What an ICO Is and What Is Left of Them in 2026

Key points

An ICO is a public sale of newly created tokens. It worked in 2017 because issuing a token takes minutes, capital was global, and no prior filter existed at all.
FDV is the concept most people get wrong: a token with a $2bn diluted valuation and only 8% circulating needs to absorb $1.84bn of future supply to hold its price.
The SEC made Telegram return over $1.2bn in 2020 and fined Block.one $24m in 2019: the regulatory risk of a token sale is not theoretical.
In the EU, MiCA has required since 2024 that a white paper be notified to the regulator, and it makes the issuer liable for its content. Notified does not mean approved.
Buying at listing when most of the supply is locked means taking the short side of an unlock schedule that is already written and already public.

Market data: Market data on this page was collected on 28 August 2026 and changes constantly. Check the linked sources before relying on it.

An ICO (initial coin offering) is a public sale of newly created tokens to fund a project that does not yet exist. The comparison everyone reaches for is an IPO, and it is the comparison that confuses most: in an IPO you buy a share of a company with audited accounts and shareholder rights; in an ICO you buy an accounting entry in a smart contract whose value depends on someone later building what they promised.

Worth saying plainly at the outset: taking part in a token sale is a high-risk speculative transaction, with a serious probability of total loss and none of the recovery mechanisms that exist in a regulated securities market. This guide does not tell you whether to do it. It explains what exactly you are signing up for when you do, and which numbers you should have checked first.

Why the model worked so well in 2017

Three conditions lined up, and none of them was accidental.

Issuing a token is trivial. An ERC-20 token on Ethereum is a contract of a few dozen lines: a map of addresses to balances plus functions to transfer. No permissions, no intermediaries, no infrastructure. If you want the technical distinction between a token like that and a coin with its own chain, it is in token versus coin.

Capital was global and frictionless. Anyone with a wallet could send ether from any country at any hour. No bank account, no broker, no market hours, no jurisdictional restrictions beyond whatever the issuer chose to apply.

There was no filter. No approved prospectus, no audit, no capital requirements, no clear issuer liability. In a securities market someone reviews the document before you read it. Here the document was a PDF with a diagram and a team page of profile photos.

Add an underlying asset that kept rising and a scarcity narrative manufactured by countdown timers, and you have the complete model.

What is actually known about how they ended

Here comes a methodological warning that no optimistic article includes.

Widely quoted percentages circulate about what proportion of 2017–2018 ICOs ended in failure, abandonment or fraud. Almost all of them come from private consultancy reports with different methodologies and mutually inconsistent definitions of "failure," and they get repeated from article to article without anyone checking the original source. I am not going to cite any of them, because I have no verifiable primary source to back them up. A figure you cannot trace is not data. It is decoration.

What is documented and verifiable are the enforcement files. Two very large ones bracket the end of the cycle:

Telegram (Grams / TON). The SEC sued Telegram on 11 October 2019 over the sale of roughly 2.9 billion Grams to 171 initial purchasers, which had raised approximately $1.2 billion. On 26 June 2020 the court approved the settlement: Telegram returned more than $1.2 billion to investors and paid an $18.5 million civil penalty.
Block.one (EOS). On 30 September 2019 the SEC ordered Block.one to pay a $24 million penalty for an unregistered ICO of 900 million tokens conducted between June 2017 and June 2018, which raised several billion dollars. The SEC's finding was that it had not given investors the information they were entitled to.

Two projects with enormous visibility, real teams and abundant capital, both sanctioned. The pattern you can assert without inventing figures is this: most tokens from that cycle stopped showing development activity, market activity or both, and people who bought at the peak of the euphoria did not get their money back. That statement is deliberately qualitative.

What it became: IEOs, IDOs and private rounds

Format	Where it is sold	Who filters	What improves	What it does not fix
Classic ICO	The project's own website	Nobody	—	Everything
IEO	A centralised exchange (launchpad)	The exchange, on its own undisclosed criteria	Some diligence and KYC	The exchange charges to list and is not liable for the project
IDO	A DEX or on-chain launchpad	The code and sometimes a community	Open access, immediate liquidity	Bots, manipulated listings, removable liquidity
Private round + listing	Bilateral agreements beforehand	The project itself	Stable capital, better terms for the fund	Retail buys last, and dearer
MiCA-compliant public offer	Anyone, with a notified white paper	The regulator receives, it does not approve	Mandatory disclosure and issuer liability	No guarantee the project is worth anything

The evolution has not been towards less risk but towards more layers. In a private round followed by a listing, by the time the token reaches the open market the early investors are already in at a far lower price, and the listing price reflects expectations rather than results. Retail is not in the front row of that sale. It is in the last one.

The terms that decide the outcome

The six concepts below appear in any token sale document and explain more about what is going to happen than the entire "vision and mission" section.

Tokenomics. How total supply is split: public, team, investors, treasury, incentives, reserves. If team and investors together exceed half, you are exit liquidity by design, not by accident.
Cliff. Months during which an allocation releases nothing. With no cliff, whoever bought at $0.05 can sell the day you buy at $2.
Vesting. The release pace afterwards, usually linear. A 24-month vest spreads the pressure; a 6-month vest concentrates it.
Unlock schedule. The specific calendar of unlocks, with dates and quantities. It is public information and almost nobody reads it before buying.
Circulating market cap. Price × tokens already circulating.
FDV (fully diluted valuation). Price × total supply, including what does not yet exist in the market.

The relationship between the last two is where most money is lost.

Worked example: $2bn FDV with 8% circulating

Suppose a token lists with these numbers:

Total supply: 1 billion tokens.
Listing price: $2.00.
FDV: 1bn × $2 = $2 billion.
Circulating: 8% → 80 million tokens.
Circulating market cap: 80m × $2 = $160 million.

The headline will say "project valued at $2 billion." The money actually invested to hold that price is $160 million. The difference, $1.84 billion in tokens, exists and will reach the market on a calendar that is already written.

Now the calculation that matters. Assume a 12-month cliff and then linear vesting over 24 months for the remaining 920 million:

920m / 24 = 38.3 million tokens a month entering circulation.
At $2, that is $76.6 million of new supply every month.
Initial circulating market cap was $160 million. Every month, supply arrives equivalent to nearly half of all the capital holding the price up.

And the limit case: if net demand freezes at that $160 million and all supply eventually circulates, the equilibrium price would be $160m / 1bn = $0.16. That is 92% below the listing price, with no fraud, no hack and no bad news. Supply arithmetic alone.

A high FDV with little circulating supply is not a sign of success: it is a supply debt. For the price to hold, someone has to bring new capital at the pace tokens unlock. If somebody tells you "only 8% is circulating, imagine the potential," they are describing the risk and calling it an opportunity.

What MiCA actually changes

Since Title II of Regulation (EU) 2023/1114 came into application (30 December 2024 for service providers; the public-offer obligations sit within the same framework), anyone offering a crypto asset to the public in the EU or seeking its admission to trading must produce a white paper with mandatory content, notify it to the competent national authority and publish it. The issuer is legally liable for information that is misleading or incomplete, and consumers buying directly from the offeror have a 14-calendar-day right of withdrawal in public offers that are not admitted to trading.

Three things are worth understanding properly.

Notified is not approved. The authority receives the document. It does not validate it, does not bless it and expresses no view on the quality of the project. Any campaign implying otherwise is misrepresenting the rule.

Liability is the real novelty. Before MiCA, a white paper was literally a PDF with no legal consequences. Now there is an identifiable party answerable for its content. That changes the calculus of whoever writes it.

There are exemptions. The white paper regime does not apply equally to every offer: the Regulation carves out offers made for free, offers to fewer than 150 persons per member state, offers solely to qualified investors, and offers whose total consideration does not exceed one million euros over twelve months. A project publishing no white paper may be exempt rather than in breach, and the distinction matters before accusing anyone of anything.

Separately, MiCA governs who may sell you the token. National transitional periods ran out through 2026 — in Spain, on 1 July 2026 — and from then on a provider needs a CASP licence or an EU passport. Outside the EU the licensing question exists too, under different names and different regulators; we go through the landscape in the regulation guide.

The United States: the March 2026 taxonomy

On 17 March 2026 the SEC and the CFTC published a joint interpretation setting out a taxonomy of crypto assets: digital commodities, digital collectibles, digital tools, stablecoins and digital securities. The central point is the explicit recognition that most crypto assets are not themselves securities, alongside clarifications on airdrops, protocol mining, protocol staking and wrapping.

That does not absolve token sales. The doctrine that caught Telegram and Block.one pointed at the investment contract wrapped around the sale — the promise of the promoter's future efforts — rather than at the token as an object. A token can be a digital commodity and its initial sale still be an unregistered securities offering. The taxonomy clears the ground; it does not clear the way.

Concrete warning signs

The unlock schedule is not published, or exists only as an image with no exact figures.
The team is anonymous and simultaneously controls the treasury and the contract's admin key.
There are bonuses for inviting other buyers: that is a referral structure, not a distribution.
The white paper devotes more space to "appreciation potential" than to what the product does.
There is a countdown timer with a shrinking discount: the purpose is to stop you thinking.
An advertised audit turns out to be a one-day report on a contract different from the deployed one.
A celebrity promotes the sale without disclosing that they are paid to do it.

What to check before committing money

Calculate FDV and compare it with circulating market cap. If the ratio exceeds 5x, understand why.
Find the unlock schedule and mark the dates of the large tranches in your calendar.
Check what percentage the team and investors keep, and what price they paid.
Read the notified white paper, not the marketing summary, and check who is named as liable.
Verify in your national regulator's official register that the platform selling it to you is authorised where you live.
Write the full amount off before you send it: if that assumption makes you uncomfortable, the amount is too large.
Remember that selling or swapping tokens is a taxable disposal in most jurisdictions, and that receiving tokens for free has its own rules, as we explain in [the airdrop guide](/en/what-is-a-crypto-airdrop/).

The tax question nobody asks first

Two separate events matter, and people usually think about only the second.

The first is acquisition. If you paid for the tokens, that is normally just a cost basis and nothing is due. But if you paid in another crypto asset — sending ether to a sale contract, for instance — most jurisdictions treat that as disposing of the ether, which can crystallise a taxable gain at the moment of purchase, before the new token has done anything at all. People who bought in 2017 with heavily appreciated ETH discovered this the hard way.

The second is disposal, when you eventually sell or swap the token. The gain is generally measured against the cost basis in your local currency, and the applicable rate depends on your country, your residence status and often the holding period. Some jurisdictions also let realised losses offset gains, which matters a great deal in a category where total losses are common — but the offsetting rules, and whether a worthless token even counts as disposed of, vary widely.

Locked tokens are the awkward case. If an allocation vests to you over years, when the tax event occurs — at grant, at unlock, or at sale — is treated differently across jurisdictions, and some have published nothing on it. Check your own tax authority's guidance and take professional advice before assuming the answer.

What is left of ICOs in 2026 is a market with more regulation in the wrapper and the same arithmetic underneath. The white paper now has a named party answerable for it; the unlocks are still the same, still published, and still the thing that decides the price.

Frequently asked questions

Do ICOs still exist in 2026?

The 2017 format — a website, a white paper and an address to send ether to — has all but disappeared. What remains are variants with an intermediary: IEOs on an exchange, IDOs on a decentralised launchpad, and private rounds with vesting followed by a listing. The underlying mechanic is identical: someone creates a token and sells it to you before a product exists.

What is FDV and why does it matter so much?

Fully diluted valuation is the token price multiplied by total supply, including the part not yet circulating. Circulating market cap counts only tokens already issued. When FDV is ten or twenty times circulating market cap, you are buying at a valuation that only holds if the market absorbs all future supply at the same price.

Does MiCA make a token sale safe?

No. MiCA requires disclosure and creates legal liability for the issuer over the white paper's content, but no European authority approves or validates the project. The white paper is notified, not authorised. A token can be fully MiCA-compliant and worth zero a year later.

What are a cliff and vesting?

The cliff is the initial period during which the team or investors receive no tokens at all. Vesting is the gradual release afterwards, usually linear over months or years. A short cliff, or none, means whoever bought at a far lower price than you can sell almost immediately.

Sources and references

Important notice. InfoCrypto is an educational site. Nothing here is financial, tax or investment advice. Crypto assets are high risk: you can lose all the capital you invest. Always verify data against the original sources before making any decision, and consult a licensed professional if you need to.