The Biggest Upgrade in Crypto History
Ethereum’s transition from proof-of-work to proof-of-stake — known as “The Merge” — completed on September 15, 2022. It was the blockchain equivalent of replacing a jet engine mid-flight. More than two years later, we have enough data to assess whether it worked. The short answer: yes, with some important asterisks.
Energy: The Unambiguous Win
The Merge reduced Ethereum’s energy consumption by approximately 99.95%. Pre-Merge, Ethereum consumed roughly 78 TWh annually — comparable to Chile. Post-Merge, that number dropped to about 0.01 TWh — comparable to a few thousand households. The Cambridge Centre for Alternative Finance, which tracked Ethereum’s energy consumption religiously, essentially discontinued their Ethereum index because the numbers became insignificantly small.
This is genuinely one of the largest decarbonization events in history — accomplished not through government mandate or carbon offsets, but through a protocol upgrade voted on by the community. Whatever you think of crypto’s environmental track record, eliminating 99.95% of a major blockchain’s energy consumption in a single day is an engineering achievement worth acknowledging.
Economics: The Staking Boom
As of early 2025, approximately 34 million ETH — roughly 28% of all ETH in circulation — is staked across more than 1 million validators. At current prices ($3,000-4,000 range), that’s $100-135 billion in staked value. Validators earn approximately 3-4% annual yield, coming from a combination of new ETH issuance (about 0.5% annually, significantly less than the 4%+ under proof-of-work) and priority fees/MEV from transaction processing.
The shift to proof-of-stake made ETH net-deflationary during periods of high network activity. The EIP-1559 upgrade (August 2021) burns a portion of every transaction fee. When network activity is high, the burned ETH exceeds the newly issued ETH to validators, reducing the total supply. Since the Merge, approximately 1.5 million ETH has been burned, reducing total supply by roughly 1.2%. This “ultrasound money” narrative has become central to ETH’s value proposition.
Liquid Staking and Centralization Concerns
Here’s the asterisk. Liquid staking protocols — primarily Lido — allow users to stake ETH and receive a liquid token (stETH) that can be used in DeFi while earning staking rewards. This convenience has made Lido dominant: as of 2025, Lido controls roughly 28% of all staked ETH, with its node operators collectively running about 30% of Ethereum validators.
This concentration raises legitimate concerns. If a single liquid staking protocol controls 33%+ of validators, it could theoretically disrupt the network (though economic incentives would make this self-destructive). More practically, Lido’s governance token (LDO) determines which node operators are included in its set, creating a governance attack surface. The Ethereum community has debated solutions — including staking ratio caps and changes to the protocol’s economics — but consensus has been elusive.
Coinbase, Binance, and Kraken collectively control another 10-15% of staked ETH, adding centralized exchange concentration to the mix. The concern isn’t abstract: OFAC-compliant blocks (which censor transactions from sanctioned addresses) have at times exceeded 30% of blocks, driven by regulated validators. The network continues to function, but the censorship resistance that is supposed to be crypto’s defining feature is being eroded by the very entities that make the network accessible.
Technical Performance
Post-Merge Ethereum has been remarkably stable. There have been no major consensus failures, no extended downtime, and no critical bugs in the proof-of-stake mechanism. Transaction throughput didn’t increase significantly (that’s the job of layer 2s, not The Merge), but block times became more predictable at exactly 12 seconds, and finality (the point at which transactions can’t be reversed) occurs in about 13 minutes — faster than the probabilistic finality under proof-of-work.
The real scalability improvements came from layer 2 rollups, which benefited from EIP-4844 (“proto-danksharding” or “blobs”) implemented in March 2024. This upgrade dramatically reduced rollup transaction costs by providing dedicated data storage space. As of early 2025, layer 2 networks (Arbitrum, Optimism, Base, zkSync, StarkNet) collectively process about 5-10x more transactions than Ethereum mainnet, with fees typically under $0.01 per transaction. The “modular blockchain” thesis — Ethereum as a settlement layer, rollups as execution layers — is working in practice.
Verdict
The Merge was a technical success, an environmental triumph, and a governance stress test that Ethereum passed. The centralization concerns around liquid staking are real and unresolved, but they’re manageable problems, not existential threats. Ethereum’s multi-year roadmap — The Merge, The Surge (scaling), The Scourge (MEV and censorship resistance), The Verge (statelessness), The Purge (state expiry), and The Splurge (miscellaneous improvements) — is ambitious to the point of hubris, but so far, the execution has been credible.