Buterin said Ethereum was moving from a system in which all nodes download and re-execute every operation toward an architecture based on data sampling, SNARK verification and parallel computation. He said the transformation over the next three years could be large enough to make Ethereum a qualitatively different type of technology.

Although Ethereum is often described as a blockchain by analogy with Bitcoin, Buterin said the technology had changed significantly over the past 15 years. Ethereum now supports general-purpose computation, Proof-of-Stake, onchain applications using zero-knowledge proofs, and layer-2 networks for scaling and privacy.

The future architecture would use a flexible computing model designed to balance scalability and generality, several forms of multi-proposer block construction, optimized Proof-of-Stake and zero-knowledge proofs built into the base layer, according to Buterin.

Comparing Ethereum’s architecture in 2015, 2025 and 2030, he said almost every fundamental component had undergone or would undergo substantial changes. Block verification is shifting from downloading and re-executing data toward PeerDAS and SNARK verification, while consensus has moved from Proof-of-Work to Proof-of-Stake and is expected to undergo further optimization. Block production would move from a single miner to multi-proposer construction.

Changes for Ethereum users

Buterin said Ethereum would preserve 100% availability, transaction irreversibility and guaranteed execution according to programmed rules. The network is also expected to gain stronger censorship resistance through FOCIL, which would provide guaranteed real-time transaction inclusion.

He said specialized operations would carry significantly less overhead than executing a large amount of computation within a single sequential transaction. Although latency within Ethereum itself would remain higher than on centralized servers, infrastructure surrounding the network could offer much lower latency without changing the blockchain’s fundamental properties.

Users would still need to operate their own nodes to obtain the strongest guarantees, Buterin said, but the requirements for running those nodes should decline significantly.

Parallel computation and decentralization

Buterin identified the ability to distribute computation more efficiently among network participants as a key architectural change. In a classic blockchain, every unit of data and computation must pass through a single sequential processing model. Under the proposed architecture, operating costs would depend on how computation is organized.

Well-isolated dependencies could run in parallel or be filtered out before a transaction enters the final block, potentially affecting how decentralized applications are structured. Buterin said onchain data could eventually consist mainly of information directly related to state changes and their ordering, while other computation would be aggregated before inclusion in a block.

Modern cryptography makes it possible to verify the results of distributed work without trusting individual committees, according to Buterin. He said this could remove one of the main constraints faced by earlier efforts to scale decentralized systems through distributed work.

A “cryptographic world computer”

Buterin also discussed the possible development of obfuscation, including indistinguishability obfuscation, or iO. He said a viable form of iO could reduce the trade-off between privacy and generality, while weaker forms could have practical uses such as encrypted mempools. Most of the other changes he described would occur well before such technology emerged, he added.

Buterin described the resulting system as a “cryptographic world computer” combining blockchains, cryptographic privacy and verification, and powerful decentralized off-chain components.

Technical challenges remain, including improving the efficiency and security of zero-knowledge proofs and managing and parallelizing access to large volumes of network state. Buterin said the Hegota fork planned for next year would probably be Ethereum’s last “ordinary” upgrade containing features familiar to network developers since 2015.

He said subsequent priorities would include recursive STARKs, automated formal verification, optimized consensus algorithms and quantum security. In his view, PeerDAS has already started Ethereum’s transition from a simple blockchain system toward a more complex architecture, with the next stage focused on making cheaper, more scalable and more private computation the network’s main development direction.

Source: Incrypted