Ethereum

Buterin circulated the white paper in November 2013. Gavin Wood turned the proposal into the Yellow Paper's formal machine and helped create Solidity; Jeffrey Wilcke built Geth, while Mihai Alisie worked on the Swiss foundation and sale. The official history names eight co-founders, not a single corporate inventor.

Those people did not remain under one command. Some formed other companies or protocols, while the nonprofit Ethereum Foundation became one contributor among client teams, researchers and application builders. Ethereum therefore launched with recognizable authors but without a CEO able to order every node to upgrade.

The 42-day sale opened on July 22, 2014 and brought in about 31,000 BTC, then worth roughly US$18 million. Its announcement called ether a product for using network services and explicitly withheld voting rights and any promise of value—a narrower bargain than the investment stories that later accumulated around ETH.

Frontier went live on July 30, 2015. It gave developers a general-purpose execution layer: contracts could hold balances, call one another and update a common state, while ETH priced the computation as gas.

That openness moved failure outward. A faulty application could lose assets even when Ethereum consensus worked as designed, and popular applications could compete for the same scarce block space. The platform's freedom and its hazards came from the same decision to execute other people's code.

In 2016, an attacker diverted more than 3.6 million ETH from The DAO into a child structure. The proposed recovery did not simply patch an application; it altered the state that future Ethereum clients would accept.

The Foundation's pre-fork note said no single entity could make the decision. The recovery code activated at block 1,920,000 on July 20 and directed DAO balances to a withdrawal contract, while opponents kept running the unforked rules. Exchanges, miners, users and developers supplied the names afterward: the recovery chain retained Ethereum, and the other history became Ethereum Classic.

The episode established a lasting governance fact. Smart-contract rules are executed by machines, but the software that defines canonical history is still adopted by people. A hard fork can coordinate a majority without making dissent disappear.

As applications filled blocks, first-price gas auctions made fees hard to predict. London activated EIP-1559 on August 5, 2021: blocks gained a protocol base fee that is burned, plus a priority fee users can offer to the block producer.

Burning made network use part of ETH's supply equation, but did not create a cap. Consensus rewards continue to issue ETH; activity burns varying amounts. Calling ETH permanently deflationary hides that the balance can reverse.

Beacon Chain began producing proof-of-stake consensus in December 2020 while the existing chain still used miners. For almost two years, Ethereum developed and rehearsed a replacement engine beside the live execution system.

On September 15, 2022, The Merge joined Mainnet execution to Beacon consensus. Balances, contracts and transaction history remained in place, so there was no new ETH to claim or swap. Validators took over block proposals and attestations, and Ethereum estimates that energy consumption fell about 99.95%.

The Merge did not let validators withdraw. Shapella supplied that missing exit on April 12, 2023, separating the consensus transition from the later release of staked principal and rewards.

Dencun followed on March 13, 2024 with EIP-4844 blobs. The change created temporary, separately priced data space for rollups; it lowered their publishing cost without pretending the base EVM could process unlimited transactions.

Pectra on May 7, 2025 and Fusaka on December 3 changed different boundaries of the same live system. EIP-7702 let an ordinary account delegate to executable code for features such as batching; the delegation persists but can be replaced or revoked, making the chosen code and authorization a new security boundary. EIP-7251 let validators consolidate up to a 2,048 ETH effective balance while retaining the 32 ETH entry minimum, and PeerDAS let nodes sample blob data rather than require every node to download every blob in full. None became an Ethereum rule merely because an EIP or client implementation existed; operators had to install compatible releases and continue on the activated chain.

The later BPO changes—from a blob target/maximum of 6/9 to 10/15 on December 9 and 14/21 on January 7, 2026—show the same process at smaller scale. Specifications define a candidate change, client teams ship it, testnets expose failures, and validators and node operators decide whether the running network adopts it. Ethereum’s continuity therefore comes from repeated coordination, not from a roadmap promise or a single institution’s command.