Ethereum no longer uses proof of work to decide which blocks extend the chain. The change is called the Merge. It is history: the protocol rules changed on dates that are already in the record. Treating it as a rumor, a company rebrand, or a promise about the price of ether confuses a consensus change with a market story.
Two chains, then one system
Proof of stake did not appear on the morning of the Merge as a new network that discarded the old one. The Beacon Chain launched on 1 December 2020 and ran proof of stake in parallel, without handling the execution of ordinary Ethereum transactions. The execution layer continued under proof of work. The Merge was the moment those pieces were joined: the Beacon Chain began choosing the blocks of the execution layer.
Accounts, contracts, and transaction history stayed. Ethereum.org’s description of the Merge is explicit that the upgrade changed consensus and did not expand the execution capacity of the chain by itself. Gas, contract bytecode, and the account model remained the subjects they were. Anyone who woke up to a different token address or a request to “migrate ETH to the new chain” was not looking at the protocol upgrade.
The dates and the trigger
The upgrade was specified across consensus and execution documents, including EIP-3675, which proposed the Merge, and EIP-4399, which replaced difficulty with a previous beacon-block root in the execution header. The consensus-layer upgrade named Bellatrix activated on the Beacon Chain at epoch 144896, on 6 September 2022. The execution-layer upgrade named Paris followed.
Paris was not scheduled by a clock time alone. It triggered when the proof-of-work chain first reached a terminal total difficulty of 58750000000000000000000. Ethereum.org records that this happened at block 15537393 on 15 September 2022. The next execution block, 15537394, was the first produced under proof of stake. Those are the figures to check against a node or against the project’s own history page, not a round number from a secondary recap.
What changed in the way blocks are chosen
Under proof of work, miners competed to find a header hash under a target. Under proof of stake, validators who have deposited 32 ETH propose and attest to blocks. The protocol assigns that work and applies penalties, including slashing for specified violations. The economic idea is different: the cost of misbehavior is bound up with a deposit that the protocol can destroy, rather than with electricity already burned on a losing header.
The project’s own energy claim should be quoted at the size it is made. Ethereum.org’s Merge page says energy consumption fell by about 99.95 percent because the proof-of-work competition ended. Its energy page separately cites a CCRI estimate of a reduction of more than 99.988 percent, and it publishes a current network estimate that moves as nodes join and leave. Those are statements about electricity. They are not statements that fees fell by the same percentage, that ether became “official,” or that validators cannot refuse a transaction. A validator set can still decline to include something. Inclusion rules are a different subject from the energy statistic.
What the Merge did not settle
Issuance changed because miners no longer receive a proof-of-work subsidy, and validators are paid under the proof-of-stake rules. This article does not pin a dollar value or a current issuance rate on that change. Both depend on definitions and on later parameter choices. Withdrawals of staked ether were a later upgrade, Shanghai and Capella, in April 2023. They were not part of the Merge itself. Reading “the Merge let everyone out” is a calendar error.
The Merge also did not turn ether into a security or a commodity by itself. Those are legal classifications in particular jurisdictions, decided under statutes and by regulators and courts, not by a difficulty total. A technical history can say what the protocol did on 15 September 2022. It should stop before it pretends to be a legal ruling.
For a reader checking old writing, the useful test is mechanical. Does the text claim that ETH had to be swapped, that a new chain replaced the state, or that proof of work is still how Ethereum produces blocks? Those claims do not match the record. The state continued. The consensus mechanism did not.
What the sequence of upgrades actually separated
The Merge was not one undifferentiated weekend. The Beacon Chain had already been producing blocks since 1 December 2020, after the deposit contract reached the minimum validator count. Altair, in October 2021, was a consensus-layer upgrade while proof of work still ran the execution layer. Bellatrix, at epoch 144896 on 6 September 2022, prepared validators for the terminal total difficulty and turned inactivity penalties up to their full values. Paris, on the execution side, was the block after total difficulty crossed the threshold. A history that says “Ethereum became proof of stake in 2022” is directionally right and still too coarse for anyone checking a block number.
Withdrawals were a later, separate upgrade. Ethereum.org’s history lists Capella, the consensus-layer half of the Shanghai and Capella upgrade, on 12 April 2023. Until that upgrade, a validator’s deposit was locked by the protocol even though the chain was already proof of stake. “Staked” after the Merge did not mean “free to leave.” It meant deposited under the rules then in force. Readers who treat every 2022 announcement as the day withdrawals opened are merging two upgrades.
Gas fees were not what the Merge rewrote. Ethereum.org says this in so many words: the Merge was a change of consensus, not an expansion of execution capacity, and it was not designed to lower fees. Fees move with demand for block space. Later upgrades, including blobs for rollup data, are the ones that changed how data is priced. Attributing a fee chart in 2024 or 2026 to the Merge is a chronology error. The energy chart is the chart the Merge itself claims.
One more boundary is worth keeping because it shows up in old instructions to holders. Nobody had to send ether to a new address, sign a migration, or redeem a new token. The account state, the contracts, and the history of transactions stayed on the same chain. A message that said otherwise was not a description of Paris. It was a request for keys or for a transfer, and those requests are outside the protocol change this article is about.
The block numbers are the cleanest way to keep the story from drifting. Block 15537393 is the last execution block under proof of work, the one at which terminal total difficulty was met. Block 15537394 is the first execution block under proof of stake. Epoch 144896 is Bellatrix, nine days earlier, not the Merge itself. 1 December 2020 is the Beacon Chain’s start, not the day execution moved. 12 April 2023 is Capella, the day withdrawal credentials became usable, not the day staking began. A paragraph that sorts those four dates correctly is already a better history than a headline that treats “the Merge” as a single undifferentiated switch.
Sources
- Ethereum.org, The Merge, including the energy figure of about 99.95 percent.
- Ethereum.org, Ethereum energy consumption, including the CCRI estimate.
- Ethereum.org, history, including Bellatrix, Paris, and the terminal total difficulty.
- Ethereum.org, The Beacon Chain.
- EIP-3675 and EIP-4399.




