Bitcoin vs Ethereum: Built for Different Jobs

On 8 August 2026, at block height 961,632, Bitcoin nodes running software for a proposal called BIP-110 started rejecting every block that did not signal support for it. BIP-110 would have kept pictures, text and other non-payment data out of transactions for a year. Only 51 of the previous 2,016 blocks had signalled, 2.53% against the 55% it needed. So the enforcing nodes split onto a chain of their own, which managed two blocks in roughly eight hours while the main chain added about 50.

Everyone else carried on. Bitcoin's consensus rules have not changed since Taproot in November 2021, and that weekend showed again how hard they are to move.

Ethereum was built on the opposite assumption: running arbitrary programs, and storing whatever data those programs need, is the job it was designed for, and its rules change on a published timetable. It shipped two named upgrades in 2025 alone, and the next one, Glamsterdam, was being tested in September 2026.

Those two attitudes to change are a fair guide to the rest of the comparison. What follows takes the real differences one at a time, and none of it is financial advice.

What each was built to do

Bitcoin's genesis block was mined on 3 January 2009, and the system it started still does one main thing: it tracks a fixed-supply currency and lets people move it without a middleman. Its scripting language can demand two signatures instead of one, or lock coins until a future date. It cannot loop. According to the Bitcoin wiki, that limit is there on purpose.

Ethereum's whitepaper, written by Vitalik Buterin before the network launched, explained the reasoning (verifying a transaction must never get stuck in an infinite loop) and still counted the missing loops among Bitcoin's limitations. When Ethereum's mainnet went live on 30 July 2015 it carried a Turing-complete language instead, so a contract can run almost any program, provided someone pays gas for every step. On Ethereum, tokens, stablecoins, lending pools and NFT collections all run as contracts of that kind.

The same design choice reads as discipline from one side and as a missing feature from the other. Nobody got it wrong. The two projects were after different machines from the start.

A fixed ceiling and a moving total

Bitcoin's supply schedule was set before the first block. The reward for mining a block started at 50 BTC and halves every 210,000 blocks, roughly every four years, so issuance tapers towards a ceiling just under 21 million, due to be reached around 2140. The April 2024 halving cut it to 3.125 BTC. The 20 millionth bitcoin was mined on 9 March 2026, and about 20.09 million existed by 28 September 2026.

Ether has no ceiling, and new ether goes to validators, while the base fee on every transaction has been burned, removed from circulation for good, since the London upgrade of 5 August 2021. Around 4.64 million ETH had gone that way by late September 2026. Whether the total rises or falls depends on how much gets burned against how much gets issued.

Both have happened since the Merge in September 2022, when the supply stood at about 120.52 million. It shrank through 2023 as burning outpaced issuance and bottomed out near 120.06 million on 5 April 2024, three weeks after the Dencun upgrade made it far cheaper for layer 2 networks to post their data to Ethereum. Cheaper data meant less burning, and by 27 September 2026 the supply had climbed to about 122.08 million.

Anyone can work out how many bitcoin will have been issued by a given block height decades from now, but nobody can do that for ether.

Mining and staking

Bitcoin is still secured by proof of work. Miners run specialised machines that burn electricity for the right to add each block, and rewriting history would take more computing power than the rest of the network put together. Ethereum ran the same way until 15 September 2022, when the Merge swapped its miners for validators. Each validator locks up ether as collateral, 32 ETH to start, and loses part of it for breaking the rules.

By ethereum.org's estimate, the switch cut Ethereum's energy consumption by 99.95%. It also changed what an attack costs: against Bitcoin an attacker needs hardware and power, and against Ethereum they need a large stake of ether that the protocol can destroy.

Ten minutes against twelve seconds

Bitcoin aims for one block every ten minutes and adjusts its mining difficulty every 2,016 blocks, about a fortnight, to hold that pace. Ethereum keeps time in 12-second slots grouped into 32-slot epochs, so roughly 50 slots pass for every Bitcoin block.

Settlement differs as well: a Bitcoin payment never becomes final in any formal sense, because each block added on top only makes reversal less likely, and many exchanges wait for six, about an hour. Ethereum has an explicit finality step that takes about 15 minutes. Reverting a block after that, according to ethereum.org, would cost an attacker at least a third of all staked ether.

How the rules change

Bitcoin's rule changes arrive as soft forks, which only narrow what counts as valid, so nodes that never update still accept the new blocks. Even those are rare, and slow to arrive. Taproot was first described on a developer mailing list in January 2018, locked in during June 2021 once 90% of blocks in a two-week window signalled for it, and switched on at block 709,632 on 14 November 2021. Nothing has changed the consensus rules since, and that is the wall BIP-110 hit.

Ethereum changes by hard fork, where old software cannot validate blocks made under the new rules, so every node operator has to update before the upgrade goes live. Its own timeline at ethereum.org shows at least one named upgrade in every year from 2019 to 2025, with recent examples:

• London, on 5 August 2021, rebuilt the fee system and started the burn.

• The Merge (15 September 2022) ended mining on Ethereum.

• Dencun, 13 March 2024: blobs, a cheap home for layer 2 data that is deleted after about 18 days.

• Pectra went live on 7 May 2025, letting a single validator hold a stake of up to 2,048 ETH instead of 32, and letting ordinary accounts run smart-contract code.

• Fusaka followed on 3 December 2025, so that each node stores only a slice of the blob data.

The chain that refused

One Ethereum fork was on nobody's roadmap: on 17 June 2016 an attacker drained 3.6 million ETH from The DAO, a code-run investment fund, taking roughly a third of the 11.5 million that backers had committed. On 20 July, at block 1,920,000, Ethereum adopted a one-off change that moved the drained funds into a contract where the original depositors could withdraw them. Some miners refused, kept the untouched chain running, and it became Ethereum Classic.

Bitcoin's nearest parallel was a bug: on 15 August 2010 block 74,638 created more than 184 billion bitcoins through an overflow error, a patched client was out within five hours, and a corrected chain soon overtook the bad one. Those coins broke Bitcoin's intended rules, whereas the DAO contract had done exactly what its code said, flaw and all.

Ethereum Classic is still running: it kept proof of work when Ethereum moved to staking, caps its supply at about 210.7 million coins, and describes itself as combining the technology of Ethereum with the philosophy of Bitcoin. Same engine, different rules.