Article 5 of the NATO treaty — the collective defence clause that an attack against one member is an attack against all — has existed since 1949. Applying it to cyberspace has taken seven decades longer, primarily because the questions it raises are genuinely hard: What constitutes an attack? How destructive must the effects be? How definitively must attribution be established? The Vilnius Cyber Protocol, signed in March 2025 by all 32 member states, provides binding answers.

Under the protocol, a cyberattack that causes physical damage to critical infrastructure, results in loss of human life, or renders essential government functions inoperable for more than 72 hours is formally classified as an armed attack for Article 5 purposes. Attribution does not require the evidentiary standard of a criminal prosecution — a classified intelligence assessment agreed by a qualified majority of the North Atlantic Council is sufficient to trigger the consultation mechanism that precedes any collective response.

The response options available under the protocol are explicitly not limited to cyber means. The document states that the Alliance may choose to respond "by such action as it deems necessary, including in other domains" — language that military legal scholars have interpreted as preserving the option of conventional military response to a destructive cyberattack, a deliberate ambiguity designed to maximise deterrence.

Russia, China, and Iran have each publicly condemned the protocol as "escalatory." The United States Cyber Command has described it as "the most significant evolution in Alliance deterrence posture since the adoption of nuclear sharing arrangements in the 1960s."

"Deterrence only works if the adversary believes it," said NATO Secretary General Mark Rutte. "Ambiguity about whether we would respond has been an invitation. We have withdrawn it."