An ICO is a token sale. A project writes a whitepaper, deploys a smart contract that mints a new ERC-20 token, opens the contract to buyers who send in ETH or BTC in exchange for tokens at a pre-set ratio, and then uses the raised funds to build whatever they said they were going to build. The structure is conceptually similar to an IPO, except that what you receive is a freshly-invented token rather than equity, the terms are whatever the project feels like offering, and there is usually no regulatory oversight. The term was coined around 2013 by analogy to IPO, and the format exploded in 2017 when the combination of Ethereum’s ERC-20 standard and a rising market created both the technical infrastructure and the speculative appetite.
The numbers from 2017-2018 are genuinely staggering. Filecoin raised $257 million in an hour. Tezos raised $232 million. EOS ran a year-long sale that brought in over $4 billion. Over the course of 2017 and 2018, ICOs collectively raised something like $30 billion, most of it from retail investors who had no way to evaluate whether the projects they were funding had any chance of delivering anything.
Why They Worked, Briefly
A combination of factors made ICOs a natural vehicle for that particular moment. Ethereum made token issuance trivially easy — you could deploy a working ERC-20 contract in an afternoon. There was no meaningful regulation in most jurisdictions, so projects could sell tokens without needing to do the legal work of a real securities offering. The market was rising, which meant that tokens bought in the sale typically listed on exchanges at a premium and early participants made quick profits. Those profits attracted more buyers, which pushed prices higher, which attracted more projects running sales, which pulled in more buyers. The loop was self-reinforcing for about eighteen months.
The quality of the projects being funded was mostly terrible. A standard ICO pitch in 2017 was a whitepaper, a slick website, a few stock-photo team members (sometimes fake), a roadmap full of vague milestones, and a promise that the token would be used for “governance” or “payments” or “access” in some future product that did not yet exist. Vetting was minimal. Diligence was usually someone reading the whitepaper and a Telegram group’s enthusiasm level. Anyone who had participated in a traditional startup fundraise would have been horrified at the information asymmetry, but the buyers mostly did not come from that world.
The Crash and the Regulatory Hangover
Most ICOs from that era have failed. Studies tracking the 2017 cohort have found that something like 80 percent of them were either scams, abandoned, or technical failures within two years. The tokens, even for projects that technically delivered, mostly trade at fractions of their ICO price. Prices peaked in early 2018 and then fell through the rest of that year as the mania unwound.
The regulatory response was swift and ongoing. The SEC began classifying many ICO tokens as unregistered securities and suing the projects that issued them. Cases against Telegram (TON), Kik (Kin), and Ripple (XRP) set precedents that made it legally dangerous for US-domiciled projects to run traditional ICOs, and the format largely disappeared from the US market. It was replaced by IEOs (exchange-run sales), IDOs (DEX-run sales), airdrops, and whatever the current workaround for securities law happens to be. The spirit is the same — sell newly-minted tokens to retail — but the legal packaging changes every cycle to stay ahead of enforcement.
The ICO era left a permanent mark on crypto’s reputation. For years afterwards, “crypto” and “scam” were more or less synonymous in mainstream coverage, and a lot of that association traces back to the specific damage done by 2017 token sales.