History of Forks: Chapter 2
In 2016, an attacker drained the Decentralized Autonomous Organization (The DAO) of roughly 3.6 million ETH. Ethereum's core team executed a hard fork to reverse the theft. A minority kept mining the original, unaltered chain on principle, and it became Ethereum Classic. This is the only fork in the series where the side that changed the protocol won decisively.
When Ethereum launched in July 2015, it arrived with a radically broader ambition than Bitcoin. It didn't just want to be digital money; it wanted to be a "world computer."
While Bitcoin was built as a highly secure digital ledger, Vitalik Buterin's 2013 whitepaper proposed something closer to a decentralized computing platform. Ethereum featured a built-in programming language capable of running arbitrary computations. The goal wasn't just to transfer value, but to build an entirely new foundation for applications where anyone could deploy custom tokens, decentralized insurance, and financial contracts directly onto the blockchain.
Much of the early community rallied around a simple creed: code is law. These applications, called smart contracts, would execute automatically across thousands of nodes. Ethereum promised applications resistant to downtime, censorship, and third-party interference.
That uncompromising promise was about to face its first catastrophic test.
By spring 2016, a project called The DAO emerged as a showcase for the world computer idea. Designed as a decentralized venture capital fund, it allowed anyone to send ETH in exchange for voting rights over future project funding. Its crowdsale raised roughly 11.5 million ETH, worth about $150 million at the time.
The scale was extraordinary. Roughly 14–15% of all ETH in circulation was now locked inside a single, experimental smart contract.
Then, on June 17, 2016, an attacker exploited a re-entrancy vulnerability in The DAO's code. In simple terms, the attacker could trigger another withdrawal before the contract updated its balance, repeating the process until approximately 3.6 million ETH, worth around $50–60 million, had been moved into a child DAO.
The Ethereum protocol itself hadn't been hacked. The attacker exploited a flaw in The DAO's code, and the blockchain kept processing transactions exactly as designed. The incident exposed a fundamental tension: if smart contracts were supposed to execute exactly as written, what would happen when the code produced an outcome the community considered unacceptable?
The attacker still faced one obstacle. The DAO's rules gave Ethereum a window of roughly four weeks.
Suddenly, a ticking clock hung over the entire Ethereum ecosystem. Panic spread, ETH's price fell sharply, and developers, miners, investors, and exchanges were thrown into frantic debate.
As the countdown approached zero, Ethereum faced an existential question:
If "code is law," what happens when the law allows a hacker to walk away with a fortune, and the network's credibility with it?
The DAO's code gave Ethereum four weeks. Any ETH moved into a child DAO was locked for 28 days, so the attacker couldn't touch the 3.6 million ETH right away.
Vitalik Buterin and other core developers proposed a soft fork that would blacklist the attacker's addresses. The funds would remain frozen without changing Ethereum's history.
Researchers, including Cornell's Emin Gün Sirer, found a problem: enforcing the blacklist meant miners had to check every transaction for links to the frozen funds. Because rejected transactions pay no fees, an attacker could flood the network with them at no cost, forcing miners to waste resources on spam and potentially bringing Ethereum to a halt.
With the soft fork abandoned and the clock running down, Ethereum took the more radical option: a hard fork. Instead of freezing the stolen ETH, the new rules moved it, along with the rest of The DAO's funds, into a recovery contract, allowing the original DAO investors to withdraw their money.
But a hard fork requires more than code. It requires people to run it.
Most developers, exchanges, miners, and users upgraded. The new chain recovered the funds and continued as Ethereum.
A minority refused. They believed no amount of money justified rewriting the ledger. They kept mining the original chain, which survived as Ethereum Classic.
The DAO fork left Ethereum with two competing answers to the same question: should a decentralized network preserve its rules, or preserve the purpose those rules were meant to serve?
Five Twitter (X) posts from August 2016 through October 2017, showing how Ethereum Classic supporters argued that an immutable chain would outlast the one that reversed The DAO hack.
Monthly figures since ETC's fork (July 2016) through September 2026. Price and fees come from CoinMetrics' public dataset; developer activity is a first-party count of monthly unique commit authors across each chain's GitHub ecosystem (per Electric Capital's open taxonomy of which repos belong to which chain).
Average daily close per month, log-scaled so both chains' full range stays readable on one chart.
Total network fees paid each month, converted to USD day-by-day (native fee × that day's price, summed) rather than applying one month-end price to the whole month.
Distinct GitHub accounts committing to each chain's tagged repositories that month. ETH is capped to its 300 most-starred repos (its full ecosystem is far too large to crawl in full); ETC's much smaller repo list is used in full, uncapped.
From Ethereum's own launch through The DAO, the hard fork, and Ethereum Classic's first exchange listings. Events are evenly spaced in date order, not to scale in time. The date under each point gives the actual timing.