Launch Autopsy #1: Ethereum's ICO
The sale that wrote the token-launch playbook: 42 days, ~31,500 BTC, 60 million ETH — priced in a currency that then halved. What the first great token launch got right, what it got away with, and what only worked because it was 2014.
Bruno Pierri GalvaoThis is the first study in a series with a simple premise: token launches are among the highest-stakes, least-repeatable events in crypto — most teams get exactly one — yet the record of what actually worked is scattered across old forum threads and survivor memory. Each study takes one chain and runs the same autopsy: how the token launched, who got what, the crypto economy it launched into, and whether the design or the era decided the outcome. Same rubric every time, so the studies add up to something.
Starting where the playbook was written: Ethereum, July 2014.
The world it launched into
Nothing about mid-2014 said “good time to sell a new token.” — then the largest Bitcoin exchange — had collapsed five months earlier, taking roughly 850,000 BTC and most of the market’s confidence with it. Bitcoin had fallen from its late-2013 high above $1,100 to around $600 and was grinding lower. The regulatory landscape wasn’t hostile so much as absent: no framework, no precedent, no safe harbor — just a handful of lawyers advising founders to hope.
“Selling a token” barely existed as a concept. had run the first recognizable a year earlier; had run a messy one that April. The dominant opinion in the room that mattered — the Bitcoin community — was that altcoins were scams by construction, and Ethereum was called exactly that, loudly, by serious people.
That’s the first thing to hold onto: the most successful token launch in history happened in a bear market, into open hostility, under total regulatory uncertainty. The era did not hand Ethereum its outcome.
The mechanics
The sale ran 42 days, July 22 to September 2, 2014, priced in BTC: 2000 ETH per BTC for the first two weeks, then stepping down to a final rate of 1337 ETH per BTC — an early-buyer discount of about 50% (and yes, the closing rate is a meme; 2014 was like that).
Design choices worth naming plainly:
- Uncapped. The sale had no ceiling; it would absorb whatever came. This was criticized then and would be radioactive now — an uncapped sale converts hype directly into insider-controlled treasury with no price discovery.
- Premine on top. Beyond the 60.1M ETH sold, two further pots of 9.9% of the sold amount each — one for ~80 early contributors, one for the Ethereum Foundation — brought supply to ~72M ETH. Insiders held roughly 16.5% of genesis: modest by later standards (VC-era chains routinely reserved 40–60%), scandalous by the standard Bitcoin had set.
- No . Nothing at genesis carried a protocol-enforced lockup — every balance, the endowment pools included, was transferable from block 0. The Foundation’s pot was a “long-term endowment” by stated policy, not by code.
- No , no , no jurisdiction gating. Anyone with BTC could buy. This is the single least repeatable feature of the whole launch.
- BTC-denominated treasury. The raise was collected and largely held in BTC, unhedged. This decision nearly killed the project — next section.
The whole package on one card — the anatomy every study in this series will repeat, so the panels stay comparable from token to token:
What happened next
The gap between sale and chain was almost a full year: genesis came July 30, 2015. In that year, Bitcoin kept falling — from ~$600 at the sale to under $250 by mid-2015 — and the foundation’s unhedged BTC treasury roughly halved in dollar terms before much of it could be spent. The organization that had just run an $18M raise spent 2015 in genuine austerity, cutting staff and, by its own later accounts, coming uncomfortably close to running out of money before the chain had proven anything.
BTC/USD · monthly close · Jul 2014 – Aug 2015
ETH listed in August 2015 around $2.77, then did what almost every new listing does: fell hard, bottoming around $0.42 that October — down ~85% from listing, yet still ~35% above the ICO price. That detail rewards attention. The sale price was low enough that even the post-listing capitulation never put ICO participants underwater. Every 2014 buyer who held was made whole at the very bottom and life-changingly rich within two years. That — more than any mechanism — is what built Ethereum’s famously durable holder base: the launch never gave its earliest believers a reason to become sellers.
ETH/USD · Aug 2015 – Jan 2016
Demand caught up to the token in stages, none of them present at launch: was the only sink in 2015; the 2017 ICO boom made ETH the reserve currency of an entire asset class; made it collateral in 2020; made it partially deflationary in 2021; made it yield-bearing in 2022. The launch design deserves credit for one thing here: it didn’t foreclose any of these. It sold a commodity (“fuel for computation”) and let the sinks arrive.
What worked
- Wide, cheap, global distribution. Roughly 8,400 purchases at ~$0.31 with no gatekeeping produced a holder base that doubled as a user base, an evangelist base, and eventually a validator base. No -farming industry existed to game it.
- Low insider share, low expectations. ~16.5% insider allocation and an $18M raise look quaint against later launches — and that modesty was load-bearing. There was no multi-billion to grow into, no hanging over year two.
- The bear-market filter. Selling into hostility meant the buyers were believers with long horizons, not momentum tourists. The people who bought because of a 2014 forum argument were still building on Ethereum in 2020.
- Honest framing. ETH was sold as fuel, priced like fuel, and the chain shipped. The gap between what was promised and what was delivered — the gap that later defined the 2017 class — was small.
What broke, or nearly did
- The unhedged BTC treasury. Raising in a volatile asset and holding it through a bear market halved the war chest before the product existed. This is the most directly transferable negative lesson in the study: treasury denomination is a tokenomics decision, and it nearly ended Ethereum before the genesis block.
- Uncapped and unvested. Ethereum got away with both because the amounts were small and the buyers were believers. The same two choices, re-run at 2017 scale by projects with none of Ethereum’s discipline, produced the ICO bubble’s worst outcomes. A playbook isn’t just what worked — it’s what worked only because of its context.
- A year of regulatory luck. The sale predated any framework and was never retroactively punished; the SEC’s 2017 and the 2018 posture arrived as after-the-fact absolution. Nothing about that sequence is reproducible. Treating Ethereum’s regulatory path as precedent for a new launch is the most common way this case study gets misread.
- No supply cap. Uncapped issuance was a live criticism for years and only stopped mattering when EIP-1559 and rebuilt the supply schedule in flight — a rescue available only to a chain that had already won.
Scorecard
| Dimension | Grade | Note |
|---|---|---|
| Distribution fairness | A− | Global, cheap, ungated; early-buyer discount the only tilt |
| Insider overhang | B+ | ~16.5%, but unvested; saved by small absolute size |
| Demand sinks at launch | C | Gas only, and the chain was a year away |
| Demand sinks, eventual | A | Fuel → reserve asset → collateral → burn → yield |
| Emission schedule | C+ | Uncapped PoW issuance; fixed only years later |
| Treasury management | D | Unhedged BTC through a bear market; near-death |
| Regulatory posture | Incomplete | Unpunished, then absolved; not reproducible |
| Macro timing | B | Bear-market launch filtered for believers and set a low bar |
Next in the series: the era Ethereum’s success created — the 2017 ICO boom, where every one of these choices was re-run at 100× scale, without the context that made them safe.