Vitalik proposes a more sophisticated strategy: the "pause button" to control AI computing power

Ethereum co-founder Vitalik Buterin has responded to political pressures to limit the expansion of mega data centers dedicated to artificial intelligence, presenting an alternative vision to outright development bans.

A more nuanced solution than a simple ban

Instead of resorting to a total shutdown switch, Vitalik argues that the priority should be developing the technological capability to implement a sophisticated “pause button.” This mechanism would allow reducing available computational capacity by 90-99% in critical or emergency situations, providing granular and reversible control rather than an outright prohibition.

The context of Bernie Sanders’ proposal

Vitalik’s stance comes in response to a call from U.S. Senator Bernie Sanders, who has urged to slow down the construction of massive infrastructure facilities dedicated to AI data centers. The underlying concern involves the concentration of computational power and the potential geopolitical risks associated with AI technology control.

Decentralization as a viable path

A key element of Vitalik’s reasoning is the distinction between large-scale AI hardware used by mega clusters and consumer-level AI devices. From this perspective, promoting a more fragmented and distributed AI development ecosystem could serve as an effective compromise between innovation and risk management. A decentralized approach would naturally reduce dependence on a few giant computational hubs.

Vitalik’s proposal suggests that control mechanisms and safety buttons could represent a more realistic and flexible path than total bans, while still leaving room for distributed technological growth.