Layer 1

What makes a chain a layer 1

Three properties, and a chain needs all of them. It runs its own consensus, so block production and validation happen among its own participants. It holds the authoritative copy of its own state, so a transaction is settled when that chain says it is settled. And it funds its own security budget, out of issuance, fees, or both.

The third property is the one pitches skip and the one that costs money. A layer 1's security is inherited from nowhere. It is bought every block from validators or miners who could be doing something else with the same capital or the same hardware. That makes launching an L1 a decision about a permanent operating expense before it is a decision about technology, which is why the first honest question is what pays for it in year three.

The scalability trilemma, stated properly

Vitalik Buterin's own statement of the constraint rewards reading in full rather than in paraphrase. The trilemma says there are three properties a blockchain tries to have, and that if you stick to simple techniques you can only get two. Scalability: the chain can process more transactions than a single regular node, meaning a consumer laptop, can verify. Decentralization: the chain runs without trust dependencies on a small group of large centralized actors. Security: the chain can resist a large percentage of participating nodes trying to attack it.¹

The definitions carry teeth that summaries drop. Security is quantified, ideally at 50 percent of participants, with anything above 25 percent described as fine and 5 percent as definitely not fine.¹ Decentralization is defined by whether you can join without trusting a set of nodes you cannot join with a consumer laptop, not by a validator count on a slide. The post also sorts the easy answers: traditional chains where every participant verifies every transaction get decentralization and security but not scalability, while high throughput chains that lean on a few dozen nodes buy scalability by giving up the first one.¹ Read any L1 pitch against those three definitions and the traded corner becomes obvious.

Where the L1 and L2 boundary actually sits

The clean test comes from the rollups post. Instead of putting all activity on the blockchain directly, users perform the bulk of their activity off chain in a layer 2 protocol, and there is a smart contract on chain with only two tasks: processing deposits and withdrawals, and verifying proofs that everything happening off chain is following the rules.² Two jobs. That is the entire L1 surface of a genuine layer 2.

Ethereum's documentation states the security consequence directly. Rollups perform transaction execution outside layer 1 and post the data to layer 1, where consensus is reached, and because that data sits inside L1 blocks the rollup is secured by native Ethereum security.³

Sidechains fail that test, and the same page says why: they run under their own chosen rules of consensus and block parameters.³ A sidechain is a separate layer 1 with a bridge attached, and its users are trusting its validators rather than Ethereum's. Validium sits between the two, using validity proofs like a zk-rollup but keeping the data off the main layer 1 chain.³ When you write a risk disclosure, the sentence that matters is which validator set a user's funds actually depend on, and it is frequently not the one on the logo.

The modular endgame is already the plan

The architectural direction is public and it is not neutral. Ethereum's community favours rollup centric scaling over sharding the base layer, largely because blobs of rollup data can be verified efficiently by validators.³ The L1 gets deliberately conservative while execution moves upward.

Buterin's Endgame post follows that logic somewhere uncomfortable. What he describes as the destination is a chain where block production is still centralized, but block validation is trustless and highly decentralized, with anti-censorship mechanisms stopping block producers from censoring, and he calls the result somewhat aesthetically ugly.⁴ Sit with that for a moment. Decentralized block production is not the property being defended. Verifiability and censorship resistance are. A great deal of L1 marketing is still selling the property that was traded away, and once you have read the post you cannot unsee it.

What picking or launching an L1 commits you to

If you are choosing an L1 to build on, four things follow you into production. Finality behaviour, and what it does to the moment your product can credit a user. Fee volatility, and whether your unit economics survive a congested week. Validator or sequencer concentration, and whether it belongs on the risk register your investors read. And upgrade governance, because the rules underneath your contracts can change without your consent.

If you are launching one, the trilemma is your budget line. Name the corner you gave up, in the whitepaper, in your own words. Reviewers will find it regardless, and letting them find it first costs credibility you cannot spare. Then answer the security budget with a revenue model instead of an emission schedule, because a layer 1 with no application demand funds its own security through dilution for as long as it runs. That is a business problem wearing a protocol costume, and no consensus design fixes it.