Ethereum (CRYPTO:ETH) co-founder Vitalik Buterin has described 2026 as a turning point for restoring computing self-sovereignty, beginning with changes to his own personal technology setup.
In a post shared on Friday on X, Buterin outlined how he has reduced dependence on centralised, data-driven platforms in favour of privacy-focused alternatives.
He said the shift reflects a broader effort to reclaim control over personal data, software, and digital autonomy from large technology firms.
Buterin highlighted two major changes made during 2025, starting with a near-complete switch to Fileverse, a decentralised and open-source document platform.
He described Fileverse as a privacy-preserving alternative to Google Docs that avoids centralised data storage.
The second major change involved moving decisively to Signal as his primary messaging application.
Signal provides default end-to-end encryption for individual and group conversations while retaining minimal metadata.
Metadata stored by Signal is limited to basic information such as account creation dates and last connection times.
Buterin contrasted this approach with Telegram, which only encrypts optional secret chats while storing most messages and metadata on its own servers.
Telegram’s data practices have faced increased scrutiny amid rising law enforcement requests in countries including France.
Looking ahead to 2026, Buterin said he has replaced Google Maps with OpenStreetMap through the OrganicMaps application.
He also confirmed a move away from Gmail in favour of Proton Mail, citing privacy and encryption benefits.
Decentralised social media platforms have become a growing priority in his personal technology stack.
Buterin discussed experimenting with locally hosted artificial intelligence models rather than relying on cloud-based services.
He argued that sending personal data to third-party AI providers is increasingly unnecessary as local hardware capabilities improve.
Sending all your data to third-party services is unnecessary when users can increasingly run AI on their own hardware.
Vitalik Buterin said.
He acknowledged that local AI tools still require better interfaces, integration, and efficiency to become mainstream defaults.
Buterin added that progress in local model usability has been significant compared with one year ago.