Journalist
Hassan Shittu
Journalist
Share
Ethereum cofounder Vitalik Buterin has outlined a personal shift away from Big Tech platforms, framing 2026 as a pivotal year for what he calls “computing self-sovereignty.”
This is a concept that extends beyond blockchain and into how individuals use everyday software, communication tools, and artificial intelligence.
In a post shared on X, Buterin described a series of changes he has made across his devices to reduce reliance on centralized, data-intensive services.
Vitalik Buterin Replaces Google, Telegram With Privacy-Focused Alternatives
He claimed that the process started in 2025 when he transferred nearly completely to the open-source and encrypted and decentralized document platform Fileverse, which is purported to be a privacy-focused alternative to Google Docs.
At roughly the same period, he reported having changed to Signal as his main messaging application, abandoning Telegram.
Signal supports end-to-end encryption on all conversations by default and only retains a little metadata, whereas Telegram encrypts messages only in optional so-called secret chats.
He further stated that it otherwise stores messages and metadata on its servers, a design that has attracted increased attention due to an increase in law enforcement requests in some jurisdictions.
The changes made in 2026 included more extensive ones, which were due to those initial adjustments.
Buterin claimed to have substituted Google Maps with OpenStreetMap, with Organic Maps on mobile phones, citing the advantage of local and offline use that restricted the volume of location data sent to third parties.
He also left Gmail for Proton Mail, citing encrypted messaging as a more powerful tool to use for confidential communication.
Simultaneously, he claimed to have started giving priority to decentralized social media and still tests the idea of running large language models locally, as opposed to using cloud-based AI services.
Buterin Sees Local AI as the Future of User Privacy
The rationale of these decisions is not purity in ideology but practicality, as Buterin wrote.
According to him, it is not necessary to send big amounts of personal information to centralized services, as there are tools that can be used to minimize this exposure.
He admitted that local AI systems still have usability and integration issues, especially for translation, transcription, and document search, but noted that there has been a lot of improvement in the last year.
He explained a more ambitious long-term vision where local models are integrated with cryptographic schemes like zero-knowledge proofs, trusted execution environments, and local data filtering to restrict the information that ever leaves a user’s device.
Rising AI Demand Puts Self-Sovereign Computing Back on the Map
The remarks made by Butterin come amidst the wider revival of interest in self-sovereign computing, which is a model that highlights the importance of individualized controls of data, identity, and computing resources.
The idea integrates identity systems based on decentralization, personal servers, and privacy-by-design software with the aim of minimizing reliance on platforms that are controlled by corporations.
Privacy activists, including Naomi Brockwell, have long held the position that the most consistent approach to ensuring autonomy is to run software and AI models in place.
The time also stands out due to the re-evaluation of the strategic value of computing infrastructure by governments and corporations.
Analysts believe that 2026 will be the year of a change in the treatment of AI data centers, energy-backed compute and GPU capacity.
With the demand of large-scale AI continuing to outpace its supply, compute and energy are becoming viewed as a source of geopolitical power.
In that environment, the appeal of local and decentralized computing models has grown, particularly as concerns mount over surveillance, data residency, and platform dependence.