2019. The Jacobson brothers launch RealT and promise to democratise real estate through the blockchain. The pitch is simple: buy houses in Detroit, split them into tokens, sell them all over the world. One token = a fraction of a property. Rents paid out automatically. Instant liquidity. Every retail investor’s dream.
Why Detroit? Jean-Marc Jacobson explains:
It was, he says, the best time and the best place to get both yield and capital growth, which is rare in real estate: usually, you get one or the other.
The technical model: one American LLC per property. The ERC-20 tokens represent shares in that company, not chunks of wall24.
Why an LLC? It is almost the only structure that makes tokenisation simply feasible in the United States.
In France, attempts rely on even more convoluted structures. Crowdfunding platforms issuing perpetual bonds, like Tantiem. Atoa comes close to genuine tokenisation through a fiducie, the French equivalent of a trust. But these are only workarounds.
Besides, frankly: who cares.
The real value proposition here is cutting out intermediaries, offering more transparency, and letting you build your own portfolio by picking your own assets. The technical feat only interests the people who want to pull it off.25.
Jean-Marc Jacobson walks me through their philosophy:
Their business model, he says, was built so that they did not do the property management themselves. That was the foundation of everything.
Management? Systematically outsourced.
They had property managers, he says, well-known firms with a proper storefront.
RealT raises, tokenises, distributes. The managers take care of the rest. Or were supposed to.
In six years, RealT raises $160M. 2,000 homes tokenised. Investors all over the world buying fractions of properties they will never see. The startup takes 10% on acquisition and 2% of rents. A flawless business model, on paper.
On paper, this digitised bricks and mortar is just one more step towards ultra-financialisation. In France, the financialisation of property probably began in 1964 with the first SCPI, the French retail property fund, created shortly after REITs were invented in the United States. A system that lets you distribute 90% of property income without paying corporation tax26, and above all with surprising stability.
2008? SCPIs become a safe haven. Their capitalisation goes from €11bn in 2000 to €100bn in 2022. The ultimate paradox: the financial crisis makes them more attractive. While CDOs collapse and Lehman Brothers disappears, French savers discover that their SCPIs keep paying out rent.
The history of REITs, by contrast, is a festival of crashes and resurrections. 1974: -50%27. Developers had overbuilt, rates were climbing, petrol was rationed. Apocalypse. 2008: -67%, with General Growth Properties28, America’s second-largest shopping mall owner, going bankrupt. 2020: -42% in three weeks when Covid hits29.
As for the more recent period, since rates went up, it has not really called the model itself into question. But it has exposed the (serious) management mistakes of those who bought a lot, and expensively, to deploy huge inflows of savings. On top of showing that themed vehicles were structurally flawed.
And that is exactly when the tokenisation of anything and everything exploded: we even saw cars and Pokémon cards, at companies that have since vanished. It does make certain investments accessible, and lets you build a portfolio yourself without going through a managed fund. But fees are generally higher, and the promised liquidity (very relative: you still need a buyer) is a first problem. One of the big advantages of real estate is precisely its stability, which comes from the fact that assets react to trends over the medium term and to their environment over the long term. But very little in the short term.
Listing a token means giving these assets a real-time, volatile price.
The other major problem, obviously, is the risk of stuffing thingamajigs into whatsits, and whatsits into doohickeys, until nobody really knows what corresponds to what. Like Robinhood, which says it tokenises shares but in reality tokenises derivatives of shares. Or Primonial, a French asset manager, whose SCPIs invest in its in-house OPCIs (another type of French property fund), and vice versa, or park millions of euros in shareholder current accounts. All of this is explained in hundreds of pages of annual reports, perfectly unreadable for the average retail investor.
Even though we now know perfectly well that this kind of mess was one of the main causes of the 2008 crisis. The report of the FCIC30 on the subprime crisis says so plainly31.
The financial engineering behind these investments, the commission writes, made them harder to understand and to value than the individual loans.
The growing complexity of housing-related assets and the many steps between the borrower and the final investor, it adds, increased the importance of the credit rating agencies and made it harder for investors to assess the risks independently.
I wanted to try it for myself. €50 for a token in 1610 State Fair, an apartment block in the north of Detroit. The photos on the website looked fine. Google Street View showed a building standing up, windows intact. Promised yield: 14.47% net.
The maths was simple: $608,751 raised for a property bought for $520,000. Estimated rents: $129,000 a year, or $672 per apartment. After costs, my token was supposed to bring in $7.28 a year. A dream return on €50 invested.
Three years later, I had received $20.04. That is an 11.6% yield, 20% less than promised. Not a disaster. Quite the opposite. Except that in December 2024, a reader warned me about the actual state of the building. I went back to Google Street View, the recent imagery this time.
And no, it is not just the light or the season.
The windows have been replaced with cardboard. The front door gapes open. Graffiti everywhere. The rusted exterior stairs look ready to collapse. What I found next was worse: the City of Detroit has listed the building as a “dangerous building” since January 2025, yet my rent kept landing in my wallet. $20 while someone was living in those ruins.
Except that my 1610 State Fair is not an exception. It is the model.
RealT’s properties are not scattered at random. Lay the map of its holdings over the map of poverty in Detroit: they match perfectly. Half of the investments are concentrated in the north-west, in the most deprived neighbourhoods of the city.
Overlay of the map of properties owned by RealT (the circles), provided by the City of Detroit, and 2009-2013 poverty data from the city’s open data portal, mapped by Alex B. Hill32
Petoskey-Otsego: crime six times the state average. One resident in eight belongs to a gang. RealT owns six properties there. Yorkshire Woods, Denby, Mapleridge, Moross-Morang: so many forgotten names where RealT keeps piling up properties.
Jean-Marc Jacobson defends that choice.
They decided, he says, to grow in concentric circles in Detroit because they thought it was the best moment: it was one of the cities best placed to deliver both yield and capital growth, which is rare in real estate. Usually, it is one or the other.
Brightmoor accounts for a share of the investments. Built in 1921 to house factory workers, the neighbourhood has become a symbol of urban decay. Residents avoid the petrol stations (too dangerous). Businesses barricade their car parks. Visitors get survival tips: never stop with less than half a tank, remove anything valuable, fit a steering-wheel lock. RealT owns dozens of properties there.
The other big cluster sits around Eight Mile, the boundary Eminem made famous. At his childhood home, 19946 Dresden, there has been nothing left for decades. Around it, the neighbourhood hasn’t changed: stacked tyres serve as benches, rubbish piles up. This is where Karnail “Bugz” Pitts was murdered in a park after an argument in 1999, a few hours before he was due on stage with D12, Eminem’s original crew. Where DeShaun “Proof” Holton, another D12 member, was murdered outside the CCC Club at 15308 8 Mile Road, a venue now bricked up like so many others around it.
Between Evergreen and Telegraph, north of Highway 96, stretch the most rundown blocks. That is precisely where RealT invests heavily. In streets where the poverty rate exceeds 60%, where 81% of the population is Black, where public services left long ago.
Overlay of the map of properties owned by RealT, provided by the City of Detroit, and crime data33
The strategy is clear: buy in poverty, rent to poverty.
On screen, a double-digit yield. On the ground, an apartment block listed as a “dangerous building”. And we investors click “buy” from our sofas in Paris or London. We collect our dividends without ever seeing the cardboard in the windows. The abstraction is perfect: Detroit is no longer a city, it is an Excel spreadsheet.