What Was the ICO Boom?

In the summer of 2017 a developer launched the Ethereum and put the warning on the front page: the token does nothing, it is worth nothing, and the money goes to the developer. Do not buy it, the site said, more or less. People sent real ether anyway.

That was the ICO boom in miniature: for about eighteen months, from early 2017 into 2018, a PDF and a wallet address could raise nine figures. This piece is history, not advice. The pattern keeps returning under new names, which is the reason to know the original.

How an ICO actually worked

An initial coin offering was crowdfunding with the paperwork deleted. A project wrote a whitepaper, published a wallet address or a smart contract and set a sale window, and most ran discount tiers on top of that, presale first, with early-buyer bonuses shrinking week by week. You sent ether or bitcoin to the address, newly created tokens came back the other way, and no broker or prospectus ever stood between the project and your money.

Almost none of these sales used the word share, and that was no accident, because a token pitched as fuel for a future network or as access to a product is not, on paper, a security. Most tokens were issued on Ethereum, which is one reason the boom and ether's own 2017 rise fed each other.

You can trace the template back to Ethereum itself, whose 2014 crowdsale raised around $18 million in bitcoin and whose promised platform actually got built and shipped, and every sale afterwards leaned on that one success, because if Ethereum could bootstrap from a public token sale then the next whitepaper could as well.

The numbers, loosely

Thousands of ICOs launched through 2017 and early 2018, and the totals are estimates rather than facts, because nobody agreed on what counted as an ICO or which claimed raises were real. Most counts put the take well over $20 billion across 2017 and 2018 all the same.

The landmarks are firmer: Filecoin raised about $257 million in 2017 to build a decentralised storage network, and its sale already looked different from the free-for-all around it, running under a legal wrapper called a SAFT and restricted to accredited investors, an early sign of where things were heading. EOS kept its sale open for a full year across 2017 and 2018 and took in around $4 billion, more than 200 times what Ethereum's own crowdsale had raised, for a platform that did not exist yet, and Telegram raised about $1.7 billion privately in 2018 for its TON blockchain before refunding investors once US SEC action ended the project's original form in 2020.

The froth

Diligence did not keep up, and the clock was part of the design. Basic Attention Token raised about $35 million in under a minute in May 2017, and windows that short leave no time to read anything. Projects raised eight figures on whitepapers with no working code behind them, some of the papers were plagiarised from other projects' papers, and few buyers checked any of it, because checking meant missing the window.

Celebrity promotion turned up too: in 2017 Floyd Mayweather and DJ Khaled both promoted Centra Tech, a crypto payments startup whose founders were later convicted of fraud. Both promoters settled separately with the SEC over not disclosing they had been paid to post, which puts the Useless Ethereum Token in rare company, because it was at least honest about what it was.

The turn

Regulation moved earlier than people remember: in July 2017, with the party still loud, the US SEC published its DAO Report. It said existing securities law can apply to token sales, judged by the Howey test, a standard from 1946 that asks whether buyers are investing in a common enterprise expecting profit from the efforts of others. A large share of 2017's tokens fit that description comfortably. China went further and banned ICOs outright in September 2017, and the US moved case by case instead.

Enforcement built from there through 2018, 2019 and 2020, and Telegram's refund was one of the biggest results it produced. Projects took to blocking US buyers from their sale pages, and Block.one, the company behind EOS, settled unregistered-offering charges in 2019 for $24 million, which worked out at roughly half a percent of what it had raised.

The market did the rest of the work. Most 2017-era tokens lost the large majority of their value through the 2018 bear market, and a long tail turned out to be exit scams from day one, projects that raised, went quiet and deleted their websites. By 2019 new projects avoided the acronym entirely.

What it paid for

Ethereum, the platform most of the boom ran on, was itself crowdsale-funded, and Filecoin spent three years building before its network went live in 2020, so you cannot call the mechanism a pure waste, it paid for real infrastructure and wrapped it in an enormous amount of vapour. Nobody ever cleanly accounted for how much of the $20 billion produced anything at all, and the list of visible survivors is short.

The format died and the pattern survived it, because exchanges repackaged public sales as IEOs, where the exchange did the listing and some of the vetting, and Europe eventually wrote MiCA, its rulebook for how crypto assets get offered and sold in the EU.

Seeing it coming next time

The durable skill from the whole episode is document-level scepticism: does working code exist, who holds the token supply, and what, in writing, does the token entitle you to. Those three questions would have filtered most of the 2017 field, and they are the same checks our guide to reading a whitepaper walks through, because that is the era they were learned in.

The next version of this will not call itself an ICO, because that name is burnt. Expect a new acronym, a cleaner website and the same mechanism underneath. Having seen it once, you get the second look at a discount.