Ethereum: Vitalik Buterin Wants to Import Utreexo, the Anti-Storage Gambit of Bitcoin

A borrowed idea, almost claimed. Last Sunday, Vitalik Buterin paid an unusual tribute to Bitcoin developers. On X, the co-founder of Ethereum credited them with the invention of Utreexo, a technique that drastically reduces what nodes need to store to verify a blockchain, and described what he calls the scalability strategy currently being considered for Ethereum. The core idea is to combine multiple state models rather than choosing just one. Nothing is finalized, Buterin himself clarifies, but the outlined direction deserves attention.
Key points of this article:

• Vitalik Buterin praised Bitcoin developers for the invention of Utreexo, a technique that lightens node memory.
• Buterin outlined a scalability strategy for Ethereum, inspired by Bitcoin's UTXO model, without abandoning the existing account model.

Utreexo: How Bitcoin Lightens the Memory of Its Nodes

Some context is necessary before diving into the heart of the matter. Utreexo was designed in 2019 by Bitcoin developer Tadge Dryja, as part of the MIT Digital Currency Initiative.
Instead of requiring each node to retain the entirety of all UTXOs (unspent transaction outputs, the unspent transaction outputs, the basic unit of Bitcoin's ledger), Utreexo represents it through a compact accumulator based on cryptographic hashes.
Each transaction carries its own proof of inclusion, verifiable without the node keeping all the data in memory. The structure thus grows logarithmically rather than proportionally to the number of outputs, according to the original MIT DCI research paper. A lighter Bitcoin node, therefore easier to run independently.
Ethereum Eyes the UTXO Model, Without Abandoning Its Accounts

However, this is not a new whim from the co-founder of Ethereum. On July 6, Ethereum Foundation researcher Toni Wahrstätter even published a proposal for native UTXOs for Ethereum on ethresear.ch, without removing the existing account model.
The idea targets simple payments, those that do not require the persistent state of a smart contract, and shifts them to a lighter format like Bitcoin. The estimated gain approaches 99.8% on this type of load: at one billion entries, the proposal mentions about 300 MB of permanent state, compared to 100 to 150 GB for the equivalent in traditional accounts.
A nuance must be made here, however. These are design estimates, not measurements obtained on the main network. The setup depends on EIP-8141, known as Frame Transactions, which currently only appears as an option in the official specification of Hegotá, far behind FOCIL (EIP-7805), the only proposal formally scheduled for this expected update in 2027.
Buterin is simultaneously working on a completely different bottleneck, that of proof bandwidth. His January research on a mempool aggregating recursive STARKs (cryptographic proofs that can themselves be combined) starts from proofs of about 128 KB and proposes to merge them periodically rather than attaching one to each propagated object. With eight peers and intervals of 500 milliseconds, the additional bandwidth would remain around 2 MB per second per node, without increasing even if the volume of objects rises. Some commentators have already seen in combining the two ideas an architecture capable of absorbing an unlimited volume of transactions. Nothing confirms this reading: these are two distinct research projects, addressing two different bottlenecks.
This convergence of ideas from Bitcoin fits into the broader logic of Lean Ethereum, the protocol overhaul initiated by Buterin, which aims to simplify rather than stack layers. Nothing is set on the Hegotá calendar for now, and client teams must first decide between the dozens of competing proposals already on the table.