The Complete History of Crypto Presales (2013–2026)
In July 2013, someone sent Bitcoin to an address posted on a forum and hoped. That was Mastercoin — the first crypto presale — and roughly USD 5 million arrived. Thirteen years later, token presales have moved tens of billions of dollars, minted a handful of the most important networks in computing, and torched fortunes at a scale nineteenth-century railway promoters would have admired.
This is the full history, told from data rather than memory. Wave 1 of our research dataset documents 2,300 token sales from 2013 to 2026 — every row backed by at least one public source, with money figures recorded only where the cited source actually shows one (864 rows). Where a number cannot be sourced, we say so. That rule matters more in this corner of crypto than anywhere else, as you are about to see.
All figures computed from the dataset. Aggregator-sourced figures can include private rounds alongside public sales; the methodology section explains how each tier of the data is verified.
Era one: the genesis years (2013–2016)
Mastercoin proved the mechanism; Ethereum proved the point. Its 42-day sale in mid-2014 took in about USD 18.3 million in Bitcoin at roughly 31 cents per ETH — and used the money to ship the platform on which nearly every later sale in this article would run. Augur (USD 5.1 million, 2015) became the first big sale on Ethereum; DigixDAO sold out in 12 hours in March 2016; Lisk, Waves, Golem and FirstBlood filled out the first real cohort. China's Antshares — later NEO — took in USD 4.7 million across its 2015–16 crowdsales and briefly became a top-10 network.
Then came the era's morality tale. The DAO pulled in roughly USD 160 million in ETH in spring 2016 — the largest crowdfunding of anything, ever, at that point — and was drained within weeks through a contract flaw. Ethereum hard-forked to claw the funds back, splitting the chain in two. The lesson the industry would spend a decade relearning was already on the table: the code that holds the money is the product, and nobody had audited it hard enough.
Genesis era in the dataset: 20 sales, median sourced figure USD 5.5 million.
Era two: ICO mania (2017–2018)
In June 2017, Bancor collected USD 153 million in ETH in about three hours. Basic Attention Token took USD 35 million in under thirty seconds. Status brought in USD 103 million and congested all of Ethereum doing it. Filecoin set a record with USD 257 million; Tezos beat the uncapped-sale record with USD 232 million and then spent a year fighting itself in court before shipping. TRON closed its USD 70 million sale days before China banned ICOs outright in September 2017.
Two sales towered over everything. EOS ran a year-long continuous sale that took in roughly USD 4.1 billion — still the largest ever — and later settled unregistered-offering charges with the U.S. securities watchdog for a USD 24 million penalty. Telegram collected USD 1.7 billion for its Gram token in early 2018; the same watchdog halted the offering, and Telegram returned USD 1.2 billion to purchasers and paid a USD 18.5 million penalty. The chain survived anyway — revived years later by its community as TON.
The frauds scaled with the era, and pretending otherwise would falsify the history:
| Scheme | Years | Money taken | What happened | |
|---|---|---|---|---|
| OneCoin | 2014–17 | ~$4,000,000,000 | A "coin" with no blockchain at all, sold through multi-level marketing; founder Ruja Ignatova vanished in 2017 and is on the FBI's Ten Most Wanted list. | source |
| BitConnect | 2016–18 | ~$2,400,000,000 | A lending-scheme Ponzi wearing a coin costume; the token fell 92% in a day when it collapsed in January 2018. | source |
| Pincoin / iFan | 2018 | ~$660,000,000 | Vietnam's Modern Tech vanished with the money of ~32,000 buyers — one of the largest exit scams on record. | source |
| Petro | 2018 | claimed $3,300,000,000 | Venezuela's state oil token; the pre-sale figure was never independently audited, and the token was shut down and liquidated in January 2024. | source |
| Centra Tech | 2017 | >$32,000,000 | Celebrity-promoted payments card that did not exist; the founders were charged with fraud in 2018 and convicted. | source |
| Envion | 2017–18 | ~$100,000,000 | Swiss authorities found the sale unlawful and a court ordered the company liquidated. | source |
| Prodeum | 2018 | almost nothing | Promised fruit-and-vegetable tracking on Ethereum, exit-scammed leaving a single rude word on its homepage — the era in miniature. | source |
Mania era in the dataset: 234 sales with a median sourced figure of USD 20 million — the highest of any era. Counting the documented frauds beside the genuine sales, 40% of every sourced 2013–2018 dollar went into schemes now known to be frauds.
Era three: winter, launch platforms and auctions (2019–2020)
After the crash, the public sale went respectable — and shrank. Sales moved onto vetted launch platforms and price-discovery formats: Algorand ran a Dutch auction that brought in USD 60 million in June 2019; the company behind Bitfinex sold USD 1 billion of LEO privately in ten days to plug a USD 850 million hole. The most telling number of the era is Solana's: its March 2020 public auction collected just USD 1.8 million — of roughly USD 359.7 million the project took across seven rounds, nearly all was private. The crowd that funded Ethereum had been replaced by venture funds, with the public sale kept as a marketing sliver.
Winter era in the dataset: 73 sales, median sourced figure USD 5 million.
Era four: the IDO years (2021–2023)
Launchpads on DEXes made token sales permissionless again — and small. The 787 sales our dataset holds for 2021–2023 have the lowest median sourced figure of any era (USD 3.6 million): thousands of gaming and DeFi tokens collecting a few million each, most now untraceable. The era's most human moment wasn't an IDO at all: ConstitutionDAO crowd-collected USD 49.5 million in ETH in under a week to bid on a printed copy of the U.S. Constitution, lost the auction, and gave the money back.
Era five: the presale era (2024–2026)
The current cycle brought the direct-to-website presale: no launchpad, no venue, a widget on the project's own site selling through dozens of stages. More than half our dataset — 1,186 sales — belongs to this era. Meme presales set its tone: Pepe Unchained (reported ~USD 73 million), Wall Street Pepe (reported >USD 73.9 million at sell-out), Solaxy (reported ~USD 55 million), Qubetics (reported ~USD 18.4 million), and wallet-token sales like Best Wallet. At the institutional end the format got enormous again: World Liberty Financial sold USD 550 million of tokens, Pump.fun collected roughly USD 500 million in twelve minutes, MegaETH's auction drew USD 1.39 billion in bids for a ~USD 450 million sale, and Plasma took USD 373 million against a USD 50 million target.
But note the word that recurs in that paragraph: "reported." Most presale-era totals are project-reported numbers on a website widget — not figures you can check on a block explorer. Across our whole dataset, project-reported and aggregator figures dominate the modern era, while the sale history's best-documented figures are, ironically, a decade old. That is the single biggest data-quality problem in presales today, and it is fixable, because the fix already exists: publish the total on-chain.
Full disclosure, since this is a BMIC research piece: BMIC's own presale is one of the 1,186 era-five sales in this dataset, and it is the only one in the set whose money figure anyone can verify in one click — the contract at 0xf36523f1d4ed392E5426aaf06e376Ba9042dAaaB exposes totalUSDRaised() on Ethereum, reading $637,490 across 1,354 purchases at $0.0528542 per token as this page was built. If the era's lesson is "verifiable beats reported," we hold ourselves to it first.
The biggest sales, sourced
Frauds excluded (they got their own table). Every figure below links to the archive or press source that shows it.
| # | Sale | Year | Amount | Outcome | Source |
|---|---|---|---|---|---|
| 1 | EOS | 2018 | $4,100,000,000 | listed | source |
| 2 | Telegram Open Network (Gram) | 2018 | $1,700,000,000 | refunded | source |
| 3 | UNUS SED LEO (Bitfinex/iFinex) | 2019 | $1,000,000,000 | active | source |
| 4 | World Liberty Financial | 2025 | $550,000,000 | listed | source |
| 5 | Pump.fun | 2025 | $500,000,000 | listed | source |
| 6 | MegaETH | 2025 | $450,000,000 | listed | source |
| 7 | Plasma | 2025 | $373,000,000 | listed | source |
| 8 | Monad | 2025 | $269,000,000 | listed | source |
| 9 | Hdac (Hyundai BS&C) | 2017 | $258,000,000 | listed | source |
| 10 | Filecoin | 2017 | $257,000,000 | listed | source |
| 11 | Tezos | 2017 | $232,000,000 | listed | source |
| 12 | The DAO | 2016 | $160,000,000 | failed | source |
| 13 | Sirin Labs | 2017 | $158,000,000 | inactive | source |
| 14 | Bancor | 2017 | $153,000,000 | listed | source |
The numbers by era
| Era | Sales in dataset | With sourced figure | Sum of sourced figures | Median |
|---|---|---|---|---|
| 2013–2016 | 20 | 17 | $298,176,862 | $5,500,000 |
| 2017–2018 | 234 | 151 | $10,956,900,261 | $20,000,000 |
| 2019–2020 | 73 | 35 | $1,806,089,999 | $5,000,000 |
| 2021–2023 | 787 | 268 | $3,027,006,101 | $3,590,000 |
| 2024–2026 | 1,186 | 372 | $7,394,702,777 | $5,615,000 |
Sums count only figures visible at a cited source, with the documented frauds excluded, and aggregator listings can bundle private rounds into their totals — treat era sums as a documented floor-with-caveats, not a market total. Nobody, including us, has a true total for this industry; that is partly the point of building the dataset.
What separated the survivors
Reading 2,300 rows of outcomes, the pattern is unglamorous:
- Product before promises. Ethereum, Cosmos, Chainlink and Filecoin all shipped the thing the sale paid for. Nearly every mega-sale that died had shipped a whitepaper and a countdown timer.
- Verifiable numbers. The DAO's collapse, Petro's unauditable claim, and a decade of "reported" totals all point one way: totals that can be checked on-chain age better than totals that must be believed.
- Independent code review. The DAO was the first nine-figure lesson that unaudited contracts are unexploded ordnance. It was not the last.
- Honest wind-downs happen — and get remembered. Cofound.it returned remaining funds when it shut in 2018; ConstitutionDAO refunded; Telegram gave back USD 1.2 billion. The teams that respected buyers on the way down are the era's quiet heroes.
- The red flags never change: anonymous teams, celebrity promotion, yield promises, and totals that only ever appear on the project's own widget.
That reading list is why our related work exists: our ten-minute due-diligence checklist operationalizes the survivor pattern, and our regularly updated best crypto presale rankings apply it to sales running right now. If you want the wider field in one place, the crypto presale list is the living companion to this history, and the current presale rankings show which of era five's sales pass the same tests the 2017 class failed.
Methodology and sources
The rule: no source, no row. Wave 1 of the dataset was compiled on August 27, 2026 from public archives, and every row carries at least one URL where the claim can be checked:
- Tier 1 — hand-verified (61 rows): figures checked against the page text of Wikipedia's list of highest-funded crowdfunding projects and its initial-coin-offering history, plus ICODrops project archives and live on-chain data.
- Tier 2 — research-lane verified (164 rows): era-by-era rows compiled from news archives (CoinDesk, TechCrunch, CNBC, The Defiant and others), each source-checked before acceptance.
- Tier 3 — aggregator-listed (2,075 rows): bulk sale listings from the CryptoTotem past-sales archive. These document that a sale existed and when; their money figures can bundle private rounds and are labeled per-listing.
Money figures are recorded only when visible at the cited source — 864 of 2,300 rows. Duplicates across sources were merged (best source wins); 130 candidate rows were rejected for missing years or unusable sourcing, and the rejection log is kept with the dataset. Known gaps: pre-2016 coverage is thin outside the famous names, dead projects with dead sources are underrepresented by construction, and "every presale ever" is an asymptote — wave 2 grows the long tail.
Corrections with sources: support@bmic.ai. The dataset is CC-BY 4.0 — reuse it with a link back to this page.
FAQ
What was the first crypto presale?
Mastercoin, July 2013 — about USD 5 million in Bitcoin sent to an address on a forum post. Ethereum's 2014 sale made the model matter.
What is the biggest crypto presale ever?
EOS: roughly USD 4.1 billion across a year-long sale ending 2018. Telegram's USD 1.7 billion (2018) was halted and largely refunded; among sales that completed and kept the money, nothing else comes close.
How many crypto presales have there been?
Nobody knows — thousands, many of which left no reliable record. This dataset documents 2,300 with sources, and grows by wave.
This article is historical research, not a recommendation to buy anything. Token sales are high-risk; most of the 2,300 sales in this dataset did not make anyone wealthy. Do your own research — the sources are all here.