Vitalik Targets Five-Year Protocol Shrink for Ethereum (ETH) at ETH Taipei 2026

Vitalik's CROPS Framework in Taipei

Ethereum co-founder Vitalik Buterin told the ETH Taipei 2026 audience on September 13 that blockchains should be judged as tools rather than treated as destinations, framing the point around a value set he calls CROPS: Censorship and Capture Resistance, Open Source, Privacy and Safety. Under that lens, the proof-of-stake network he helped create is merely one of several technologies that can deliver those four outcomes, and teams should define the problem first before picking a chain. Much of what the industry celebrated in past cycles fails that test, he argued. He revisited the 2020-2021 bull market, when moving personal data, identities and social graphs on-chain was pitched as innovation, and NFTs were valued largely because they existed on a blockchain — joking that an identical NFT product launched by Twitter or Google would likely have been disliked by the same crowd. Smart contract finance drew similar scrutiny: markets trusted protocols branded DeFi even where oracles were highly centralized and cross-chain bridges were governed by multisig wallets that could harbor backdoors. Rising prices and a limited toolbox reinforced the habit, he said — when the only visible hammer is a blockchain, every problem starts looking like a nail. The full Ethereum picture he sketched goes well beyond philosophy.

The AI Choke-Point Problem

Buterin also assessed that the “crypto-first” mindset has effectively reached its limit this year: walk into a university, an Asian technology company or a large enterprise and ask which technology draws the most attention, and the answer is more likely to be artificial intelligence than blockchain. The problems crypto set out to solve have not disappeared, though — in his view they have sharpened. Workflows dependent on proprietary models from OpenAI or Anthropic carry a centralized single point of failure, since the provider can change policy, restrict access or halt service outright. Query content and usage logs can be analyzed, and users cannot easily confirm whether a next-generation model's training might retain sensitive inputs such as passwords. Deanonymization compounds the risk: AI can infer identity from writing style, metadata and posting times, while encrypted messengers like Signal still expose who contacted whom, how often and when. His provocative conclusion was that the most advanced “freedom tech” today may not be a blockchain at all, but open-weight large language models. Even so, he expects chains to endure — the real frontier, in his telling, is the fusion of blockchain with cryptography, from zero-knowledge proofs to fully homomorphic encryption and eventually indistinguishability obfuscation.

A Roadmap Grown Five to Ten Times

The protocol's own roadmap is becoming more ambitious, not less. Buterin said the development to-do list — previously planned out to roughly 2030 — has grown to around five to ten times its former scope, with AI-assisted formal verification keeping that acceleration safe by mathematically confirming generated code behaves as intended. On scaling, the next Glamsterdam hard fork, the kind of generational upgrade traced in our Ethereum 2.0 guide, is expected to make a higher gas fee ceiling safer, support parallel verification and rework gas pricing to blunt denial-of-service vectors, while EIP-8288 aims to cut the cost of zero-knowledge proofs and quantum-safe signatures — a defense theme explored in depth by Optimum's co-founder Médard. zkEVM integration is planned over the next three years, expanding computation without state and node burdens ballooning, and individual transaction proofs are to be recursively aggregated until a block needs just one proof — a trend he called the “proof singularity,” extending to signature aggregation across tens of thousands of validators. Privacy work includes stronger transaction inclusion guarantees and the Kohaku access-layer project, whose demo already routes payments through Tornado, Privacy Pools or Railgun. Long term, he wants RISC-V to serve as the instruction set for both user proofs and protocol logic, with the EVM demoted to an intermediate representation — and he hopes the protocol starts getting smaller within five years. Readers tracking the market in real time can follow live spot and futures prices on Gate.

Five Years to a Leaner Ethereum

Read together, the threads of the Taipei keynote form one arc: CROPS defines the ends, and the roadmap merely serves them. In our reading, the notable shift is that Ethereum's most aggressive technical commitments — recursive proofs, a higher gas limit, cheaper quantum-safe signatures via EIP-8288 — are explicitly justified by censorship resistance, privacy and safety rather than throughput for its own sake. That inverts the “put it on-chain” logic of prior cycles and hands developers a test any design can fail. For teams choosing between a Layer 2 and a centralized database, the question is no longer whether a blockchain is involved, but whether it measurably advances those four values.