Ethereum 101

Bitcoin is a ledger that does one thing extremely well: track who owns bitcoin. Ethereum started from a different question — what if the ledger could run programs, not just record payments? Everything confusing about Ethereum follows from that one change, and everything it is criticised for is the price of it.

This page assumes you have the basics of how a blockchain works. If not, Bitcoin 101 covers them, and almost all of it carries over.

The one idea: programmable money

A Bitcoin transaction says “move this amount to that address”. An Ethereum transaction can say that, or it can say “run this code”. The code lives on the chain, every node executes it, and they all reach the same answer — which is what makes the result trustworthy without a company standing behind it.

Those programs are called smart contracts, which is a poor name: they are neither smart nor legal contracts. A smart contract is a program with a balance, deployed to an address, that runs exactly as written whenever someone calls it. It cannot be argued with, and — unless its author deliberately built in a way to change it — it cannot be amended.

That is the whole appeal and the whole danger in one sentence. Nobody can quietly alter the rules of a lending market or a token sale. Nobody can fix an obvious bug afterwards either. There is no clause for “this is clearly not what we meant”, which is why contracts get audited before they hold serious money and why the failures, when they come, are total rather than partial.

The EVM: one machine, thousands of copies

The Ethereum Virtual Machine is the runtime every node uses to execute contract code. Because every node runs the same instructions over the same state, they all arrive at the same result independently — that agreement is the network's answer, not a report from any one machine.

The practical consequence is that Ethereum is slow and expensive by the standards of an ordinary computer, and that is not a flaw being optimised away. Thousands of machines redundantly executing the same code is the redundancy you are paying for. If you only needed the computation done, a server would be enormously cheaper; you are paying for the property that no one operator can lie about the outcome.

Gas: paying for computation, not for value moved

Bitcoin fees scale roughly with how much block space a transaction occupies. Ethereum charges by how much work it asks the network to do, measured in gas. A plain transfer costs a fixed small amount; a contract call that touches a lot of state costs much more. Moving a fortune can cost less than a fiddly interaction moving nothing.

Gas prices are quoted in gwei, a billionth of an ETH. Because block space is finite, the price rises when the network is busy and falls when it is quiet — the same action can cost very different amounts depending on the hour. Since 2021 a portion of each fee (the base fee) is destroyed rather than paid to a validator, so heavy use removes ETH from circulation.

Two things follow that catch people out. A failed transaction still costs gas, because the network did the work before hitting the error. And a wallet asking you to approve a “gas fee” on a page you did not expect is one of the most common shapes an attack takes — the fee is real, but what you are signing may not be what you think.

The Merge: Ethereum changed how it agrees

Ethereum originally used proof of work, like Bitcoin. In September 2022, in an upgrade known as the Merge, it switched to proof of stake — validators lock up ETH as collateral instead of miners spending electricity. It cut the network's energy consumption dramatically, and it was a genuinely difficult piece of engineering: a live network carrying a great deal of value changed its consensus mechanism without halting.

Running a validator requires 32 ETH. Smaller holders usually join a pool or stake through an exchange, which is convenient and quietly reintroduces the concentration the design was meant to avoid. Our Proof of Work vs Proof of Stake guide compares the two mechanisms properly, including the criticisms of each.

Layer 2: doing the work somewhere cheaper

Because space on Ethereum itself is scarce, most activity has moved to layer 2 networks — chains that process transactions separately and periodically settle a summary back to Ethereum, inheriting its security while charging a fraction of the fee.

The trade-offs are real and specific to each design: how long withdrawing back to Ethereum takes, who can censor or delay you in the meantime, and the fact that the bridge between the two is itself software that can fail. Bridges have been among the most profitable targets in crypto's history, so “it's on a layer 2” is a statement about cost, not a statement about safety.

ETH and BTC are different kinds of asset

This is the comparison that generates the most heat and the least clarity, so state it plainly: they are not competing versions of the same thing.

• Bitcoin has a fixed supply of 21 million, a deliberately conservative development culture, and one job. The case for it is monetary — scarcity guaranteed by rules that are extremely hard to change.
• Ethereum has no fixed cap. Issuance pays validators and a share of fees is burned, so net supply can rise or fall depending on how busy the network is. The case for it is utility — the value of the applications that need it — which makes it closer to a productive asset than to digital gold.

You can think one is a better investment than the other. What you cannot sensibly do is evaluate them against the same criteria, because they are not making the same promise.

What Ethereum is genuinely bad at

An honest reference page owes you this section. Ethereum is complex, and complexity is where money is lost: contract bugs, bridge exploits, and interfaces that ask you to approve permissions you cannot read. Fees remain unpredictable at the base layer. The roadmap moves, so what is true of the network changes more often than it does for Bitcoin — which is precisely the trade its community chose. And a large share of what gets built on it has been worthless, because the same openness that lets anyone deploy something useful lets anyone deploy anything at all.