Ethereum ETH

The review

Ethereum is the most ambitious live software project in the world, and reviewing it means confronting an uncomfortable fact: it keeps doing things that ought to be impossible. In September 2022 it replaced its entire consensus mechanism — the beating heart of a network securing hundreds of billions of dollars — while it was running, with user funds in flight, and the transition was so clean that most people found out from a headline rather than an outage. There is no comparable feat in commercial software. Airlines take systems offline for weekends to change a database schema. Ethereum swapped proof of work for proof of stake mid-flight and did not drop a block.

That single event tells you most of what our team score is measuring. Ethereum's contributors are not one company. They are multiple independent client teams, a research organisation that publishes its arguments in public, and a sprawl of implementers who disagree loudly and then ship compatible software anyway. Client diversity is not a slogan here; it is an operational reality that has repeatedly prevented a single implementation bug from becoming a network failure. When we score team and track record, we ask what happened the last time something broke. On Ethereum, the answer is consistently: multiple clients, rapid public post-mortems, and fixes that shipped without anyone needing to pause the chain.

The technology story since the Merge has been relentless. Shanghai enabled staking withdrawals, closing the one genuinely uncomfortable open loop in the proof-of-stake design and proving the exit path works at scale. Dencun introduced blob-carrying transactions, a dedicated cheap data lane for rollups that collapsed layer-two costs by orders of magnitude essentially overnight. Subsequent upgrades have widened blob capacity, improved validator economics, and steadily reduced the cost of running a node. Each of these landed on schedule-ish, in public, after months of testnet rehearsal, with clear rollback thinking. That is what mature protocol engineering looks like, and almost nobody else in this industry does it this way.

Tokenomics deserve careful attention because Ethereum's are genuinely unusual. Post-Merge issuance dropped by roughly ninety percent, since the network no longer pays miners to burn electricity. Simultaneously, the fee-burn mechanism introduced in 2021 destroys the base fee of every transaction. The result is a supply curve that responds to actual network usage: heavy demand burns more ETH than validators earn, and net supply contracts. Light demand allows modest inflation to compensate the validator set. We have never seen a monetary design that ties issuance to real economic throughput this cleanly, and critically, it is not discretionary. No committee sets the burn. It is a consequence of usage, enforced in consensus.

The validator set is the other half of that story. Nearly a million validators secure the chain, and the entry point is genuinely accessible — a solo staker with modest consumer hardware and a home connection participates on identical terms to an institution. That property is rare and worth defending. Yes, liquid staking providers hold a large aggregate share, and we watch that number closely; it is the one structural item on our monitoring list. But the protocol's response has been substantive rather than rhetorical: distributed validator technology, the growth of independent node-operator sets within those providers, and ongoing research into correlation penalties that make concentration economically punitive rather than merely frowned upon.

Where Ethereum's design decisions look most vindicated is the rollup-centric roadmap. Rather than scaling by raising base-layer throughput and pricing out home validators, Ethereum chose to become the settlement and data-availability layer for an ecosystem of execution environments. Today that ecosystem processes the overwhelming majority of the network's transactions, at fractions of a cent, while inheriting Ethereum's security through proofs posted to layer one. Optimistic rollups have matured their fraud-proof systems into live, permissionless operation. Zero-knowledge rollups have gone from research curiosity to production systems verifying general-purpose execution. The blob market Dencun created is now the economic engine underneath all of it.

For builders, nothing else is close. The EVM is the closest thing this industry has to a universal instruction set — implemented across dozens of chains, targeted by every serious toolchain, and understood by more auditors than any other execution environment. The consequence is a compounding safety advantage: the audited contract libraries, formal verification tooling, static analysers, fuzzing harnesses, and institutional knowledge about failure modes all concentrate here first. When we review a project on a niche runtime, we routinely find the same class of bug that Ethereum's ecosystem catalogued and tooled against years ago. That accumulated defensive knowledge is an asset Ethereum owns outright.

Real economic activity backs the abstraction. Stablecoin settlement on Ethereum and its rollups now moves volumes comparable to major payment networks. Tokenised treasuries and money-market funds from regulated issuers settle here because the audit trail is public and the finality is unambiguous. Lending markets that have survived multiple violent liquidation cascades without socialised losses continue to operate under the same contracts. Every one of those systems has been attacked, forked, copied, and stress-tested in production. The ones still standing on Ethereum are standing because the base layer never gave way underneath them.

Our honest reservations are about surface area, not soundness. A rollup-centric world means users navigate bridges, differing finality assumptions, and layer-specific quirks, and the UX work to hide that is genuinely hard — though account abstraction and chain-abstraction standards have made large strides. Staking concentration remains a thing to monitor rather than a thing to panic about. Neither of these is a protocol defect; both are ecosystem coordination problems being addressed in public with measurable progress each year.

Against our rubric, Ethereum returns maximum marks across the board. Technology: shipped behaviour matches specification through the most complex upgrade sequence ever attempted on a live chain. Team: multi-client, multi-organisation, no single point of capture, with public post-mortems every time. Tokenomics: fully disclosed, usage-responsive issuance with the original allocation long since vested and a burn mechanism nobody controls. Security posture: no admin key, no upgrade multisig, no pause switch — consensus changes require coordinated client releases and social consensus. Community: governance happens in open research forums and public calls, and it has rejected changes favoured by powerful participants more than once. Ten out of ten. Ethereum did not just pass our rubric; several of our rubric's questions exist because Ethereum answered them first.