I thought it was a joke the first time someone told me they had spent forty thousand dollars on a piece of land that didn’t exist. A friend of a friend had purchased a package close to a Decentraland virtual nightclub in late 2021. He discussed it in the same manner that people used to discuss Miami waterfront real estate. View, neighbors, and location. In this instance, the view was rendered. The neighbors were avatars who only ever logged in once or twice before disappearing forever.
It seems like that conversation took place in a different decade. Despite all of its noise, the metaverse land rush has subsided in a way that very few anticipated. No single crash occurred, nor was there a dramatic morning when the floor collapsed. It simply became thinner. Listings ceased to move. There were no bids at the end of auctions. The kind of gradual erosion that eventually destroys a town but does not garner much attention.
| Subject | Detail |
|---|---|
| Topic | Collapse of metaverse virtual land markets |
| Peak Boom Period | 2021–2022 |
| Major Platforms Affected | Decentraland, The Sandbox, Otherside |
| Sandbox 2021 Transactions | Around 65,000 deals worth $350 million |
| Reported Top Plot Prices | Over $200,000 per parcel at peak |
| Brands That Entered Early | Gucci, Snoop Dogg, various luxury and crypto firms |
| Primary Failure Modes | Low foot traffic, clunky access, weak community |
| Where Users Actually Went | Roblox, Fortnite, private AR experiences |
| Recent Funding Signal | Somnia Network’s $270M Series B (Oct 2025) |
| Current Outlook | Infrastructure phase replacing speculation, per IMF metaverse research |
A portion of the story is told by numbers. In 2021, the Sandbox alone handled about 65,000 land transactions worth about $350 million. Many parcels that once exchanged hands for the price of a Brooklyn apartment now sit unsold at any price, and secondary sales on the majority of these platforms had fallen to a fraction of their peak by 2024. Those who closely observed this seem to believe that the speculation was never truly about the land. It had to do with having faith that someone else would have faith.
Looking back, I’m struck by how convincing the optics were. Gucci launched an online store. Snoop Dogg hosted promotional events and constructed a mansion. JPMorgan cut a press release as they entered a virtual lounge. The sense of inevitability that large corporations create when they want a trend to be genuine permeated every announcement. However, the one metric that truly mattered—foot traffic—never appeared. It was like strolling through a partially constructed mall after it had closed in 2022. The lights are on. No one was present.
There was a mechanical component to the issue. A cryptocurrency wallet, a good graphics card, and tolerance for cumbersome controls that felt at least a generation behind anything on a console were necessary for logging into Decentraland. The avatars appeared to be remnants of an MMO from 2008. The loading times were prolonged. There were audio glitches. The discrepancy between pitch and product was too great for consumers who came hoping for something similar to what the marketing videos advertised.
However, it was people, not technology, that was the deeper miscalculation. The metaverse was marketed as the internet’s next layer, but the internet itself had already imparted a valuable lesson: the value of real estate, whether digital or not, is determined by the local community. The crowd, not the pavement, is what makes Times Square so valuable. A virtual plot adjacent to an empty plaza is merely a coordinate. The early purchasers may have recognized this and anticipated large crowds. They didn’t.
The real online crowds, on the other hand, had always been elsewhere. More people use Roblox every day than the population of most nations. Tens of millions of people attend Fortnite concerts. These platforms may have been successful because they were never marketed as metaverses. They began by playing games with friends and finding reasons to return. The land issue was raised later, if at all.
Instead of pulling back, investors appear to be subtly recalibrating. Improbable and MSquared co-led Somnia Network’s $270 million Series B in October 2025, indicating that funding is still going into the underlying infrastructure rather than the actual parcels. Land ownership has given way to rail ownership in the narrative. The pattern is recognizable. Any technology’s initial wave eventually rewards plumbers and speculators.
What’s left is more difficult to classify. Cyber risk in virtual environments continues to be a topic of discussion for insurance actuaries. For training, some companies have switched to private corporate metaverses, which do produce quantifiable results. A small number of sincere believers carry on building, frequently with less money and greater attention than they did at the height. It’s still unclear if any of this leads to a long-term solution.
It’s difficult to ignore how fast the conversation progressed. Early in 2023, the same venture capital firms that had been making virtual land pitches were now making generative AI pitches, frequently to the same audiences and occasionally in the same decks with a few slides changed. The pixels are still present. Most of the owners are not. Furthermore, if there is a lesson, it predates the technology that created it. Without demand, scarcity is just empty space with a cost.
