ÷ how much each worker gets done ) × space per worker

Now there is a fraction in the middle, and that fraction is the headcount. This fraction is the beating heart of the whole thing. Everything interesting that is about to happen, happens inside it.

Here is how the three pieces behaved for the seventy years that trained Carol and everyone like her. The bottom of the fraction, how much each worker gets done, went up every decade, because tools kept getting better. You would expect that to mean companies need fewer workers, the way a bigger bottom shrinks any fraction. The opposite happened. The top of the fraction, work to be done, went up even faster, because every time a kind of knowledge work got cheaper, companies discovered they wanted a mountain more of it. So the fraction grew. More workers, decade after decade. Meanwhile the third piece, space per worker, slowly shrank, through cubicles and open plans and the tiny daily humiliation of hot-desking.

The rising headcount beat the shrinking space, and office demand drifted gently up, year after year, dull and dependable, a savings bond cast in concrete.

And now the equation shows you something that plain words tend to hide. Carol was never really betting on offices. She was not even betting on jobs. She was betting on a relationship between two numbers: that the top of the fraction would always outrun the bottom. That bet had a two-hundred-year winning streak.

Which raises a question the winning streak let everyone skip for two centuries. Why did the top always win?

This is the part I could not stop thinking about, so stay with me, because the answer is the soul of the whole thing.

Humans and machines had a deal

For about two hundred years, humans and machines had a Deal.

Nobody wrote the Deal down. Everybody honored it anyway. The Deal went like this: a machine could take over a task, but a human still had to stand one step above it, aiming the machine, judging what it spat out, deciding what to do next, owning the result. The spreadsheet would do the arithmetic, sure, but a person had to know which numbers mattered and what the answer meant.

So whenever a machine ate a task, it did not push the human out the door. It pushed the human up, onto the next step, the one the machine could not reach. Cheap calculation freed people into analysis. Cheap analysis freed people into judgment. There was always higher ground, just past where the machine could climb, and the freed-up worker climbed onto it. Multiply that across the entire economy, and that upward climb is the reason the top of the fraction kept exploding. Every machine created a new floor of human work standing on top of it.

The Deal, in one line: machines take the low ground, humans climb to higher ground, and there is always higher ground.

And the Deal held on exactly one condition, which almost nobody said out loud because it never needed saying. The machines cannot climb. As long as there was always one more cognitive rung above the tool, a place where only a person could stand, the people stayed employed, the fraction kept growing, and Carol’s tower kept filling up with bankers.

Read that condition one more time, because Carol’s forty billion dollars is sitting on top of it. The whole bet was safe for as long as, and not one day longer than, machines could not climb the ladder by themselves.

You can probably see where this is going.

AI is the first machine we have ever built specifically to climb the ladder.

So about that one condition

The good thing about having a real equation is that you do not have to argue about AI with your hands. You can just take each piece of the equation and ask, calmly, what AI does to it. So let’s do that, one piece at a time.

Space per worker. Already hit. This is the crash you have already lived through. Working from home in 2020 did not touch the fraction at all, because the bankers and lawyers kept their jobs, so headcount held steady. What it hit was this last piece. People came in three days a week, then two, companies quietly let leases lapse, and space per worker fell off a cliff. Then 2022 interest rates blew up everything bought with debt, which was most of it. That is the famous office crash. It hit one piece, it is mostly baked into prices now, and the bleeding has slowed.

Quick aside, because it matters for whether Carol was a fool. By late 2021, the market had already priced in the work-from-home hit. What actually destroyed people was the interest-rate shock stacked on top, which nobody in 2015 could have put in a spreadsheet. Carol saw the obvious blow coming. The thing that got her was the ground moving underneath it.

The bottom of the fraction, how much each worker gets done. AI shoves it straight up, harder than any tool before it, and for the first time it shoves at the judgment and the next step, the exact higher ground that used to be the human’s safe perch above the machine.

The top of the fraction, work to be done. This is the only piece still up for grabs, and Carol’s entire future hangs on it. When AI lifts the bottom, does the top race ahead to keep up, the way it has for two hundred years, so the fraction stays full? Or does the work stay roughly where it is, now done by way fewer people, so the fraction shrinks and the desks empty?

That one question is the second shoe, and right now the market is pricing it at zero.

Now, you might be thinking: hang on, the AI jobs apocalypse keeps getting promised and the unemployment rate is fine. You are right, and I want to be straight about it, because the loud version of this story is everywhere and most of it is junk. There is no white-collar job collapse in the big numbers yet. What there is, is an edge starting to fray, and it is the exact edge the equation says should go first. The bottom rung. Workers aged 22 to 25 in the most AI-exposed jobs have already seen their employment drop about 16 percent relative to everyone else. Young software developers sit nearly a fifth below their 2024 peak. AI is not coming for the senior people first. It is quietly removing the entry-level rung, the one that used to turn graduates into the very headcount the fraction is built from.

The same equation, two completely different answers

Run the equation across the whole stock of buildings and it does a strange thing. It does not give you one answer. It gives you two, and they point in opposite directions.

Picture the workers who survive the thinning. Fewer of them, more senior, more valuable, and their companies want them in the same room, in a place that does their best work justice and helps recruit the next ones. Those companies fight over the best towers, the ones with light and a good location and a cafe that does not depress you. For trophy buildings, the fraction holds, and space per worker might even rise as a few precious humans spread out. Some hit record rents in the middle of a so-called office apocalypse.

Down the street, the older building, the one whose whole reason to exist was to hold a lot of ordinary workers at a reasonable rent, watches its fraction fall through the floor. It loses tenants, then financing, then any future at all. It does not get cheap and find a second life as a bargain. It empties out, gets converted into apartments if its owner has the cash, and gets demolished if not.

One cause, two opposite outcomes. People draw the shape as a K. And the K is the giveaway that this is structural and not just a bad year, because a normal cycle drags the whole market down together and then lifts it back together. Only a change in the actual terms of the equation rips it cleanly in half like this.

Wait. This was never about offices.

Up to now we have been reading the equation as a fact about office buildings. It is not a fact about office buildings. It is a fact about anything whose income depends on how many knowledge workers there are.

Swap the left side of the equation. Take out “office demand” and put in the revenue of the sandwich shop jammed into the base of Carol’s tower. Or the contract to clean forty floors every night. Or the commuter rail line whose entire reason to exist is the towers filling up each morning. Or the software company that charges by the seat and quietly assumes the seats keep multiplying. Or, one level up from all of them, the city itself, which wrote its whole budget around a downtown packed with high-earning people who pay taxes and buy lunch.

The right side of the equation barely changes. Every one of these is the same fraction in a different outfit: money riding on a rising count of human knowledge workers.

Carol’s tower is just the loudest member of this family. It is the most expensive, the most heavily borrowed against, and the only one with a price stapled to it that updates in public every quarter. So when the thing underneath the whole family starts to shift, the office is the one that screams first.

What the equation was measuring the entire time

Now strip the equation down to the bone and ask what it has actually been tracking, all these years.

Office value goes up and down with the amount of paid human thinking that needs a place to sit. That is what the whole asset class is. It is a meter. It measures one thing: how much the economy pays for human cognition.

A falling office value is that meter twitching. It is the price system noticing that the premium it has always paid humans for thinking, the premium that filled the towers and the sandwich shops and the trains, is starting to leak out.

And I promise you this is not a metaphor, because you can watch the money physically move. The same pools of capital that spent decades building towers to house human minds are now building windowless sheds in the desert to house GPUs. Office construction has fallen to its lowest in years while data center construction goes vertical. A dollar that used to become a desk is becoming a server rack. The returns to thinking are relocating, in poured concrete, out of the buildings we made for human cognition and into the buildings we are making for machine cognition. The office crash is the shadow that move throws on the wall.

We can even put a rough number on the one piece still in play, and this is the chart that holds the whole argument in one picture. It turns on a single dial: of all the time AI saves a company, how much do they pour back into doing more work, which keeps the top of the fraction climbing and the desks full, versus how much do they just keep as a smaller payroll?

Today’s office prices sit far to the right of that dial. They are a bet that companies will pour almost all of the AI savings back into more work, that the Deal holds, that the next rung is still human. Then you go look at how AI is actually getting used inside real companies, where enterprise deployment runs about 75 percent full automation, the hand-it-the-task-and-take-the-output kind, not the friendly sit-beside-you-and-help kind. The dial the real usage points to is nowhere near where the prices are. That gap, very roughly 19 percent of value across office and closer to 25 percent for the older buildings, is the second shoe, still unpriced, sitting on top of the crash that already happened.

One honest catch, because the sharp reader is already raising a hand. You cannot read that dial to a decimal point. Going from how people chat with an AI to the labor content of the entire economy is a real leap, and anyone who hands you a precise number is bluffing. But you do not need a precise number. You only need to notice that the market’s number is sitting outside the range the evidence allows. The bet baked into office prices today is not “AI will be mild.” It is “AI will behave like every machine before it, and keep the Deal.” That is a specific claim, it can be proven wrong, and it gets lonelier every quarter.

Carol’s real trade

Take the elevator back up to the street and stand next to Carol in 2015 with her gleaming tower, and let’s finally say what she was holding.

A long bet on office was a long bet on the size of the human knowledge workforce. Which was a long bet on the Deal, the two-hundred-year truce where machines complement minds. Which was, although it appeared on no page of the contract, a short bet against the arrival of a machine that thinks.

For seventy years, “long the economy” and “short AI” were the same trade, the exact same trade, because all the growth ran straight through human headcount. More output meant more workers meant more desks. The equation tied them together with a knot. AI is the moment the knot comes undone. The economy can boom while the office dies. Output can soar while the desks empty. And Carol’s tower, the biggest and most leveraged and most public bet on the old knot, is the first big asset to end up on the wrong side of the split and reprice in broad daylight, in front of everyone, every quarter.

That is what makes it the canary. Not the most poisoned thing in the mine. The first one with a voice.

The thing that gets me about this story is that nobody in it was stupid. Carol did the careful thing, the spreadsheet thing, the thing two hundred years of history paid out on every single time. The careful thing and the catastrophic thing turned out to be the same thing, and the only way to have dodged it was to bet against the most reliable pattern in the modern economy before the pattern broke. Try getting that one past your investment committee in 2015.

And the rest of the family is still standing on the same cracked piece of ground. Most of it does not trade on an exchange, so it will not reprice in a clean number on a screen. It will reprice slowly and quietly, as a town that thins out, as a tax base that sags, as a career that loses its bottom rung, as the soft two-hundred-year assumption that making more of anything will always, somehow, take more of us.

Carol was not wrong about the building. The building is fine.

She was wrong about the Deal.

And so, still, is almost everyone else.

*On the numbers: the dial in the last chart is a simulated model, not a crystal ball. It calibrates the three-part equation to the historical record, office-using employment from the BLS and space per worker from the big brokerages, then turns the single dial and reads off the result. The point is the shape of the relationship, and where today’s prices are standing on it, not a precise forecast.

The video went viral. Griffin threatened to move Citadel’s investments to Miami. The familiar gears of the American culture war engaged.

I’ve been thinking about this scene for weeks. The politics aren’t the reason. The politics are familiar: rich guy versus populist mayor, tax-the-wealthy versus capital-flight threats, left versus right. You’ve seen this movie.

I keep coming back to it because I think the frameworks people are using to interpret the moment are too small.

The obvious frame: socialism versus capitalism. A DSA mayor takes on a hedge fund billionaire. Class warfare, American edition.

A step up in sophistication: populism versus the establishment. A young outsider channels resentment against an out-of-touch elite. The pendulum swings.

Both frames fit. Both miss the thing underneath.

The thing underneath is an engine. It has been running for about five hundred years. It powered the rise of the Dutch Republic, the British Empire, the American century. It won the Cold War. It built the city that Mamdani and Griffin are both standing in.

That engine is undergoing a mutation. AI is reaching into its internal machinery and altering the conditions that made it work. Neither Mamdani nor Griffin has the vocabulary to describe what’s happening, and I’m not sure anyone else does either. But I’ve spent enough time looking at the pieces that I want to try to lay them out.

To see the mutation, you first need to see the engine. And to see the engine, you need to start in Amsterdam, in the year 1602.

The Original Term Sheet

The Dutch East India Company did something in 1602 that no one had done before. It sold shares to the public on the Amsterdam Stock Exchange. Ordinary citizens could walk in, put down guilders, and own a fractional piece of a trading venture to the East Indies.

This gets taught in business schools as a financial innovation. A clever capital-raising mechanism. That framing undersells it by three orders of magnitude.

Think about what the Dutch East India Company was asking people to do. Hand your money to strangers. The strangers will use it to outfit wooden ships. The ships will sail to the other side of the planet, through storms, past pirates, across oceans where diseases no European doctor could treat ran through crews like fire. The round trip takes two years. The ship might not come back. Your money might end up on the ocean floor somewhere off the Cape of Good Hope.

Who takes that deal?

You take that deal if you believe a specific promise. The promise, encoded in a share certificate, says: whatever profits this venture generates, your proportional cut belongs to you. Not the king’s. Not the company director’s to skim. Yours. Courts will enforce this. You can sell the share to someone else if your nerve fails. You can leave it to your children. The returns are yours.

This was, if you want to think about it in modern terms, the original term sheet. And it set something in motion that had been stuck for thousands of years of human commerce. Once people believed that the future returns belonged to them, they put capital at risk. And once capital was at risk in sufficient volume, it started flowing toward the most productive available ventures, pulled by each investor’s pursuit of their own upside. No king pointed the way. No committee allocated the funds. The credibility of the property right, the trust in that single promise, was the entire activation mechanism.

The reverse makes the point cleaner.

In the early 1930s, Joseph Stalin’s government forced millions of Ukrainian and Russian farmers into collective farms. The arrangement: your land, your animals, your tools now belong to “the collective.” You’ll work the collective’s fields. The output goes to the state plan

The farmers processed the economics of this arrangement faster than the ideologues who designed it. If the extra wheat you grow by working twice as hard goes to the collective, the incentive to work twice as hard vanishes. And if the cow in your barn will become collective property next week, the rational move is to slaughter it tonight, salt the beef, and feed your family before someone in Moscow decides where the meat goes.

They did this across the Soviet countryside. Livestock herds dropped by close to half within two years. Millions of animals, killed by their owners, in a wave of collective liquidation that no central planner had anticipated.

The farmers weren’t saboteurs. They were doing the math that every investor, every entrepreneur, every person with a stake in the future does. When the future doesn’t belong to you, you eat the future today. You pull the money off the table. You slaughter the cow.

Property rights are the first ingredient in the engine. They’re the reason anyone bothers to invest, build, take risk, or think about tomorrow rather than today. Remove them, and the downstream machinery goes still.

The Poker Table

The second ingredient is harder to see. It works in the background. But it carries as much weight as the first.

Picture a poker table the size of the world. Millions of seats. Each player holds different cards: different information, different judgment, different appetite for risk. They bet against each other, raise, fold, call. No player sees all the cards in play. But because millions of them are betting, each one acting on a different sliver of reality, the prices that emerge from all these overlapping wagers turn out to be more accurate, in aggregate, than the best estimate any single player could produce.

This is what the Austrian economist Friedrich Hayek figured out in 1945. He published it as an academic paper. The plain-language version: a crowd of ordinary people, each acting on local knowledge, will price goods and risks with more precision than a panel of geniuses working from a central office.

Think about what a single price contains. The price of a dozen eggs in a Manhattan grocery store encodes information about the number of laying hens within trucking range, the cost of feed corn in Iowa, diesel prices for the delivery fleet, warehouse rent in New Jersey, wages for the drivers, seasonal demand shifts around Easter, competition from the bodega down the block, and about four hundred other variables that no individual person tracks. All of that information, compressed into one number on a shelf tag. No committee compiled it. No bureau calculated it. Millions of independent decisions by farmers, truckers, wholesalers, grocers, and shoppers, each person processing one tiny piece of the picture, fed into the system. The price is what came out the other end.

This is where a Soviet mathematician named Viktor Glushkov enters the story, and I find his story difficult to put down once I’ve picked it up.

In 1962, Glushkov pitched the Politburo a plan. Build a computer network covering the entire Soviet Union. Wire every factory, every warehouse, every rail depot, every collective farm into a single data grid. Run mathematical models on the data to compute optimal resource allocation in real time, across the whole economy. No prices needed. Pure computation.

He called the system OGAS: the All-State Automated Management System. Estimated cost: about 20 billion rubles, in the same range as the Soviet space program. Projected timeline: ten to fifteen years.

Glushkov’s reasoning was clean. Hayek says central planners can’t match the information-aggregation power of prices? Fine. Hayek was writing in 1945, before computers. Build a computer big enough, feed it enough data, and you can do what the price system does, without the price system. Replace the poker table with an equation.

The Politburo killed the project. The reasons were a cocktail of the bureaucratic and the structural. Planning ministry officials didn’t want a computer deciding their budgets; a computer that optimizes resource allocation is a computer that makes middle managers redundant, and middle managers, then as now, tend to notice when someone’s building a machine to replace them. The military didn’t want to share resources. And the most basic problem: the computers of 1962, room-sized arrays of vacuum tubes and early transistors, couldn’t come close to processing real-time data from a 200-million-person economy.

Glushkov was right about the concept. He was sixty years too early for the hardware.

The hardware exists now. We call the systems running on it AI.

But they haven’t rescued central planning. The Soviet Union collapsed decades before ChatGPT launched. There is no centralized economy left to optimize.

Instead, these systems are doing something Glushkov never considered. I’ll come back to this.

The Math That Won the Cold War

Property rights and market pricing are the engine’s two structural components. Time is the force that reveals what the engine can do when you let it run.

In 1626, Peter Minuit paid about $24 in equivalent Dutch guilders to the Lenape people for the island of Manhattan. If the Lenape had invested that $24 at 6% compound annual return, the sum in 2024 would exceed $200 billion, a figure in the same neighborhood as Manhattan’s current assessed real estate value.

Twenty-four dollars. Four hundred years. Two hundred billion.

Let that math sit for a moment, because it contains something counterintuitive. Nobody did anything brilliant during those four centuries. Nobody hit a home run. Nobody made a killing. The 6% return is boring. It’s average. It’s the kind of return you don’t write home about. And yet, compounded over enough time, average performance at a boring rate turns $24 into a sum that can buy a significant fraction of the most valuable island on earth.

A 7% annual return doubles your money in ten years. Quadruples it in twenty. Multiplies it about a thousandfold in a century, which is four generations. A hundred years of quiet, consistent, boring 7%.

Human brains are bad at this. Our intuitions, calibrated on the African savanna where everything important scaled in straight lines (walk twice as far, cover twice the ground; carry twice the weight, burn twice the energy), have no built-in module for exponential growth. You have to override your instincts and run the numbers. The numbers are always bigger than you expect.

Capitalism won the Cold War on these numbers. In any given year of the 1950s and 1960s, Soviet GDP growth sometimes matched American growth. Single-year snapshots made the race look close. But compound interest doesn’t care about snapshots. It cares about accumulation. Each year’s small efficiency edge on the capitalist side, stacked on top of itself, compounded for four decades, produced a gap so visible that you could see it with your eyes the night the Berlin Wall opened. East Germans walking through the checkpoints into West Berlin in November 1989 were walking into forty years of compound interest differential made visible in neon signs, stocked shelves, full parking lots, and shop windows displaying goods that didn’t exist in any store in Dresden or Leipzig.

The engine works like this: property rights give people the incentive to invest (the returns are mine). Market pricing tells them where to invest (prices carry information about where capital is most needed). Time lets compound interest do the rest.

The Soviet system broke both preconditions. No real property rights destroyed the incentive. No market prices destroyed the information. The engine couldn’t start.

Hold these three pieces in your head. Property rights. Market pricing. Compound interest. AI is reaching into the first two at the same time, and what that does to the third is the thing I keep circling back to, the thing that keeps me up at night.

Zero Marginal Cost, Infinite Returns

For the entire history of industrial capitalism, about five centuries if you start counting from the Dutch merchants, “capital” has carried one defining physical trait.

It is scarce.

A factory takes up a fixed parcel of land, houses a fixed number of machines, and requires a fixed number of workers. Richard Arkwright built the first water-powered spinning mill on the banks of the River Derwent in 1771. Its output capacity was set by the power of the water wheel and the size of the building. Arkwright wanted more output. He built a second mill. Then a third. Each one cost about as much as the first.

Henry Ford installed the first moving assembly line at Highland Park, Michigan, in 1913. A single line produced a fixed number of Model Ts per day, determined by the line’s length, the belt speed, and the number of workers stationed along it. Ford wanted more cars. He built a second line. Then River Rouge, the largest factory complex in the world. Even River Rouge had a ceiling.

A container ship carries one cargo on one route. A building sits on one lot in one city. A railroad connects two points with a fixed throughput capacity. Every physical capital good in history has scaled with its customers: serve twice as many, spend close to twice as much.

This linear scaling is the reason capital earns a return. You’re deploying a scarce resource. Your competitors need years and comparable investment to copy it. The gap between what your capital produces and what it costs to replicate is where your margin lives. Piketty’s famous r > g, the tendency of capital returns to outpace economic growth, rests on this scarcity. If anyone could copy your factory for free, competition would push your profit to zero. Econ 101.

For five hundred years, that “if” stayed in the textbook. No meaningful capital good could be copied at zero cost.

Then someone trained a large language model.

A trained AI model is, in economic terms, a capital good. It’s a productive asset. Feed it an input, it produces an output. Training one requires enormous upfront investment: hundreds of millions of dollars for a frontier model. In that respect, it looks like a factory. Large fixed cost, productive output, big investment to build.

But it has one property that no factory, no ship, no building, no railroad has carried in the history of economic production.

Serving one user and serving a hundred million users costs almost the same.

Arkwright’s mill couldn’t do this. Each additional customer meant additional spindles, additional cotton, additional labor, additional cost. Ford’s line couldn’t do this. Each additional Model T needed its own steel, rubber, paint, man-hours. But a trained AI model can serve a million users at once for a fractional increase in compute cost. Training the model is the fixed cost. Running it for each additional user is near-zero marginal cost. The model is an information good, and information goods follow different economics than physical goods. You can copy them. The copy performs as well as the original. The copy costs almost nothing to produce.

Consider what this does to Piketty’s equation.

r > g holds across five centuries of data because capital is scarce, hard to replicate, and risky to deploy. These constraints cap the return. Your factory earns well, but capacity limits the upside, and competitors can build their own. r exceeds g within bounds.

AI capital breaks those bounds.

When your “factory” is a model and your “product” copies to a billion users at near-zero cost, the numerator of your return (revenue from all those users) grows without limit, while the denominator (additional capital required to serve them) sits pinned near zero. For the handful of organizations that own frontier models and the compute infrastructure underneath, r is accelerating toward a number that the system, the institutions, the social contract, none of it was engineered to handle.

But the other side of this split matters more.

AI is compressing the returns on every other form of capital by dissolving the scarcity premiums those assets depend on.

Office towers earn their rents because knowledge workers need to gather in one physical room. If AI-coordinated distributed collaboration matches face-to-face productivity, the scarcity of prime urban office space fades. You’re paying $150 per square foot for a premise that weakens every quarter.

Big law firms bill high rates because legal analysis is scarce. Three years of law school, a decade of practice, and you’ve built expertise that’s hard to copy. If AI handles 80% of the legal research and document review that junior associates bill at $400 an hour, the billing pyramid that supports partner compensation loses its base. The senior partner’s judgment may retain value for a while longer. The structure funding it is contracting.

SaaS companies sell subscriptions because most businesses can’t build their own software. If AI assembles a custom application tailored to a specific workflow in an afternoon, the value of generic pre-built software drops. Why subscribe to a one-size-fits-all tool when you can have a tailored one for less?

Each case follows one logic: AI dissolves the scarcity that made the traditional asset valuable. The scarcity premium compresses. The return falls.

The picture that forms is a split in the nature of capital itself. One class of capital (AI models and the compute behind them) sees returns accelerating toward the upper bound of what the system can contain. Every other class (real estate, professional expertise, packaged software, conventional IP) sees returns falling as AI erodes the scarcity those assets depended on.

Piketty described a slow, generational process of wealth concentration. AI is switching it to fast-forward, and the acceleration runs in one direction: toward the handful of people who own the models.

Ownership Without Founders

I need a detour through history here. It connects to AI in a way that might not be obvious at first, but the connection is the crux of the problem.

Property rights feel natural, like gravity. You make something, you own it. But this feeling is recent. For most of recorded history, “ownership” meant something so different from the modern concept that the same word misleads.

In medieval Europe, land “ownership” was a layered, ambiguous web. The king “owned” all land in the realm in theory. Nobles “held” land from the king, as grants tied to military service. Peasants “used” land the nobles held, in exchange for labor and a share of the harvest. Nobody “owned” land in the modern sense: an exclusive, tradeable right, enforced by independent courts, that you can sell, bequeath, or pledge as collateral without asking permission from anyone above you.

The transition from feudal web to modern property took about three centuries. England’s Enclosure Acts, stretching from the 1400s through the 1800s, converted common grazing land into private farms. This process uprooted millions of families from the land their ancestors had worked for generations; those displaced families became the labor force for the Industrial Revolution’s factories. The English Civil War of the 1640s was, among other things, a property fight. The Glorious Revolution of 1688 settled key disputes about the relationship between crown and private ownership. France chose a faster, bloodier path: the Revolution of 1789 abolished feudal property in a single act, then spent a generation and the Napoleonic Code building a replacement.

The payoff of all that upheaval: clear, tradeable, court-enforced property rights. And those rights turned out to be the on-switch for the compound interest engine. England finished first. England industrialized first. The Homestead Act of 1862 turned the American West into private land, and capital flooded in behind the settlers. Japan’s Meiji Restoration rebuilt the property system, and within twenty years Japan had factories. The pattern repeats: property rights don’t follow economic growth. They precede it.

Now bring this to AI, and the problem comes into sharp focus.

A piece of code can be copyrighted. You wrote it; you own it. An invention can be patented. You designed it; the patent office grants exclusive rights for twenty years. A trade secret can be protected. You keep it locked up; law penalizes anyone who steals it. These tools handle 20th-century assets well. Clear categories. Clean boundaries.

But the capabilities that a frontier AI model develops after training on trillions of tokens: who owns those?

I keep pulling at this thread, and it doesn’t resolve.

The capabilities come from training data: trillions of tokens of text, books, papers, code, web pages, forum posts. Billions of people produced this material over decades. Most never knew their work was feeding a training pipeline. None gave consent. None received payment.

The capabilities come from architecture: researchers at Google Brain published the Transformer in a 2017 paper, built on decades of academic work in attention mechanisms and backpropagation, most of it funded by public grants and university budgets.

The capabilities come from compute: tens of thousands of GPUs running for months. NVIDIA designed the chips. TSMC fabricated them. ASML built the lithography machines. A global supply chain made the training run possible.

And the most valuable capabilities, general reasoning, creative generation, the ability to handle tasks nobody anticipated, nobody programmed in. They emerged. They arose from the interaction of data, architecture, and compute at a scale that crossed thresholds nobody predicted. No engineer coded “the ability to reason about contract law.” That ability showed up when the system got large enough and trained long enough on varied enough text.

These emergent capabilities can’t be traced to specific training data. They can’t be pinned to a specific architectural choice. They are collective phenomena, the whole exceeding the sum of its parts, in a way that existing intellectual property law was never designed to handle.

We’re looking at a new class of asset, a strong candidate for the most valuable asset class of the coming century. And the legal system of every country on earth lacks the tools to assign ownership of it.

In the absence of a framework, raw power fills the gap. The companies that train the largest models hold the capabilities the way a feudal lord held land: through possession and force of position. A lord’s title rested on armed men, not statute books. “This land is mine because my soldiers say so.” It worked for centuries. It was incoherent in principle.

As economies grew and more parties developed stakes in how land was used (merchants wanting to buy it, creditors wanting to lend against it, tenants wanting security), disputes accumulated with no legitimate resolution mechanism. The pressure built until the system either reformed through legal invention (England’s path, slower, less blood) or shattered through revolution (France’s path, faster, far more blood).

AI’s property vacuum is the same type of problem. “Whoever has the compute” settles the ownership question today. But that settlement won’t hold. Too many stakeholders have too much at stake: the billions of people whose data trained the models, the taxpayers who funded the foundational research, the governments that want oversight, the competitors who want access, the users who depend on the output. As the economic value of AI capabilities grows, the gap between de facto ownership and legitimate ownership will widen. And widening property gaps don’t resolve themselves on their own. They never have.

The Player With X-Ray Glasses

Back to the poker table.

Millions of players, different cards, overlapping bets. The table works because of one condition so basic that nobody at the table thinks about it:

Each player reads and processes cards at about the same speed and depth.

You might be sharper than me. Better instincts, more experience, faster math. But the gap between us is a gap of degree. We’re both human. Both running the same type of biological hardware. Both contributing to the price signal through our bets.

Now picture a new player taking a seat. This player wears glasses that let him see through the backs of every other player’s cards.

Nobody else knows about the glasses. The game continues. Cards dealt, bets placed, chips moving. Walk into the room and you’d see a normal poker game.

But the game’s structure has changed. The player with X-ray glasses reads every hand, calculates every probability, bets with certainty against opponents who bet with uncertainty. No skill required. No insight. The glasses do the work. Chips flow, with mechanical regularity, from every other seat toward his.

The table “works” in the sense that transactions occur. It has stopped working in the way that once mattered: as a fair contest where the prices that emerge reflect the collective judgment of many independent minds.

AI is the glasses.

A hedge fund’s AI system scans the full order book across global markets in milliseconds. It cross-references satellite images of retail parking lots, credit card spending data from consumer panels, sentiment analysis of earnings call transcripts, supply chain shipping logs. Then it executes. Ten rounds of trades in the time it takes a human analyst to read a single page of a 10-K filing.

The gap between this system and a human analyst is a gap of kind. A good analyst and a great analyst differ in degree; they’re both human, reading with human eyes, thinking at human speed. An AI system perceives dimensions of the data that the human brain, as a biological instrument, cannot access. The word is “cannot,” not “can, at a slower speed.”

This distinction has a sharp edge when it comes to regulation.

Traditional information asymmetry (I know something you don’t) responds to existing tools. Insider trading laws. Mandatory disclosure rules. Force companies to publish financials, and the information playing field levels.

Processing asymmetry (we hold the same data, but one of us can extract patterns from it that the other’s brain can’t perceive in principle) doesn’t respond to disclosure. The data is already public. The 10-K is on the SEC website. The satellite images are for sale. The gap isn’t in what each player has. It’s in what each player’s cognitive machinery can do with it.

The fix for information asymmetry: give everyone the same data.

The fix for processing asymmetry: give everyone the same AI.

No financial regulator on earth is set up to do that. No law on the books imagines it. The tools we built to keep the poker table fair, disclosure rules, insider trading enforcement, market surveillance, all assume that the players are the same kind of entity. Humans, with human brains, competing on human judgment. That assumption is expiring.

And the dynamic extends beyond financial markets.

In labor markets: when an employer’s AI can gauge a worker’s output, predict their future performance, and compute the minimum salary the worker will accept, all with precision no HR team can match, wage negotiations stop being negotiations. They become an optimization exercise run by one side’s machine against the other side’s gut. The “price” of labor starts reflecting the employer’s AI mapping the worker’s tolerance for being underpaid.

In real estate: when an AI models neighborhood trajectories, zoning changes, demographic shifts, and infrastructure investments with precision no individual buyer can match, property prices stop reflecting “the market’s view.” They reflect the view of whoever runs the best model. You walk through an open house and form a gut estimate. An institutional AI has already analyzed twenty years of sales on that block, every zoning filing, three bus route changes, and a Census Bureau migration dataset. It placed its bid before you turned the key in your car.

Transactions keep happening. Prices keep moving. The table looks the same. But hand by hand, it is shifting from a game where collective judgment sets fair prices toward a mechanism where the player with the best glasses extracts from everyone else.

Three Exit Scenarios

I don’t know which of these we’re heading toward. I do know that the people who sound most confident about the answer are the ones whose confidence I trust least.

But I can push the three mutations to their structural endpoints and lay out the territory each one points toward. Think of these as the way a VC might think about exit scenarios for a portfolio company. Except the company is the global economic order, and the portfolio is your life.

The Velvet Cage

The compound interest engine keeps running. Returns collect in the accounts of AI model owners. Most people live in material comfort because AI productivity keeps goods cheap. Food affordable. Housing adequate. Healthcare delivered by AI diagnostics at low cost.

But economic sovereignty evaporates.

You work on an AI-owner’s platform because outside it the economy has thinned to the point where competitive alternatives barely exist. You consume through it because AI-produced goods undercut every rival on both price and quality. Your information, entertainment, and social connections run through AI curation because the curated version, by every measurable standard, outperforms the uncurated one.

You’re not poor. You’re not hungry. No one holds a weapon on you.

But every axis of your economic life runs on infrastructure someone else owns, and no meaningful alternative exists.

The historical parallel is the medieval manor. Feudal peasants were not, as a rule, starving. They had land, communities, festivals. Feudalism’s defining trait was structural dependency: your economic life was shaped by your relationship to the lord’s land. You couldn’t choose your employer or negotiate your terms.

Replace “land” with “AI platform.” The architecture rhymes.

And this doesn’t require a villain. It’s the natural arithmetic of compound interest under AI conditions. Medieval feudalism wasn’t “designed.” It was the equilibrium that military power and land distribution produced over centuries. Nobody planned it. Everybody ended up in it.

The Fork

If augmentation technologies arrive, brain-computer interfaces, cognitive enhancement, deep AI-human integration, the gap between enhanced and unenhanced humans could shift from degree to kind.

Some people learn ten times faster, analyze with a hundred times more precision, synthesize across a thousand times more dimensions. Their interface with AI runs deeper, and that depth creates the gap.

Compound interest accelerates in the upper stratum. The lower stratum doesn’t get oppressed. It gets outpaced. Two economies, two societies, decoupling over time.

A civilization splitting along lines of cognitive architecture rather than geography, culture, or class. One species. Two trajectories.

The Bismarck Option

In 1883, Otto von Bismarck, a Prussian aristocrat who viewed socialists with open contempt, built the world’s first national social insurance system. State pensions. Accident insurance. Health coverage for workers. He did this because the Social Democratic Party was gaining seats in the Reichstag, factory strikes were disrupting the economy, and he needed a valve to release pressure before it blew the system apart. His calculation: keep the engine running, but give workers enough security that they stop voting for people who want to tear it down.

Bismarck didn’t scrap the engine. He bolted a redistribution module onto it: small enough to preserve performance, large enough to absorb the political pressure that was building toward revolution. Franklin Roosevelt ran a similar play with the New Deal in the 1930s. Postwar Western Europe ran it again with the welfare state.

AI may need a play from the same playbook, in a form nobody has invented yet. Compute as public infrastructure, like roads and power grids. Distribution mechanisms embedded in AI systems, routing a fraction of every AI-generated dollar back to the data contributors who made the system possible. A new form of property right that can handle emergent capabilities the way copyright handles creative expression and patents handle inventions.

This is the best exit. It’s also the hardest to reach, because it demands institutional innovation at the speed of technical change.

The historical scorecard on that race offers no comfort.

Industrial production methods reshaped England starting in the 1780s. Matching institutions, factory acts, union legalization, social insurance, universal suffrage, the welfare state, arrived around the 1940s. The gap between the technology and the institutions: about 160 years. In that gap: the final phase of the Enclosure Acts, the Chartist movement, the failed revolutions of 1848, the Paris Commune, two world wars, the Great Depression, and fascism on three continents.

The institutions caught up. The catching-up killed tens of millions of people.

How much time do we have this round? A generation, maybe. Maybe less.

I keep coming back to that sidewalk on Central Park South.

Mamdani pointing his camera at 24,000 square feet of empty penthouse. Griffin calling it dangerous. Ackman tweeting. The news cycle grinding. Left versus right. Populism versus capital. The same script running since the 19th century.

But the thing that makes this moment different from every previous round of this fight is the engine underneath it. The engine that made Griffin’s $238 million possible. The engine that makes Mamdani’s tax revenue possible. The engine that built the city both of them are standing in.

That engine is mutating. AI is pushing its returns past the limits it was built for. AI is creating assets that the property framework can’t define. AI is seating a player at the poker table who can see through the cards.

Mamdani thinks he’s fighting a tax fight. Griffin thinks he’s fighting a capital-mobility fight. They’re both right, in the same way that two people arguing about the arrangement of deck chairs are right about which chairs look better where.

The ship is the thing.

Glushkov believed, in 1962, that computation would save socialism. He spent a career on this conviction. The idea that computation might pose the deepest challenge to capitalism never entered his work.

I don’t know which exit scenario we’re headed for. I don’t think anyone does, and I distrust anyone who claims certainty. I do know this: the frameworks we’re using to debate the moment, socialism versus capitalism, left versus right, regulation versus free markets, were forged in the Industrial Revolution to handle that transformation’s problems. They did the job. With enormous pain and delay, they did it.

The problems arriving now are different in kind. The capital is different: it copies itself. The property is different: existing law can’t define it. The market is different: some players see through the cards.

We need new frameworks. We need a new term sheet for the 21st century. Mamdani and Griffin are both arguing over the old one, and I can’t fault either of them, because the new one doesn’t exist yet.

The question is who writes it, and how fast.

The last time the engine mutated, in the 1780s, the new term sheet took 160 years to arrive. The wait included the most destructive events in human history.

The engine is mutating again. Faster this time.

I’d prefer not to wait 160 years.

I don’t think we have them.

It’s a beautiful piece of paper. Olive-green scalloped border, the kind of ornamental engraving that no one does anymore. Across the top, a panoramic scene of Boston Harbor: sailing ships, steamships, a city skyline rising behind them. On the left side, a steam locomotive pulling a line of freight cars. In the center, the Massachusetts state seal, flanked by classical figures. Bottom right, an allegorical scene of Commerce and Industry, a winged figure and a seated goddess surrounded by the tools of trade.

On the right side, two stamps. “CANCELLED. Jul 23, 1949. State Street Trust Company.” The shares were transferred and voided eight years after they were issued.

The company that printed this certificate no longer exists. Neither does the company that acquired it. Or the company that acquired that company. Or the one after that.

The railroad it financed is still running.

I kept staring at this thing, because the more I learned about the story behind it, the more it started to feel like a message from the past about what’s happening right now. So I went down the rabbit hole. Three weeks later, I came back out, and I think this story is one of the most important things you can read if you’re trying to make sense of the AI moment we’re living through.

Let me explain.

Erie Canal Was the First-Mover Advantage

In 1825, the state of New York opened the Erie Canal. If you’re thinking “a canal, how exciting,” you’re underestimating what this thing did. The Erie Canal connected the Great Lakes to the Atlantic Ocean via the Hudson River, which meant New York City could now move goods to and from the entire American interior for a fraction of the previous cost. Freight rates dropped by 90%. Ninety percent. New York went from one of several competing East Coast ports to the dominant commercial hub in the Western Hemisphere, basically overnight.

Boston watched this happen and panicked.

Boston had no river connecting it to the interior. Worse, the Berkshire Mountains sat between Boston and Albany like a geological middle finger. You couldn’t build a canal over them. For a few years, it looked like Boston would fade into regional irrelevance while New York ate the continent.

Then someone had an idea: what about this new “railroad” technology?

Now, railroads in the 1830s were about as proven as large language models were in 2021. A few short lines existed in England. Nobody had built one over a mountain range. The engineering was uncertain, the costs were unknown, and most sensible investors thought the whole concept was speculative nonsense. (Sound familiar?)

Boston bet on it anyway. In 1831, the state chartered the Boston & Worcester Railroad. A few years later, the Western Railroad started pushing west toward Albany. By 1842, the full line was open. Boston had its connection to the interior, and it hadn’t needed a river or a canal to get it.

The cost was staggering. That one railroad from Worcester to Albany cost $8 million, about the same as the entire Erie Canal. Except the canal was publicly funded by the state of New York. The railroad was financed with private capital, through a financial instrument that most Americans had never used before:

Bonds.

This is where the story gets wild.

Railroads Were the Platform Play

Before railroads, America’s financial system was small and simple. The federal government issued some debt. Banks made loans. A handful of state-chartered companies traded shares on the New York Stock Exchange, which in the 1820s was a few dozen guys meeting under a buttonwood tree. That was it. That was the whole financial system.

Railroads broke it.

They needed more capital than any private enterprise had ever required, and they needed it for longer. You can’t build a railroad with a six-month bank loan. You need money for years before a single train generates a dollar of revenue. The time lag between investment and return was enormous.

So railroad companies started issuing bonds. Lots of them. A bond said: give us $1,000 now, and we’ll pay you 5% per year for 20 or 30 years, then give your $1,000 back. For investors, this was attractive because it paid a fixed income and (in theory) returned your principal. For railroads, it was the only way to raise the kind of money they needed at the timescale they needed it.

By 1859, American railroads had issued over $1.1 billion in bonds. In 1859 dollars. To put that in perspective: the entire federal budget that year was about $69 million. Railroad debt was sixteen times the federal budget.

That bond market transformed the New York Stock Exchange from a sleepy club into the center of global capital markets. Trading volume went from a few hundred transactions a week to hundreds of thousands.

And the money wasn’t coming from America alone. British banks underwrote railroad bond offerings. German investors, sitting on savings they couldn’t deploy in their own fragmented economies, bought American railroad debt by the trainload. Capital flowed across the Atlantic for the first time at industrial scale.

I want you to sit with this for a second, because I think most people (including me, before I went down this rabbit hole) don’t realize what happened here. The basic financial infrastructure that today’s economy runs on, corporate bonds, public equity markets, institutional underwriting, cross-border capital flows, credit analysis, even the concept of a “balance sheet,” all of it was built to finance railroads. Railroads didn’t just change transportation. They changed money. They were the platform layer underneath everything else.

$700B Is the New Railway Mania

Okay, so railroads rewired capital markets. How big was that rewiring, and how does it compare to what’s happening with AI?

Let me try to make these numbers feel real, because they’re the kind of numbers that are so large they slide right off your brain.

In the peak years of the 1850s, railroad investment hit about 3% of American GDP. When you count the knock-on spending (land development, equipment, labor camps), the total reached 4-5%. In Britain during Railway Mania in the 1840s, it was even crazier: railway investment surged to 7% of GDP, representing half of all investment in the entire economy. Half. Every other industry in Britain combined was investing the same amount as railways alone. Some scholars put it at 10-20% of private capital formation in peak mania years.

No single technology before or since has consumed that much of a nation’s savings.

Until, maybe, now.

In 2026, five companies (Microsoft, Alphabet, Amazon, Meta, and Oracle) plan to spend between $660 billion and $700 billion on capital expenditure, most of it for AI. That’s up from $162 billion in 2022. A quadrupling in four years. U.S. GDP is roughly $29 trillion, so $700 billion from just five companies is about 2.4% of GDP. To make that tangible: imagine the U.S. economy as a dollar. These five companies are taking about two and a half cents of every dollar produced in America this year and converting it into data centers, chips, and cooling systems. And that’s before you count the chip manufacturers, the energy buildout, the fiber optics, the neocloud startups like CoreWeave and Lambda, and everyone else piling in. Add all of that, and we’re approaching railroad-era levels of capital intensity. McKinsey projects that global AI capex will require $6.7 trillion by 2030 to keep pace with demand. Trillion. With a T.

But the comparison that should make you pause is on the revenue side.

In 1859, American railroads had $1.1 billion in bonds outstanding, and they were generating revenue from every train that moved. Freight paid by the ton-mile. Passengers paid by the ticket. The business model was tangible and immediate. You could stand at a station and watch your investment earn money.

In 2026, OpenAI’s annual recurring revenue is about $20 billion. Anthropic’s is about $9 billion. Combined, that’s $29 billion. The hyperscalers spending $700 billion on AI infrastructure are generating some AI-related cloud revenue on top of that, but nobody can cleanly separate “AI revenue” from “cloud revenue” in their reporting. Alphabet’s free cash flow is projected to fall nearly 90% this year, from $73 billion to $8 billion. (Read that sentence again. Ninety percent.)

The railroad investors of the 1850s could at least count the number of trains running on their tracks. AI investors in 2026 are making a bet that the revenue will come, because the technology is too important for it not to. That bet might be right. It was right for railroads, too, in the long run.

In the short run, though, the story got ugly.

Jay Cooke Was the First Blitzscale Casualty

On September 18, 1873, Jay Cooke & Company went bankrupt.

If you haven’t heard of Jay Cooke, think of him as the most trusted name in American finance. During the Civil War, he was the guy who figured out how to sell Union war bonds to ordinary citizens, basically inventing retail financial marketing. He was the federal government’s primary banker. In modern terms, he was a combination of Goldman Sachs, the U.S. Treasury Department, and a patriotic marketing campaign, all rolled into one guy.

After the war, Cooke took on the financing for the Northern Pacific Railway, a line that was supposed to connect the Great Lakes to the Pacific coast. Huge project. Very exciting. Investors loved it.

The problem was that the Northern Pacific was burning cash faster than it could sell bonds. (Again: sound familiar?) The first transcontinental railroad had already been built, and investors were starting to wonder whether America needed a second one. Bond prices dropped. Cooke’s firm couldn’t meet its obligations.

And then everything fell apart at once.

The New York Stock Exchange closed for the first time in its history. Ten days of no trading. Eighty-nine of America’s 364 railroad companies went bankrupt. Eighteen thousand businesses failed within two years.

But here’s the part that messed me up when I read about it, the part that I think is the most important lesson in this entire story:

The devastating part of the 1873 panic wasn’t domestic. German investors, who had been buying American railroad bonds for a decade, got spooked. A real estate bubble in Vienna had just burst. The German economy was opening up new domestic investment opportunities after unification. So German capital started flowing out of American railroads and back to Europe.

Read that again. A real estate bubble. In Vienna. Crashed the American railroad industry.

Nobody in 1872 was modeling “Viennese real estate bubble bursts → German capital exits American railroad bonds → 20-year depression follows.” The trigger had nothing to do with railroads. The railroads were fine. The routes were good. The trains ran. But the financial system that funded them was entangled with forces on the other side of the ocean, and when those forces shifted, everything crumbled.

Twenty years later, it happened again. Between 1893 and 1897, companies owning a third of all American rail mileage went through bankruptcy. This time the cause was domestic: too many companies had built parallel lines competing for the same routes, driving prices below the cost of service. None of them could generate enough revenue to pay their bond interest. They had built the infrastructure first and assumed the economics would follow. For many of them, it didn’t.

(If you’re keeping score at home: blitzscale, check. overbuilt competing infrastructure, check. revenue that couldn’t keep up with capital deployed, check. exogenous shock from an entangled financial system nobody fully understood, check. We’re running this playbook again right now, except with GPUs instead of rail ties.)

J.P. Morgan Invented the Rollup

So the industry is in ruins. A third of American railroads are bankrupt. Investors have been wiped out. Who shows up?

J.P. Morgan.

Morgan’s playbook was simple and ruthless: buy the bankrupt railroad’s debt for pennies. Foreclose. Create a new company with less leverage. Raise fresh capital. Install management you trust. Put your partners on the board, including the boards of competing railroads, so they stop cutting prices against each other.

Between 1893 and 1898, Morgan restructured the Southern Railway, Erie Railroad, Reading Railroad, and Northern Pacific. Other banks followed his model: Kuhn Loeb took Union Pacific, Speyer took B&O. By 1900, Morgan controlled about a sixth of all American railroad track.

They called this process “Morganization.” A restructured railroad was said to have been “Morganized.” The term carried a mix of respect and dread, because Morganization worked, but it also meant the original investors, the people who took the risk and funded the construction, were wiped out. The builders and the eventual owners were different groups of people.

If you work in finance, you recognize this playbook immediately. Morgan invented what we now call private equity restructuring, distressed debt investing, and activist board governance. He didn’t create these concepts from a textbook. He created them because the railroad industry was a catastrophe and someone had to clean it up, and it turned out that cleaning up catastrophes was the most profitable business of all.

I keep thinking about this when I look at AI. The question isn’t just “who builds the best model?” The question is: when the shakeout comes (and if the railroad precedent means anything, a shakeout will come), who will be the Morgan? Who’s sitting on the sidelines with enough capital and enough patience to buy the wreckage and reorganize it?

My guess: it won’t be a name you’re reading about in TechCrunch right now.

Right Thesis, Wrong Cap Table

Five companies are spending $700 billion this year to build AI infrastructure. Their combined cash reserves are about $420 billion. These are the most profitable companies in the history of capitalism, and they are burning through their cash at a rate that would have made Jay Cooke uncomfortable.

OpenAI has announced over $1.4 trillion in data center deals. The Stargate project alone is targeting $500 billion in infrastructure. Anthropic, xAI, and a constellation of “neocloud” companies (CoreWeave, Nebius, Lambda, Crusoe) are all building as fast as they can.

The question everyone keeps asking is: “Is this a bubble?”

I’ve spent three weeks with the railroad story now, and I think that’s the wrong question. The railroad investors of the 1860s weren’t wrong about railroads. Railroads changed everything. They connected the continent, enabled industrial agriculture, created the modern corporation, and generated trillions of dollars in economic value over the next century. The investors who bet on railroads were right about the technology.

Many of them still lost everything.

They lost because the financial structures built around the technology were fragile. They lost because capital flowed in from places they couldn’t control and flowed out for reasons they couldn’t predict. They lost because competing companies overbuilt capacity and destroyed each other’s pricing. They lost because the gap between “this technology will be transformative” and “this specific company will generate enough cash flow to service its debt” is wide enough to swallow entire fortunes.

The railroad boom peaked in 1916, when the U.S. had 254,000 miles of track. Today there are about 140,000 miles. That’s a 40% decline. The technology won. Many of the companies and investors who funded it did not.

The thesis was right. The cap table was wrong.

Seven Owners, One Railroad

The Boston & Albany Railroad survived the Panic of 1873. It survived the Panic of 1893. It survived because its route was irreplaceable (you still can’t build a highway over the Berkshires faster than running a train through them), because its management was disciplined, and because its capital structure wasn’t overextended.

In 1900, the Vanderbilt family’s New York Central Railroad was impressed enough to lease B&A for 99 years. The railroad ran under that lease until 1961, when NYC absorbed it fully. Then NYC merged into Penn Central in 1968. Penn Central went bankrupt in 1970, the largest corporate bankruptcy in American history at that time. Then Congress created Conrail. Then CSX bought Conrail’s assets.

My certificate was issued in 1941, four decades into that 99-year lease. By then B&A was a subsidiary of New York Central in all but name. Hussey & Co. bought 25 shares for $2,500. Eight years later, State Street Trust Company stamped it “CANCELLED” and transferred the shares. The certificate became a piece of paper. The railroad kept moving freight.

I’ll say it plainly: the track from Boston to Albany has been owned by seven different entities in 190 years. Seven. Boston & Worcester. Western Railroad. Boston & Albany. New York Central. Penn Central. Conrail. CSX. Each ownership change created a new set of winners and losers among the people who held the financial instruments. Some shareholders were bought out. Some bondholders were paid in full. Some got pennies. Some got nothing.

The trains kept running.

The Exit Nobody Planned

I don’t know who wins the AI race. Nobody does. If you’d asked someone in 1865 which railroad company would dominate American transportation in 1920, they would have given you a confident answer that was wrong. I’m pretty sure we’re in 1865 right now.

But I think the railroad story, if you take it seriously, tells you six things about what happens next.

First, the technology will work. That part isn’t in doubt. Railroads worked. AI works. The question was never “will trains move faster than horses?” or “will language models generate useful output?” The answer to both was always yes. The technology question is settled. Everything that follows is a finance question and an ownership question, and those are much harder.

Second, the buildout will be larger than anyone currently projects. In 1850, America had 8,879 miles of track. By 1860, it had 30,626. Nobody in 1850 would have believed that number. AI infrastructure spending has quadrupled since 2022 and shows no sign of slowing. McKinsey’s $6.7 trillion projection for 2030 might end up being low. When a technology changes the cost structure of everything, the capital required to build it out has a way of exceeding every estimate, including the ones that already seemed crazy.

Third, the financial system will change in ways we can’t anticipate. Railroads didn’t just use the existing capital markets; they created new ones. Bond markets, equity markets, underwriting syndicates, credit analysis, bankruptcy law, corporate governance, even the concept of the limited liability corporation, all got reshaped or invented to handle railroad finance. AI will do the same. We don’t know what the new financial instruments will look like yet. Maybe compute futures. Maybe revenue-sharing tokens tied to model performance. Maybe something nobody has named yet. The railroad precedent says the instruments themselves will be part of the story.

Fourth, a crisis will come, and its trigger will be something nobody is watching right now. In 1873, it was a Viennese real estate bubble. In 1893, it was the cumulative effect of rate wars nobody thought would last. Whatever hits AI won’t be “AI doesn’t work.” It’ll come from some adjacent system that’s quietly entangled with the AI buildout in ways nobody has mapped. A chip supply disruption. A sovereign debt crisis in a country that’s heavily invested in AI infrastructure. An energy bottleneck. Something nobody is talking about at Davos or on the All-In Podcast. The black swan, by definition, is the one you’re not looking for.

Fifth, the people who build it and the people who own it long-term will be different. The merchants of Boston who funded the Boston & Worcester Railroad in 1831 were not the Vanderbilt family that controlled it in 1900, and neither of them were CSX, which runs it today. OpenAI’s current investors, Anthropic’s current investors, the hyperscalers currently spending $700 billion a year: these may or may not be the entities that own and profit from AI infrastructure in 2050. Morgan got rich not by building railroads but by picking up the pieces after other people’s railroads collapsed. Someone will play that role in AI. We don’t know their name yet.

Sixth, and this is the one I keep coming back to: the infrastructure will outlast everyone’s financial projections. The tracks the Boston & Albany laid in the 1840s are still carrying freight in 2026. Data centers being built today will run computation, in some form, for decades after the companies that built them have been restructured, merged, acquired, or dissolved. The physical layer endures. The capital layer above it churns.

I keep the stock certificate on my desk. Certificate B79362. Cancelled July 23rd, 1949.

A new technology that changes the cost structure of everything. A flood of capital chasing the opportunity. Financial instruments invented on the fly to channel that capital. A period of expansion that feels like it will never end.

Then a shock. Then wreckage. Then reorganization. Then, slowly, the emergence of a new system that looks nothing like what anyone planned but works better than what came before.

The trains kept running through all of it.

I think the models will, too.

The more junior people at the same organizations are often the opposite. Harder, more transactional, quicker to let you know who they represent. You’d expect seniority to make people tougher. It seems to do the reverse.

Subscribe now

For a while I thought this was just personality. Some people are warmer than others, and maybe warmer people are slightly more pleasant to promote. But that explanation doesn’t survive contact with the data. The pattern is too consistent across too many firms, too many roles, too many markets. I see it in New York and in secondary cities. I see it in LPs and GPs and brokers and advisors. Whatever is producing this pattern, it’s structural, not personal.

I think the industry is selecting for trust. Continuously, and almost invisibly.

Here’s what the filtering looks like in practice. Nobody sends you an email saying “we’ve decided not to work with you because you’re difficult to deal with.” That’s not how it works. What happens is subtler. Your calls stop getting returned as quickly. Then they stop getting returned at all. Your LPs take the re-up meeting out of courtesy but the commitment never comes. You hear about a co-investment opportunity after it’s already closed, from someone who assumed you knew. Your deal flow thins out gradually, like a slow leak, and you can’t point to any single moment where things changed.

Most people who get filtered out this way never fully understand what happened. They think the market shifted, or their strategy fell out of favor, or they just had a run of bad luck. And sometimes that’s true. But sometimes the explanation is simpler: people in the ecosystem quietly decided they’d rather work with someone else.

Why is this industry’s filter so sensitive? Because of what the core transaction actually requires.

When an LP commits capital to a real estate fund, they’re handing money to a GP for seven to ten years. There’s almost no liquidity. There’s limited control. The LP gets quarterly reports and an annual meeting, but between those touchpoints the GP has enormous discretion. They choose which deals to pursue, how to manage the assets, when to sell, how to handle problems. The LPA tries to constrain this discretion, but a two hundred page legal document can’t anticipate every situation that will arise over a decade.

At some point, the LP is making a bet on a person. They’re betting that this person will deploy their capital thoughtfully, communicate honestly when things go wrong, manage conflicts of interest fairly, and not quietly optimize for their own economics at the LP’s expense. That bet is called trust, and there’s no contractual substitute for it.

This makes the institutional real estate ecosystem unusually punishment-sensitive to extractive behavior. In industries with shorter transaction cycles, you can get away with being difficult. If you’re trading securities, the counterparty might not like you, but the trade settles in two days and you move on. In real estate, you’re locked into relationships for years. The LP who commits to your fund is stuck with you. The co-investment partner on a deal is stuck with you. The property-level partners and lenders are stuck with you. Everyone is stuck with everyone for a long time. Under those conditions, people develop strong preferences about who they want to be stuck with.

And the ecosystem is small. Surprisingly small for the amount of capital it manages. The institutional real estate world in any given country is maybe a few thousand people who matter, and they all know each other, or know someone who knows each other. Reputation travels through this network faster than most people realize. When an LP is doing due diligence on a GP, they don’t just read the PPM and analyze the track record. They pick up the phone and call people. They ask around. “Have you worked with this person? What are they like? Would you commit again?” These reference calls happen constantly, informally, at conferences and over dinners and in passing. Your reputation is being discussed in rooms you’re not in, by people you may not even know are talking about you.

In this environment, every interaction is being recorded in a distributed, informal, permanent ledger. The way you treat the analyst who sets up your data room. The way you communicate when a deal underperforms. The way you behave in a negotiation when you have leverage and the other side doesn’t. None of these moments feel consequential when they’re happening. All of them are contributing to a reputation that will either compound in your favor or against you.

The people I mentioned at the beginning, the senior ones who radiate trustworthiness, didn’t start that way by accident. Some of them probably were always like that. But I think many of them simply figured out, earlier than most, what game they were playing. They realized that the industry would judge them less on any individual transaction and more on the pattern of how they treated people over time. And they adjusted accordingly.

The people who get filtered out are usually not villains. They’re not scammers or fraudsters. They’re people who treated every interaction as a discrete transaction to be won, without realizing that in this industry, there are no discrete transactions. Everything is connected. Everyone remembers.

Paul Graham once observed that mean people tend to fail in startups because the startup world is a positive-sum game, and meanness is a strategy adapted for zero-sum ones. Institutional real estate works the same way, maybe even more so. The capital is more patient, the relationships are longer, the ecosystem is smaller, and the consequences of lost trust take longer to appear but are harder to reverse.

Mean people don’t fail loudly in this business. They get quietly filtered out. The calls stop coming. The capital goes elsewhere. And usually, by the time they notice, the filtering happened years ago.

They turned on ninety power looms at once.

The sound was somewhere north of a jet engine. The docent handed out earplugs before we went in. Nineteenth-century mill workers didn’t get earplugs. They worked fourteen-hour shifts in that noise, six days a week, and a lot of them were deaf by thirty.

I walked out thinking about something that had nothing to do with deafness.

I was thinking: those looms never stopped. They just moved.

Lowell, Massachusetts → South Carolina → China → Bangladesh. Right now, this exact noise is happening in Dhaka. Women are standing at machines doing what New England farm girls did in 1830.

And here’s the weird part. Every stop along the way, people told completely different stories about what was happening.

In Lowell: “American ingenuity! National industry! Breaking free from British economic dependence!”

In 1990s China: “Socialism with Chinese characteristics! The people are getting rich!”

In Bangladesh: “Development! Globalization lifting the poor!”

These stories have nothing to do with each other. Some of them are ideological enemies. But the loom passed through all of them like they were nothing.

The loom doesn’t care about stories.

I’ve been turning this over in my head for weeks. I think the loom taught me something important about how the world actually works.

The Spy Who Remembered Everything

Let me tell you about Francis Cabot Lowell.

In 1810, this Boston merchant gets on a ship to England. He tells everyone he’s going for his health. He’s actually going to steal the most valuable industrial secret on Earth.

Britain in 1810 was basically what Silicon Valley is now, except instead of software, the killer app was cloth. The power loom—a machine that could weave fabric automatically—was the iPhone of its day. And Britain guarded it like Apple guards its source code. You couldn’t export blueprints. Skilled workers couldn’t emigrate. The penalty was severe.

Lowell didn’t steal blueprints. He just... looked at the machines really carefully. For months. Then he sailed home with the whole thing in his head.

Later, people told this as a patriotic story. American entrepreneur outsmarts British empire! Founds domestic industry! Economic independence!

Good story. Probably not what Lowell was thinking.

Here’s what Lowell was thinking: British cloth can’t get into America right now.

Jefferson had just slapped an embargo on British goods. Then the War of 1812 started. Suddenly you couldn’t buy British textiles at any price. Everyone saw this as a disaster. Lowell saw a gap in the market.

He had one piece of the puzzle in his head (the machine). Boston had another piece (rich merchants with cash looking for investments). The Merrimack River had another piece (water power). And New England farms had the last piece (labor).

Put them together. The spread is profit.

Was it patriotism? Maybe. But certainly it was arbitrage. Patriotism was the story you told investors and newspapers. Arbitrage was the math you did in private.

Lowell was what you might call a clear-eyed player. He understood what game he was actually playing, and he played it. The story was a tool. The game was the point.

The Labor Problem (And Its Creepy Solution)

Lowell had a problem, though.

British factories ran on desperate people. Peasants kicked off their land, orphan children, the absolute bottom of society with no other options.

America didn’t have that. Land was too cheap. If your life sucked, you could just go west. Nobody was desperate enough to work in a factory.

So where do you find workers?

Lowell’s answer: young, unmarried women from New England farms.

Think about the position these women were in. They couldn’t inherit the family farm—that went to sons. They couldn’t do much economically until they got married. They were, in a very real sense, leftovers of the patriarchal system. They had nowhere to be.

Patriarchy created the problem. Lowell exploited the problem.

To make it work, he had to convince their fathers to let them go. So he built dormitories. Hired chaperones. Made church attendance mandatory. Created a whole system of moral supervision that made the mills look respectable, almost like boarding schools.

This got called the “Lowell System” and people praised it as enlightened capitalism. Look how well they treat the workers! Dormitories! Libraries! Church!

I don’t think Lowell believed any of it. This wasn’t morality. It was a transaction cost. You want these workers? You have to pay their fathers. The payment comes in the form of a story about moral supervision.

By 1823, there was an entire city where nothing had existed before. A decade later, Lowell was one of the biggest industrial centers in America.

The Replacement

Francis Cabot Lowell died in 1817. He never saw any of this.

Which tells you something: he’d built a system, not a personal empire. Once the system was running, it didn’t need him anymore.

Here’s what happened next.

The mills made money. When something makes money, it gets copied. More investors, more mills, more competition. Margins got squeezed.

Suddenly all that moral supervision stuff started looking expensive. Dormitories cost money. Chaperones cost money. Libraries cost money. When profits are fat, you can afford this. When profits get thin, you start looking for cuts.

Hours got longer. Wages got cut. Conditions got worse.

The mill girls pushed back. They went on strike in 1834. Again in 1836. They quoted the Declaration of Independence. They demanded their rights.

Then history made its move.

1845: the Irish Potato Famine.

A million people died. Two million fled Ireland for America. They arrived with nothing, desperate for any work at any wage.

The mills didn’t need to court New England farm families anymore. Irish immigrants were cheaper and more desperate. They had zero bargaining power.

Within a few years, the wholesome story of the Lowell mill girls was over. The factories kept running. Different people.

Here’s what I want you to notice: nobody planned this. There was no conspiracy. No villain in a back room saying “let’s replace American women with Irish immigrants.”

The system faced cost pressure. A cheaper option appeared. The system took the cheaper option.

That’s it.

The system doesn’t know who you are. It doesn’t remember your contribution. It doesn’t care about your story. It asks one question: are you still the cheapest way to make cloth?

If yes, you stay. If no, you get replaced.

The Machine Keeps Moving

Then technology changed.

At first, one worker watched two looms. Then four. Then eight. Then twenty. Eventually the machines ran themselves and humans just monitored for problems.

Each improvement meant less human labor per yard of cloth.

Then geography changed.

Water power needs rivers. That limits where you can build. Steam engines broke that limit. Now factories could go anywhere.

New England wages crept up. The South was cheaper. The industry migrated south. Lowell started dying.

This played out over decades. Factories closed. Young people left. The city hollowed out. The grand mills became ruins. The ruins became a museum.

The system didn’t stop. It just wasn’t in Massachusetts anymore.

Plot Twist: Communist China

Now here’s where it gets interesting.

The next major stop for the textile industry was, of all places, communist China.

Mid-twentieth century China had a whole different story going. No more capitalism! Workers own the means of production! Planned economy! The people are the masters now!

The state ran cotton mills. Workers got lifetime employment, housing, healthcare, retirement—the “iron rice bowl.” This was the opposite of Lowell. There, workers were expendable. Here, workers were supposedly in charge.

Thirty years later, that story changed too.

1978: Reform and Opening. Deng Xiaoping’s new story—socialism with Chinese characteristics, market economy, getting rich is glorious.

Foreign capital flooded in. And what industry landed first? Textiles. Low barriers, labor-intensive, perfect for exploiting China’s massive cheap workforce.

Suddenly there were factories all over the Pearl River Delta. Young women from villages living in dormitories, working twelve-hour shifts, sending money home.

I want you to picture 1830s Lowell next to 1990s Dongguan. Same machines. Same noise. Same young women. Same dormitories. Same hours.

The stories couldn’t be more different. One is “capitalist exploitation.” One is “socialist market economy.” Close your eyes and listen to the looms. You can’t tell which ideology you’re in.

The system doesn’t care about ideology. It wears whatever costume the local players are wearing, as long as they play effectively.

The Cycle Repeats

Then Chinese wages rose.

Labor laws tightened. Workers gained leverage. And the system did exactly what it did in 1845: started looking for cheaper options.

Vietnam. Bangladesh. Cambodia. Ethiopia.

Beijing’s state-owned textile mills closed. The workers got laid off. The land got redeveloped. Now there are glass towers where the cotton mills used to be. The China World Trade Center. The CCTV building. Finance and tech companies.

The people working in those towers probably don’t think about what’s underneath them. They have their own story. Knowledge economy. Creative class. Human capital. Irreplaceable skills.

They’re sitting in seats that textile workers vacated.

The textile workers had stories too. They believed they were the working class, the masters of the nation. Before them, the Lowell mill girls had stories about independence and dignity. Before them, Irish immigrants had stories about new beginnings.

Every generation has stories. Every generation’s stories get overwritten by the next iteration.

What Are Stories For?

Okay, so what’s going on here?

Humans need stories. We need to believe what we’re doing makes sense, fits into something larger. The American Dream. China Rising. Workers of the World Unite. Innovation and Disruption.

Stories motivate us. They comfort us. They give us identity.

Systems don’t need any of that.

A system just needs players. It doesn’t ask why you’re playing or what you believe. It asks: can you do the function efficiently?

Lowell told himself a story about national industry while calculating arbitrage margins. Mill girls told themselves a story about independence while the system saw cheap labor. Communist China told itself a story about eliminating capitalism while becoming capitalism’s most efficient factory floor.

Stories are painkillers. They make it bearable to be a cog. They let you believe you’re the hero of the movie instead of an extra.

But stories can also be tools.

Lowell used the “moral factory” story to acquire workers cheaply. China used the “reform and opening” story to attract foreign investment. Successful players understand the instrumental value of narrative. It moves resources. It reduces friction. It makes people cooperate willingly.

The question is: are you someone who believes stories, or someone who uses them?

People who believe stories are pieces on the board. They commit fully, then feel confused and betrayed when the game changes.

People who use stories are players. They know the story is surface. The game is underneath. Surfaces change. The game continues.

Why We Should Respect This Mechanism

I have a weird respect for capitalism now.

Not moral respect. Morality doesn’t apply here any more than it applies to gravity.

Respect for its persistence.

Think about how many ideologies have tried to kill it. Or contain it. Or replace it. Mercantilism. Colonialism. Communism. Fascism. Nationalism. Globalism. Anti-globalism.

It passed through all of them. Still running.

It ran in Puritan New England. It ran through Catholic Irish immigrants. It ran in atheist China. It runs in Islamic countries. It runs wherever people agree to play.

It’s not really an ideology. It’s more like a discovery.

We discovered gravity. We discovered evolution. And somewhere along the way, we discovered that if you combine private property, contract enforcement, and price signals, you get a system that runs itself. No central authority needed. Works across cultures, religions, languages.

It’s like finding a law of social physics.

You can like it or hate it. You can try to regulate it. But you can’t pretend it doesn’t exist.

The Current Iteration

You know what’s happening right now.

AI.

Every few days someone tells me it’s a bubble. Overvalued. Most companies will fail. Lots of money will evaporate.

They might be right. But they’re missing the point.

Nobody knows what AI is actually for yet. Nobody knows which approaches will work. We’re in unexplored territory.

When the territory is unknown, the system’s strategy is: let lots of people try stuff. Most will fail. The failures get written off. The few successes get identified, copied, scaled.

Wasteful? Sure. But trial and error is how discovery works. You pay for wrong guesses to find right answers.

The moral critiques of AI—job displacement, inequality, concentration of power—might all be accurate. But they assume we get to vote on whether this technology happens.

We don’t.

If there’s efficiency to be gained, if there’s arbitrage to be captured, the system will find a way. Block it here, it emerges there. Restrict it in one industry, it seeps in from another.

Same pattern as always. The Lowell mill girls went on strike. The system routed around them with Irish immigrants. Every attempt to use narrative—moral, political, legal—to stop the system eventually gets bypassed.

The system doesn’t process narrative.

It simply processes efficiency.

Standing by the River

I walked out of the museum and stood by the Merrimack River.

Still flowing. Dam still there. Mills silent.

Those same looms are running in Bangladesh right now.

I thought about the finance people in Beijing’s glass towers. They have their story. Knowledge work. Creativity. Human judgment. Can’t be automated.

They’re sitting where textile workers sat.

The textile workers had stories too.

What’s the difference between them?

Maybe just this: some people know they’re in a game. Some people think they’re living a story.

Knowing doesn’t mean you win. But when things shift, you’re not blindsided. You knew it was possible. You were already thinking about the next move.

Not knowing doesn’t mean you lose. But when things shift, the confusion is worse. You thought the story was real. You thought your position was secure. You were supposed to be the protagonist.

Lowell knew. He told national industry stories while doing arbitrage math. He used narratives to mobilize resources and arithmetic to make decisions.

The people who believed in “moral capitalism” didn’t know. When the system dropped the moral part, they couldn’t understand what had happened.

How many people in today’s knowledge economy know? How many don’t?

What the Loom Taught Us

This game has been running for two hundred years. It’s passed through wars, famines, revolutions, ideological collapses. It doesn’t know your name. It doesn’t remember what you contributed. It won’t make an exception for your story.

It asks one question: are you the most efficient vehicle of the function?

If yes, resources flow to you. If no, you get replaced. Not out of malice. Out of the same logic that makes water flow downhill.

You can refuse to play. But that doesn’t put you outside the game. It just means you’ve given up agency. The game continues with or without you.

You can play while believing your story. That you’re special, essential, irreplaceable. This feels good until the next iteration hits.

Or you can play with your eyes open. Knowing stories are tools, not truths. Positions are temporary. The only durable advantage is the ability to keep adapting.

The Museum’s Ending vs. The Real Ending

The last room of the museum has a plaque. Something about how Lowell’s history teaches us that industrialization brings prosperity and costs, and we should remember those who paid the price.

It’s a nice ending. Appropriate for a museum. You leave feeling thoughtful and slightly superior to the nineteenth century.

But that’s not the lesson.

The lesson is that the loom is still running.

It doesn’t care if you remember the costs. Doesn’t care about your moral reflections. Doesn’t care about your sympathy. It’s running right now, in the next cheap place, with the next batch of workers, wrapped in the next story.

You can be the person who walks through museums feeling wistful about the past.

Or you can be the person who notices the pattern and asks: where is it going next? And how do I position myself?

The loom doesn’t care which you choose.

It just keeps running.

Housing is different. When rents surge, calls for political intervention follow almost immediately. Not because people misunderstand economics — but because unlike toilet paper, they can’t add supply themselves.

In the early 1900s, Boston families built triple-deckers. They lived in one unit, rented the others, and in doing so directly expanded the city’s housing stock. Today, new supply doesn’t come from households. It comes from 400-unit build-to-rent projects, executed by institutions with capital, entitlement expertise, and scale.

When ordinary households feel priced out, they turn to the only lever available: politics.

Cambridge as a Case Study

From the early 1970s to 1994, Cambridge, Massachusetts imposed one of the strictest rent control regimes in the United States. The aim was affordability. The outcome was misaligned incentives that every investor will recognize.

Rent control capped upside while leaving downside intact. Landlords stopped reinvesting. Developers shifted capital elsewhere. Supply froze. And ironically, affordability worsened for the very households rent control was meant to help.

In 1994, voters abolished rent control. Within months, capital flowed back. Buildings were renovated, projects finally penciled again, and the city’s housing stock rebounded. Rents rose, yes — but quality and availability improved after decades of neglect.

For investors, Cambridge is more than a historical anecdote. It is a case study in how policy distortions alter the pro forma — and how reversals can unlock generational returns.

Why Housing Is Structurally Political

The toilet paper analogy is instructive. Consumer goods are elastic and individual. Housing is inelastic and institutional.

That structural difference explains why housing markets attract political pressure in ways other goods do not. Once supply requires institutional capital and long lead times, affordability crises inevitably spill into politics. Pushback — rent caps, zoning fights, ballot measures — isn’t noise. It’s intrinsic.

For institutional residential investors, this means politics must be treated as part of the underwriting model. Just as you stress-test interest rates or construction costs, you must stress-test for political risk.

Lessons for Investors

Cambridge crystallizes three enduring lessons:

1. Policy Risk Is Market Risk

   Rent control didn’t just change tenant-landlord dynamics. It reshaped the risk-return equation. For investors, regulation belongs in the pro forma as a fundamental input.

2. Distortion Creates Scarcity

   When incentives are capped, capital exits. Supply contracts. Scarcity deepens. Policy designed to protect affordability often amplifies the shortage.

3. Reversals Unlock Opportunity

   Cambridge’s post-1994 rebound shows how quickly capital re-prices when distortions unwind. Those positioned early captured generational wealth creation.

The Ultimate Investor’s Question

Housing affordability is once again headline news. Rent control proposals are resurfacing in New York, California, Minnesota, and beyond.

It’s natural to put on the Mr. Monopoly’s hat and call for more supply = good. But for serious institutional investors, the sharper question isn’t whether politics will intrude. It will.

The sharper question is: Where is today’s Cambridge — and how do you position for the reset when policy and economics inevitably converge?

Before Geltner, real estate was largely anecdotal. Prices were “what the broker said,” and risk was “what it felt like.” He didn’t try to make the field trendy. He made it honest. He pulled it closer to the asset pricing literature. He forced people to treat real estate like the financial asset it is—one with cash flows, liquidity constraints, frictions, and opportunity costs.

His biggest contribution wasn’t just a formula. It was a worldview. That you can understand this asset class in the same rigorous way we try to understand stocks or bonds. That real estate isn’t exempt from capital logic just because it’s slow-moving or illiquid or spatially messy.

That view is quietly disappearing.

Real estate academia today focuses on very different questions. It’s easier to publish a paper about zoning policy and social equity than one about cap rate behavior across rate cycles. The field has drifted toward urban studies, planning, and sustainability. There’s nothing wrong with those things. But they’re not the same thing as understanding how this market works.

Capital hasn’t changed in the same way. If anything, it has become more demanding. A pension fund manager still needs to understand how a 25 basis point change in interest rates will flow through to NAVs, debt covenants, and fund-level returns. They still need to know whether cap rates will compress next year, or whether rent growth in secondary markets is a temporary artifact of migration patterns or a durable trend. And they have to make those decisions without the benefit of daily price discovery or high-frequency trading data.

That’s the gap.

What Geltner started was a project to close it. To build a real theoretical and empirical foundation for one of the largest asset classes on earth. Not by borrowing ideas from public markets and slapping them onto buildings, but by starting from first principles: What are the actual frictions here? How do these markets clear? Why do they cycle the way they do?

We still haven’t finished that project.

A Geltnerian is someone who wants to continue it. Not because it’s fashionable, but because it’s necessary. Because someone has to understand why IRR-based compensation drives short-term behavior, or how promote structures create misaligned incentives. Because the way you structure risk in a deal affects the entire system, and the structure is often more important than the asset.

A Geltnerian doesn’t assume real estate is efficient. But they also don’t give up on modeling it. They understand that even partial efficiency has structure. Cap rates don’t move randomly. They absorb expectations about interest rates, liquidity, and long-term growth. Rent isn’t just about location. It’s about supply pipelines, zoning, tenant mix, and how long it takes to build.

You can predict some of these things. You can model others. But only if you take the asset class seriously.

There are still things we don’t understand about this market. Why public and private valuations behave the way they do. Why some asset classes have more mean reversion than others. What the actual term structure of risk looks like when you can’t mark to market.

These aren’t academic questions. They determine how billions of dollars get allocated.

The alternative is to fall back on vibes and trend reports. Or to turn real estate into a narrative exercise where every problem is solved by community engagement or better design. That might win awards. It doesn’t solve for capital.

Geltner never claimed real estate was simple. He just believed it could be understood.

Be Geltnerian.

Deep inside the vessel was the Combat Information Center — the CIC. It didn’t fire weapons or steer the ship. What it did was fuse radar, sonar, weather, satellite feeds, and weapons systems into a single, continuously updating picture of the battlespace.

Naval warfare had always been fought by line of sight. Captains squinted at the horizon, barked orders, and relied on intuition. It was slow, reactive, and often decided by who guessed better under pressure. With the CIC, that changed. Ships stopped reacting and started reasoning. They could model scenarios before contact, simulate outcomes before acting, and coordinate allies in real time.

The result wasn’t a smarter officer. It was a smarter ship.

The Navy institutionalized this insight with the Naval Tactical Data System — a digital backbone for thinking at sea. It was the birth of what we now call cognitive infrastructure: systems designed not just to store or display data, but to process reality — and act accordingly.

Subscribe now

That same transition is now beginning in institutional real estate investment.

From the outside, nothing seems different. Firms raise capital, underwrite assets, evaluate risk. But beneath the surface, the world they operate in has changed irreversibly.

Capital markets shift in real time. Spatial demand moves across geographies and formats. Labor pools swing with political borders. Policy shocks — interest rates, land use, taxation — strike with little warning. Asset performance is no longer a question of physical structure, but of interpretation under uncertainty.

This is no longer a market of stable signals. It’s a terrain of moving parts — and real estate, with all its idiosyncrasies and asymmetries, is the most cognitively demanding of all.

In many ways, managing a real estate portfolio today looks like naval combat once did: limited visibility, long feedback loops, and enormous consequences for delayed judgment.

And yet, most firms still operate like pre-CIC ships. Data is siloed, judgment is manual, decisions are episodic. Intelligence lives in the heads of individuals, rather than in the fabric of the firm. It works, in the way that maps and compasses once worked — until you meet someone with radar.

The future belongs to firms that adopt thinking capital systems — the financial analog to a CIC. Not dashboards, not automation, but reasoning infrastructure: systems that perceive signals, simulate scenarios, and surface decisions before the environment demands them.

These firms will still go to the same conferences and underwrite the same markets. But they’ll quietly exit at the right moment, rotate capital before others even sense pressure, and avoid risks that never appear on anyone else’s radar.

They won’t look different — until it’s too late to catch up.

At some point, someone will ask: How did they see the market turn before we did?

And the answer won’t be a person. It will be a system.

A room inside the ship — one that was already thinking while everyone else was still scanning the horizon.

Cognitive infrastructure will be to tomorrow’s real estate firms what CICs were to naval power: not an upgrade, but a threshold. After it appears, the game doesn’t just change — it splits. One group wins by design. The other doesn’t understand why they’re always reacting.

And the most dangerous position in that world won’t be being wrong.

It’ll be being late.

Subscribe now