The ICO Boom and Bust of 2017-2018

Key Takeaways

• The legal theory arrived a year before the enforcement did. The SEC's DAO Report on 25 July 2017 stated plainly that a token could be a security. ICO issuance kept climbing for roughly another twelve months, because a report is not a court order and nobody was yet paying a price for ignoring it.
• Money moved through a currency that was itself collapsing. Most ICOs were priced and funded in Ether, which fell 94.0 percent from its January 2018 peak to its December 2018 low on the Coinbase series published by the Federal Reserve Bank of St. Louis. A project that spent nothing improperly could still watch its runway disappear in dollar terms.
• Enforcement escalated in publicly visible steps, not one sweep. Munchee returned money before delivering tokens and paid no penalty in December 2017. Airfox and Paragon paid $250,000 each in November 2018 for registration violations alone. Kik fought a $100 million case to a court loss and a $5 million penalty in 2020. Telegram had to return $1.2 billion and abandon its network outright by mid-2020.
• Fraud and mere non-registration were different problems wearing the same wrapper. Centra Tech invented executives and a bank partnership that never existed; Airfox and Paragon built real products the SEC never accused of fraud but had still sold as unregistered securities. Treating every ICO enforcement action as proof of a scam flattens a distinction the SEC itself preserved case by case.
• Celebrity promotion was frequently a paid transaction with no disclosure, not an independent endorsement. Floyd Mayweather and DJ Khaled were the SEC's first touting cases in November 2018, both tied to undisclosed Centra Tech payments.
• The damage stayed inside the crypto sector. Individual issuers paid multimillion-dollar penalties and, in Telegram's case, gave up more than a billion dollars, but no bank, pension fund or payment system was exposed to any of it, a containment pattern this library's other crypto case studies show again later.

What Was an ICO, and How Did the Boom Start in 2017?

An initial coin offering was, mechanically, a smart contract. A project published a whitepaper describing a network, application or business it intended to build, then deployed a contract on Ethereum that accepted Ether from anyone with a wallet and returned a newly minted token in proportion to what was sent. The token was pitched as future access to whatever the project built, a share of its success, or both at once, depending on how carefully the promoters wanted to avoid the word "investment." No bank, broker-dealer or transfer agent stood between the buyer and the contract. No prospectus, audited financial statement or escrow requirement stood between the buyer and the promoters' promises either.

Ethereum's ERC-20 token standard, finalized in late 2015, is what made this cheap. Before it, creating a new tradable digital asset meant writing and securing an entire blockchain. After it, minting a token that any Ethereum wallet or exchange could already handle took a template contract and a few hours of a developer's time. The fixed cost of launching a fundraising instrument fell from something close to a startup's entire seed round to something closer to a rounding error, and the number of projects able to attempt one rose accordingly.

The other ingredient was price. Bitcoin and Ether both appreciated sharply through 2017, which did two things at once: it created a large population of holders sitting on unrealized gains who were comfortable rotating a slice of that gain into a new, riskier token, and it gave ICO issuers a currency, Ether, whose own price was rising, so a raise announced in Ether terms could be worth substantially more in dollars by the time it closed than it was worth when it opened. Both effects ran in reverse just as forcefully once prices turned, which is a mechanism this page returns to directly rather than treating as background color.

None of this required fraud to work as a description of the boom. A whitepaper, a smart contract, a rising currency to denominate the raise in, and a global pool of buyers with no securities-law gatekeeper standing between them and the contract, that combination alone explains why thousands of ICOs launched in 2017 and 2018. Fraud explains why a meaningful fraction of them were worthless from the start, which is a separate question this page also answers, but the boom itself needed no bad actors to get going.

What Did the SEC's DAO Report Actually Conclude?

The report grew out of an earlier failure that had nothing to do with securities law on its face. The DAO was a decentralized venture fund built on Ethereum in 2016, designed to let token holders vote on which projects it would fund. In June 2016 an attacker exploited a flaw in its code and drained roughly a third of its holdings before the Ethereum community intervened with a contested hard fork to reverse the theft. The SEC's Division of Enforcement opened an inquiry into whether The DAO, its German corporate sponsor Slock.it, Slock.it's founders, and related intermediaries had violated securities law by selling unregistered DAO tokens in exchange for Ether.

Statements and findings as recorded in SEC Release No. 2017-131, published 25 July 2017.

The report the SEC issued on 25 July 2017 answered a much bigger question than the hack itself raised. It found that DAO tokens were securities under the Howey test, that issuers of blockchain-based securities must register their offers and sales unless a valid exemption applies, and that trading platforms listing such securities must themselves register as exchanges unless exempt. Then, in the sentence that shaped the next eighteen months more than any other part of the document, the Commission said it had decided not to bring charges against The DAO, Slock.it or their founders "in light of the facts and circumstances," choosing to caution the market rather than punish this particular case.

SEC Chairman Jay Clayton framed the report as a balance: "We seek to foster innovative and beneficial ways to raise capital, while ensuring, first and foremost, that investors and our markets are protected." William Hinman, then Director of the Division of Corporation Finance, put the compliance point more bluntly: sponsors of offerings conducted through distributed ledger technology "must comply with the securities laws." Neither statement was ambiguous about the legal conclusion. Neither statement stopped a single ICO already underway, and the market's dollar volume, to the extent later trackers can reconstruct it, continued climbing for most of the following year.

That gap between a clear legal conclusion and an unchanged market is the single most important structural fact about this event, and it recurs throughout this page: regulatory clarity is not the same instrument as regulatory enforcement, and a market built on the assumption that nobody would test the difference kept operating on that assumption until individual issuers started losing cases.

How Does the Howey Test Turn a Token Sale Into a Security?

The Howey test comes from a 1946 Supreme Court case about orange groves, not blockchains, which is precisely why the SEC leaned on it: the test asks about economic substance, not about the technology or vocabulary wrapped around a transaction. SEC v. Howey involved a Florida hotel operator selling parcels of a citrus grove to guests, bundled with a service contract under which the seller's own company tended and harvested the fruit. Buyers were not farming anything themselves; they were passively expecting a return generated by someone else's work. The Supreme Court held that arrangement was an investment contract, and therefore a security, regardless of the real estate label on the paperwork, because "form was disregarded for substance."

The test the Court articulated has four elements: an investment of money, in a common enterprise, with a reasonable expectation of profit, derived from the entrepreneurial or managerial efforts of others. Apply that to a typical 2017 ICO and three of the four elements are close to automatic. Money moved, usually Ether. A common enterprise existed, the project itself. Profit was expected, and often explicitly marketed, since promoters routinely described how demand for the token would rise once the network launched, exchanges listed it and adoption grew.

The fourth element, effort by others, is where William Hinman's June 2018 speech "Digital Asset Transactions: When Howey Met Gary (Plastic)" did the most work to clarify the SEC's thinking. Hinman argued that a token sold when a network's success still depended on a central promoter's ongoing effort looks like a security, because "the purchaser usually has no choice but to rely on the efforts of the promoter to build the network and make the enterprise a success." A token sold once no central enterprise remained, and the token was purchased only to use the good or service the network already provided, could be a different transaction entirely. Hinman drew the comparison to Howey directly: interests in the citrus grove were sold to hotel guests, not farmers, and ICO tokens were typically marketed to a wide investing public, not to the people likely to actually use the network.

What that framework did not supply was a bright line any issuer could check against before selling a single token. "Sufficiently decentralized" is a judgment call, not a formula, and Hinman's speech was a staff view at a conference, not a rule with the force of law. Projects spent real effort in 2017 and 2018 trying to word their whitepapers around the fourth element, emphasizing a token's "utility" rather than its return, a pattern Chairman Clayton addressed directly in his December 2017 statement: "Merely calling a token a 'utility' token or structuring it to provide some utility does not prevent the token from being a security. Tokens and offerings that incorporate features and marketing efforts that emphasize the potential for profits based on the entrepreneurial or managerial efforts of others continue to contain the hallmarks of a security under U.S. law." The label changed. The Howey analysis, in the SEC's own account of it, mostly did not.

How Big Did the ICO Market Get Before It Broke?

No regulator or statistical agency ever published an audited total for global ICO fundraising in 2017 and 2018, and this page does not manufacture one from private trackers. What can be stated with precision is the size of individual deals, because those figures come from the issuers' own SEC settlements or the SEC's own complaints rather than from a market aggregator, and a small number of them were enormous.

The aggregate context comes from a different kind of source. The Financial Stability Board's October 2018 report on crypto-asset markets found that the combined market capitalization of all crypto-assets, not ICO fundraising specifically but the broader market ICOs both drew capital from and paid into, peaked at an estimated $830 billion on 8 January 2018, of which roughly 35 percent was attributable to bitcoin alone, before falling to approximately $210 billion by 4 October 2018. That figure describes the pool of speculative capital that ICOs were competing for and drawing on, not the ICOs themselves, and the two should not be conflated; a reader who wants the aggregate ICO total specifically will not find a verified one on this page, because none exists in a form this page's sourcing standard accepts.

The Congressional Research Service's October 2018 briefing adds one more scale figure worth keeping distinct from dollar totals: more than 2,000 digital assets, in the form of cryptocurrencies and crypto tokens, existed by the time of its writing, the large majority of them created through an ICO of some kind. Most of those 2,000 raised nowhere near what EOS or Telegram did. The distribution was heavily skewed, with a handful of mega-raises and a long tail of small ones, which is the shape scale tables should be read against rather than as a claim that the typical ICO looked anything like Block.one's.

What Did a Fraudulent ICO Look Like in Practice?

Centra Tech is the case the SEC itself chose to bring as a straightforward fraud, not merely an unregistered offering, and its complaint is worth reading closely because it shows exactly which claims were fabricated rather than merely aggressive. Sohrab "Sam" Sharma and Robert Farkas raised more than $32 million from thousands of investors in 2017 by selling a "CTR Token" tied to a promised suite of financial products, centered on a debit card that would let users spend cryptocurrency anywhere Visa or Mastercard was accepted.

According to the SEC's complaint, filed 2 April 2018, none of the load-bearing claims behind that pitch were true. Centra had no relationship with Visa or Mastercard, contrary to what the offering materials represented. The company's marketing featured executives with fabricated professional biographies who did not exist as described. And to move the token, Sharma and Farkas paid celebrities to promote the ICO on social media without the audience necessarily knowing the endorsement was paid, a detail this page returns to in the next section because the SEC treated it as a distinct violation on its own. Farkas was arrested attempting to board a flight out of the country; both founders were separately criminally charged and later pleaded guilty.

What makes Centra Tech useful as a contrast rather than just an example is what the SEC did not allege against other ICO issuers it also charged that same year. Airfox and Paragon, whose settlements are covered later in this page, built the products they said they would build to varying degrees and were never accused of fabricating a partnership or a management team. Their violation was purely regulatory, selling unregistered securities, and the SEC's own November 2018 press release drew the distinction explicitly, describing their cases as "the Commission's first cases imposing civil penalties solely for ICO securities offering registration violations." Reading every 2017-2018 ICO enforcement action as proof of a scam erases a line the SEC itself was careful to keep visible case by case: a company that sold an unregistered security and a company that invented a bank partnership committed different offenses, and the appropriate lesson from each is not interchangeable.

Why Were Celebrities Paid to Promote Coins, and What Happened to Them?

A paid celebrity post reads, to a follower scrolling past it, as an opinion. The SEC's position, formalized in its first ICO touting cases on 29 November 2018, is that when the celebrity was paid to post it, the post is an advertisement, and federal securities law has required advertisements for securities to disclose that payment since long before crypto existed. Section 17(b) of the Securities Act makes it unlawful to publicize a security for compensation without disclosing the nature, source and amount of that compensation, and it applies regardless of the medium.

Boxer Floyd Mayweather Jr. and producer DJ Khaled were the SEC's first targets under that theory specifically in the ICO context. Both had promoted Centra Tech's token; the connection between this case and the fraud case against Centra's founders is direct, not coincidental. The SEC's order found that Mayweather failed to disclose $100,000 in payment from Centra Tech alone, plus $200,000 more from two other ICO issuers, while telling his Twitter followers the Centra sale "starts in a few hours. Get yours before they sell out, I got mine" and posting that he should be called "Floyd Crypto Mayweather from now on." Khaled failed to disclose a $50,000 Centra Tech payment behind a post calling the token a "Game changer."

Neither settlement alleged that Mayweather or Khaled knew Centra Tech's underlying claims were fabricated. The violation was narrower and, for an ordinary reader trying to draw a lesson from it, arguably more useful: an endorsement's credibility depends on knowing whether it was purchased, and an audience with no way to find that out cannot price the endorsement correctly. As Enforcement Co-Director Stephanie Avakian put it in the SEC's release, "with no disclosure about the payments, Mayweather and Khaled's ICO promotions may have appeared to be unbiased, rather than paid endorsements." A famous name attached to a token said nothing about the token's merit and, in this instance, said even less about whether the endorsement itself was genuine.

Which ICO Warning Signs Were Visible at the Time, and Which Weren't?

The temptation with any collapsed asset class is to treat every later-revealed problem as though it had been legible from the start. Some of what unraveled ICOs was genuinely public before an investor bought in; some of it required a subpoena.

Classification by Swoopr Investment. Underlying facts are drawn from the SEC orders and the Congressional Research Service report cited throughout this page.

The honest summary is that the absence of registration, an unverifiable whitepaper and an undisclosed paid endorsement were all things a careful buyer could have weighed against a specific ICO before sending Ether. None of them told that buyer which project was Munchee, which refunded everyone before any harm, and which was Centra Tech, which had invented the executives on its own about page.

How Far Did Bitcoin and Ether Fall in 2018?

The bust is usually narrated through enforcement dates, because those are the events with press releases attached to them. The price collapse underneath the enforcement timeline was larger, faster, and mattered more directly to any ICO issuer still holding its own treasury in crypto rather than cash.

Daily closing prices from the Coinbase series published by the Federal Reserve Bank of St. Louis (FRED series CBBTCUSD and CBETHUSD), retrieved and computed on 26 August 2026. Other exchanges' quoted peaks differ by small amounts; the magnitude and direction are consistent across venues.

Ether's decline is the more important number for this specific case study, not bitcoin's, because Ether was the currency most ICOs were actually funded in and, in many cases, the currency issuers kept sitting in their own treasuries rather than converting to dollars immediately. A project that raised the equivalent of $20 million in January 2018 and simply held what it raised in Ether, spending it down gradually as development proceeded, would have watched that treasury's dollar value fall by roughly the same 94.0 percent by December, regardless of whether a single dollar had been misused. This is a structural feature of how these offerings were funded, not a story about any particular issuer's competence or honesty, and it explains why "the money ran out" and "the money was stolen" were two very different explanations for the same visible outcome: an ICO-funded project going dark in 2018.

The Financial Stability Board's October 2018 report frames the same collapse from the market-capitalization side: crypto-assets broadly went from an estimated $830 billion combined value on 8 January 2018 to approximately $210 billion by 4 October 2018, a decline of roughly 75 percent in nine months. The FSB's own characterization of the episode is worth quoting directly, because it is doing analytical work most retrospective crypto commentary skips: "the market capitalisation relative to other financial markets and comparators has remained small," and the report goes on to compare crypto's scale to two prior asset bubbles, noting that in "the Dot-com bubble of 1997-2001, the value of tech stocks peaked at $3 trillion," a peak roughly four times crypto's own. Readers who want the fuller mechanics of that earlier comparison, including how a fifteen-year recovery actually played out, will find them in Swoopr's case study on the dot-com bubble.

What Did the SEC Do to ICO Issuers After the Market Turned?

The enforcement record did not arrive as a single sweep. It built case by case, over roughly three years, with escalating stakes and, in the later cases, escalating resistance from issuers who had more to lose by settling quickly.

Read as a sequence rather than as isolated headlines, the pattern is that consequences scaled with how much an issuer had raised and how it responded once contacted, not simply with whether fraud occurred. Munchee, facing no fraud allegation and cooperating within hours, paid nothing. Airfox and Paragon, also facing no fraud allegation but having already completed their raises and spent proceeds, paid a fixed $250,000 each and accepted ongoing reporting obligations. Kik chose to fight a case the SEC ultimately won outright, converting what might have been a negotiated settlement into a court-ordered judgment. Telegram's case involved the largest sum by far, and its resolution was correspondingly the most severe: not a fine calibrated as a fraction of proceeds, but the full return of the money and the termination of the network the money was raised to build.

What Happened to the Two Largest ICOs, EOS and Telegram?

EOS and Telegram sit at very different points on the outcome spectrum despite both starting as the two largest token sales the SEC ever addressed, and the comparison illustrates how differently "the largest ICO" and "the most severely punished ICO" can diverge.

Block.one's EOS offering ran from 26 June 2017 to 1 June 2018, roughly overlapping the entire arc of this case study, and sold 900 million ERC-20 tokens for what the Congressional Research Service, citing CoinSchedule data current as of September 2018, described as the largest ICO on record at $4.2 billion. The SEC's own September 2019 order describes the sum more conservatively as "several billion dollars," without confirming the $4.2 billion figure to the decimal, and explicitly does not allege fraud: Block.one's violation, in the SEC's finding, was that it "did not provide ICO investors the information they were entitled to as participants in a securities offering," a registration failure rather than a deception. The company paid a $24 million penalty, roughly half a percent of the low end of what it had raised, and continued operating.

Telegram's Gram token offering, by contrast, ran a much shorter window, January to March 2018, and sold 2.9 billion Grams to 171 purchasers for $1.7 billion, a concentration of buyers per dollar raised that stands out against every other issuer in this case study; EOS's several billion dollars came from a broad public sale, while Telegram's $1.7 billion came from 171 people. When the SEC filed suit on 11 October 2019 to block delivery of the tokens, the US District Court for the Southern District of New York issued a preliminary injunction on 24 March 2020, finding the SEC had shown a substantial likelihood of proving Telegram's private sale to those 171 purchasers was structured as the first step of a larger, unlawful scheme to distribute Grams onward to the public without registration. Telegram settled rather than litigate further, agreeing on 26 June 2020 to return more than $1.2 billion to investors and pay an $18.5 million penalty. The Telegram Open Network itself, the entire purpose the $1.7 billion had been raised for, was abandoned.

The comparison worth sitting with is not which company was more culpable in some abstract moral sense; the SEC's orders do not support ranking them that way, since neither case alleged fraud. It is that identical starting facts, a large unregistered token sale that the SEC concluded violated registration requirements, produced a functioning company paying a modest penalty in one case and the complete unwinding of the underlying project in the other. Timing, the composition of the buyer pool, and how much of the offering had already reached the secondary market by the time the SEC acted all shaped which outcome a given issuer got, and none of those three variables were things an ordinary token buyer could have priced in advance.

Did the ICO Bust Spill Over Into Banks or the Broader Market?

No. This is one of the more consistent findings across the crypto case studies in this library, and the ICO bust fits the pattern cleanly. The Financial Stability Board's October 2018 assessment, written specifically to evaluate whether crypto-asset market stress could threaten the broader financial system, concluded that "risks to global financial stability are not significant at present, given the limited size of crypto-asset markets relative to other financial markets and the limited interconnectedness between crypto-asset markets and the regulated financial system."

The FSB's own comparison to prior bubbles is instructive on why. Even at crypto's $830 billion January 2018 peak, the report notes the dot-com bubble of 1997 to 2001 saw tech-stock value peak at roughly $3 trillion, nearly four times crypto's peak size, while the run-up to the 2008 financial crisis involved a securitized mortgage market of roughly $7.3 trillion. Crypto's smaller absolute scale mattered less to systemic risk than its lack of connective tissue into the regulated financial system: no ICO token was held as a bank's regulatory capital, wrapped into a security a pension fund was required to own, or pledged as collateral in a market other institutions depended on for their own funding. When ICO issuers failed, lost money to hacks, or were forced by the SEC to return it, the loss stayed with the buyer of that specific token, not with a chain of counterparties.

That containment is a property of how the money moved, not a permanent feature of crypto markets generally. The FSB's report flags exactly this caveat in its own terms, and other episodes in this library, including the FTX collapse five years later, show what happens when crypto-native institutions begin lending to each other and holding each other's tokens as collateral: the failures start arriving in sequence rather than staying isolated, even though the underlying financial system still avoids direct exposure. The ICO era mostly predated that lending layer; issuers held their own funds and spent them down, they did not generally borrow against token collateral from other crypto institutions the way Alameda Research would a few years later. Readers interested in how a self-referential collateral loop actually breaks a company, once that lending layer exists, will find the fuller mechanism in that later case.

Common Myths About the ICO Boom and Bust

"Every ICO was a scam." The SEC's own enforcement record contradicts a blanket claim like this. Munchee refunded investors before any harm occurred. Airfox and Paragon built real products and were charged only with a registration failure, not deception. Even the Satis Group figure sometimes cited for "81 percent of ICOs are scams" used a specific definition of scam and, as the Congressional Research Service itself noted when citing it, other studies found a lower rate. Some ICOs were frauds. Treating the category as uniformly fraudulent erases the SEC's own case-by-case distinctions.

"The DAO Report ended the ICO boom." It did not, at least not immediately. The report was published 25 July 2017, and by most available measures ICO issuance and dollar volume continued rising for roughly another year, including Block.one's EOS offering, which did not even begin until 26 June 2017 and ran until 1 June 2018, entirely inside the window after the DAO Report had already been published. A legal conclusion is not the same instrument as an enforcement action, and the market kept operating on the gap between the two until individual issuers started losing real money and real cases.

"Regulators shut the market down." The SEC's enforcement actions targeted specific issuers over specific violations; nothing in the record describes a blanket prohibition on token sales. What ended the boom was closer to a combination of falling prices removing the speculative capital that had funded new raises, and a rising cost of getting caught once enforcement outcomes like Kik's and Telegram's became public, than any single rule change.

"Bitcoin and Ether crashed because of the SEC." The timeline runs the other direction. Bitcoin peaked on 16 December 2017 and Ether peaked on 13 January 2018, both well before the SEC's first ICO-specific settlement penalty in November 2018 and more than a year before its largest case, against Telegram, was even filed. The price collapse was underway on its own terms before the enforcement wave that this page's other sections describe had meaningfully begun.

"A registered token sale would have been just as easy." Registration under US securities law requires the disclosures, audits and ongoing reporting a public company provides, which is precisely the cost and friction ICOs were structured to avoid. Airfox and Paragon's settlements required them to register their tokens after the fact and file periodic reports for at least a year, a retrofit of exactly the obligations a compliant offering would have carried from the start, and both companies' post-settlement paths show what that retrofit actually costs an issuer that already spent the money it raised.

"Celebrity endorsements meant a token was probably legitimate." The Mayweather and Khaled case demonstrates the opposite mechanism directly: a celebrity was paid specifically to make an illegitimate token look ordinary. An endorsement conveys information about whether someone was paid to say something, not about whether the thing they said is true, and Section 17(b)'s disclosure requirement exists because that distinction is otherwise invisible to the audience.

What a Reader Can Actually Carry Forward

Naming Munchee, Centra Tech or Telegram will not help a reader spot the next bad offering; the instrument itself, an unregistered token sale funded in a volatile cryptocurrency with a whitepaper standing in for a prospectus, does not repeat identically. What carries forward is the underlying pattern the SEC's own case-by-case record traces, and that pattern has already recurred in other forms since.

What generalizes

• A new instrument outrunning old rules is a recognizable moment, not a one-time event. The Howey test predates the internet by decades and still applied to a smart contract in 2017. The lag was never about whether the law reached the transaction; it was about how long it took anyone to test that in a specific case, and that lag is where the largest opportunities for both genuine innovation and outright fraud both concentrate.
• Funding a project in the same volatile asset it will later need to spend creates a hidden risk unrelated to the project's own execution. Ether's 94.0 percent decline meant ICO treasuries shrank in dollar terms regardless of what any individual team did right or wrong. Any fundraising instrument denominated in something other than the buyer's or spender's actual unit of account carries this exposure, crypto or otherwise.
• A paid endorsement without disclosure is designed to look like an unpaid one. That is the entire function of the omission, and it is why securities law has required disclosure of compensated promotion since long before ICOs existed. The specific instrument changes; the incentive to hide payment for an endorsement does not.
• Enforcement consequences track more with an issuer's size, timing and response than with the presence of fraud alone. Two issuers can commit the identical technical violation, registration failure with no fraud alleged, and land on outcomes as different as Block.one's $24 million penalty against several billion raised and Telegram's full $1.2 billion return plus network shutdown.
• Containment inside a sector is a property of connections, not size. The 2017-2018 crypto market lost most of its value without touching the regulated financial system, because ICO proceeds were not pledged as collateral into markets other institutions depended on. That containment is not automatic once a lending layer connects crypto institutions to each other, a mechanism later case studies in this library trace directly.

What does not generalize

• The specific 81 percent "scam" figure. It depends on Satis Group's particular definition and the CRS itself flagged that other studies produced lower estimates. Citing it as a settled fact about ICOs overstates what one study under one methodology actually showed.
• Munchee's zero-penalty outcome as a template for future cooperation. It required stopping the offering within hours of contact and returning every dollar before a single token was delivered. Most issuers who reach a settlement have already spent proceeds and cannot replicate that path.
• Telegram's full return as evidence that ICO investors usually get made whole. Telegram had $1.2 billion available to return because it had not yet spent most of what it raised when the injunction hit. An issuer further along in spending its treasury, especially one that had spent it in a currency that lost 94 percent of its value, would not have the same capacity.

The one question worth asking now

For any fundraising instrument marketed as new, ask what specific legal obligation its novelty is being used to argue around, and ask who bears the cost if that argument turns out to be wrong later rather than sooner. ICO buyers in 2017 were, in effect, betting that "blockchain-based" meant something different in the eyes of the law than "unregistered security." The SEC's answer, reached case by case rather than all at once, was that it mostly did not. That question is answerable today, before money moves, for any instrument currently being sold on the same argument, and it requires no special expertise in the underlying technology to ask it. Our guides to fake token presales and fake celebrity endorsement scams cover how to apply that question to a live offering today.

References

Every figure on this page was verified against the following sources, each retrieved on 26 August 2026:

• US Securities and Exchange Commission: SEC Issues Investigative Report Concluding DAO Tokens, a Digital Asset, Were Securities: the 25 July 2017 publication date, the finding that DAO tokens were securities under Howey, the decision not to bring charges against The DAO or Slock.it, and the quoted statements from Chairman Clayton and Director Hinman.
• US Securities and Exchange Commission: Statement on Cryptocurrencies and Initial Coin Offerings: Chairman Clayton's 11 December 2017 statement, including that no ICOs had been registered with the SEC to that date and the quoted passage on "utility" tokens not escaping securities analysis.
• US Securities and Exchange Commission: Company Halts ICO After SEC Raises Registration Concerns: the Munchee cease-and-desist order of 11 December 2017, the $15 million sought, the refund before token delivery, and the decision not to impose a penalty.
• US Securities and Exchange Commission: SEC Halts Fraudulent Scheme Involving Unregistered ICO: the 2 April 2018 Centra Tech complaint, the more than $32 million raised, the fabricated Visa and Mastercard partnership, the fictional executives, the paid celebrity promotion, and the arrests of Sharma and Farkas.
• US Securities and Exchange Commission: Two Celebrities Charged With Unlawfully Touting Coin Offerings: the 29 November 2018 order against Floyd Mayweather Jr. and DJ Khaled, the undisclosed Centra Tech payments of $100,000 and $50,000, the additional $200,000 to Mayweather from two other issuers, and the disgorgement, penalty and promotion-ban figures for both.
• US Securities and Exchange Commission: Two ICO Issuers Settle SEC Registration Charges, Agree to Register Tokens as Securities: the 16 November 2018 Airfox and Paragon Coin settlements, the approximately $15 million and $12 million raised respectively, the $250,000 penalty each, and the description of these as the first cases imposing penalties solely for registration violations.
• US Securities and Exchange Commission: SEC Charges Issuer With Conducting $100 Million Unregistered ICO: the 4 June 2019 Kik Interactive complaint, the more than $100 million raised, the more than $55 million from US investors, and the one trillion Kin tokens sold.
• US Securities and Exchange Commission: SEC Obtains Final Judgment Against Kik Interactive For Unregistered Offering: the 30 September 2020 summary judgment finding a single integrated offering, the 21 October 2020 final judgment, the $5 million penalty, and the three-year notice requirement.
• US Securities and Exchange Commission: SEC Orders Blockchain Company to Pay $24 Million Penalty for Unregistered ICO: the 30 September 2019 Block.one settlement, the June 2017 to June 2018 offering window, the 900 million tokens sold, the "several billion dollars" description, the $24 million penalty, and the absence of any fraud allegation.
• US Securities and Exchange Commission: SEC Halts Alleged $1.7 Billion Unregistered Digital Token Offering: the 11 October 2019 emergency action against Telegram, the $1.7 billion raised, the 2.9 billion Grams sold to 171 purchasers, and the temporary restraining order.
• US Securities and Exchange Commission: Telegram to Return $1.2 Billion to Investors and Pay $18.5 Million Penalty to Settle SEC Charges: the 26 June 2020 settlement, the more than $1.2 billion returned, the $18.5 million penalty, and the 24 March 2020 preliminary injunction finding.
• US Securities and Exchange Commission: Digital Asset Transactions: When Howey Met Gary (Plastic): William Hinman's 14 June 2018 speech, the Howey test elements, the "sufficiently decentralized" reasoning, and the direct comparison to the facts of SEC v. Howey.
• Financial Stability Board: Crypto-asset markets: Potential channels for future financial stability implications, 10 October 2018: the $830 billion peak market capitalization on 8 January 2018, the approximately 35 percent bitcoin share, the decline to approximately $210 billion by 4 October 2018, the finding that risks to global financial stability were not significant at present, and the dot-com and subprime-mortgage comparison figures.
• Congressional Research Service: Financial Innovation: Digital Assets and Initial Coin Offerings (IF11004), 17 October 2018: the $4.2 billion EOS figure attributed to CoinSchedule, the more than 2,000 digital assets figure, the Satis Group scam-rate study, the EY hacking-loss study, and the Malta and EOS US/China exclusion details.
• Federal Reserve Bank of St. Louis (FRED): Coinbase Bitcoin (CBBTCUSD): the $19,650.01 peak close on 16 December 2017 and the $3,183.00 low close on 15 December 2018.
• Federal Reserve Bank of St. Louis (FRED): Coinbase Ethereum (CBETHUSD): the $1,386.02 peak close on 13 January 2018 and the $83.00 low close on 14 December 2018.

Figures deliberately not stated. This page gives no single total for global ICO fundraising across 2017 and 2018, because no primary or institutional source consulted in preparing it supplies a verified figure of that kind; the individual deal sizes documented above are precise, but their sum is not represented as a verified market total. This page also does not state a specific ICO count for the period, beyond the Congressional Research Service's more-than-2,000 total for all digital assets existing by late 2018, which covers a broader category than ICO tokens alone.

Method note: allegations in an SEC complaint are identified as such in the text and distinguished from findings in a settled order or a court's judgment. The classification of which facts were publicly visible before a given ICO closed is Swoopr Investment's own assessment applied to sourced facts, and is labeled as such in the relevant table.

This is educational content about a historical episode. It is not investment advice, it is not a forecast, and nothing here should be read as a claim about how any future token offering will behave. Regulatory guidance discussed here, including the SEC's application of the Howey test to digital assets, reflects the legal and enforcement record through the dates cited and can be revised by future rulemaking, legislation or court decisions.

Frequently Asked Questions