Every big Tor market that died, and how it died
Every serious market operator says theirs is different. Most of the time they mean it. Almost none of them stayed different for long.
Silk Road, 2011 to 2013
The one that started the whole thing. Ross Ulbricht built it around a strong pitch, drugs bought and sold by consenting adults with no violence in the chain, and for two years it worked. What killed it was Ulbricht himself. He asked technical questions on Stack Overflow using an email address that had his real name, he ran the server admin from cafes without hiding the login times, and the FBI eventually walked into a San Francisco library and grabbed his laptop with the market unlocked on screen.
The lesson people took at the time was operational security. The actual lesson is that centralized market operators are one person, and one person makes mistakes over three years. Someone will find one of them.
Silk Road 2, 2013 to 2014
Same brand, different operator, ran for about a year. Was caught partly through an FBI operation that used a Tor deanonymization method Carnegie Mellon researchers had built for another study. The whole thing raised uncomfortable questions about who else was quietly funding Tor attacks. Nobody answered them.
Evolution, 2014 to 2015
The first big exit scam. Two operators, months of building trust, and then one weekend in March 2015 the two of them drained every wallet on the site (around 12 million dollars in Bitcoin at the time) and vanished. This was a defining moment. Every conversation about escrow, multisig and market trust that came after was shaped by Evolution.
AlphaBay, 2015 to 2017
Ran by Alexandre Cazes out of Bangkok, grew into the biggest market ever built by an order of magnitude, and died in July 2017 when Cazes was arrested. He had used his personal email in the welcome message of a forum ten years earlier, which linked back through several accounts to the market servers. He was found dead in a Thai jail cell a week after his arrest. The circumstances were officially suicide. Nobody who watched the case closely accepted that at face value.
AlphaBay users who fled after the seizure got funneled straight into Hansa. Which brings us to Hansa.
Hansa, 2015 to 2017
Hansa was already smaller than AlphaBay but still a serious market. The Dutch police had quietly taken over the Hansa servers in June 2017, weeks before AlphaBay fell. So the operators of Hansa were sitting in a police station while every AlphaBay refugee typed their credentials into a police-run site, for another month. When the operation was announced, buyers realized the mistake and the market shut down. This was probably the single most effective sting law enforcement has ever run against Tor.
Dream Market, 2013 to 2019
The one that quietly lasted longest. Six years is an eternity on Tor. It shut itself down in April 2019 with a farewell message and promised the accounts would move to a new market that never really materialized. There has never been an official explanation. Best current guess is that the operator saw the writing on the wall and left with what was left. Compared to Evolution and AlphaBay, Dream's exit was almost polite.
Empire, 2018 to 2020
Rose fast after Dream died, ran on a fork of AlphaBay's code, and exited in August 2020 with all deposits. Estimated haul was around 30 million dollars. The community response by that point was tired resignation. Empire's exit is when people started taking multisig seriously.
Hydra, 2015 to 2022
The Russian language market that ran for seven years and was, by revenue, bigger than anything before it. Not just drugs, also cashout services, forged documents and stolen data. Died in April 2022 when German police seized the servers and the US Treasury sanctioned everyone connected to it. Unlike most others, Hydra had built out a whole payments and logistics stack that could not survive one takedown.
After Hydra
The market landscape splintered. No single successor market ever reached the same scale. Instead there are eight to ten mid-sized shops running at any given time, each with their own thing, and this is roughly where we still are in 2026. The current markets learned from every death above. Multisig is the norm, PGP signed rotations are the norm, DDoS queues are the norm.
The pattern
Nobody died of technology. They died of operational security, greed, or law enforcement running a long game. The technology under Tor markets did not fail once. It was always the humans running them.
Which is why buyers should assume every market they use will end up in one of these paragraphs eventually. The question is only how soon.