GRADUS

I. What Moscow Knows — and Bets On

The head of Germany's Federal Intelligence Service, Bruno Kahl, said it publicly: an analysis by German intelligence suggests that Russian defence ministry chiefs doubt whether Article 5 would in fact be invoked in the event of an attack on a NATO member. Not because they do not know the treaty — but because they assess its political reality precisely. They have watched how NATO members responded to thirty, forty, fifty provocations below the threshold: with disapproval, with investigations, with reinforced patrols, with declarations. Not with what Article 5 promises.

That is the actual strength of hybrid warfare: it does not test NATO's military capacity. It tests its political will. And political will is structurally weaker than military capacity — because it requires unanimity, because it must be publicly justifiable, because it needs a clear casus belli that hybrid operations systematically withhold.

In 2024, 34 documented Russia-linked sabotage attacks were recorded in Europe — four times the previous year's figure. In 2025 the number rose further. Arson attacks on logistics centres in Poland and Lithuania. GPS jamming in the Baltic region. Railway damage on routes carrying military aid to Ukraine. Drones over German military facilities. Cable severances in the Baltic Sea. Assassination attempts on European politicians and intelligence personnel. Each of these acts, viewed individually, remains below the threshold. In their totality, they are a war.

The problem is not that nobody knows this. The problem is that the knowledge has no consequences — because the architecture for cumulative response is missing.

II. The Mechanics of the Grey Zone

Article 5 of the North Atlantic Treaty presupposes an armed attack. The word dates from 1949 — when an attack was defined as a tank column crossing a border. Hybrid warfare is the precise opposite: it is constructed so that no individual act is recognisable as an armed attack, even though its cumulative effect corresponds to one.

The mechanics work on three levels:

Deniability. No official Russian soldier. No Russian state mark on the weapon. A paid dock worker in Hamburg who did not know for whom he was working. A criminal group in Warsaw recruited via Telegram channels. A ship under Panamanian flag without a transponder. The chain from the Kremlin to the act is real — but not immediately provable. And as long as it is not provable, no collective response can follow.

Fragmentation. 34 attacks in one year — against twelve different countries, at thirty different targets, using ten different methods. No single country experiences enough to cross the political response threshold on its own. Each interior minister, each intelligence chief sees their part of the pattern — but the pattern as a whole is only seen by those who assemble all the pieces. And they struggle to achieve the political coordination that a collective response would require.

Desensitisation. After the twentieth cable break, the thirtieth drone deployment, the fortieth arson attack, the public grows accustomed. What triggered alarm the first time becomes routine. What is routine does not justify an extraordinary response. The grey zone functions not only through individual actions — it functions through their accumulation, which lowers the perception threshold without lowering the political response threshold.

The result: Russia gains freedom of action through repetition. Each new act comes cheaper than the first — because it has become normal.

III. What Is Currently Happening — and Why It Is Not Enough

NATO and the EU are not inactive. Operation Baltic Sentry reinforces patrols in the Baltic Sea. The NATO Cooperative Cyber Defence Centre of Excellence in Tallinn coordinates cyber defence. The EU-wide Hybrid Fusion Cell collects intelligence. Several NATO members have concluded bilateral agreements on rapid response to sabotage acts. The Vilnius Summit of 2023 agreed in principle to treat certain cyberattacks on critical infrastructure as potential Article 5 cases — case by case.

The problem lies in those last three words: case by case. Every activation requires a new political decision with all 32 members. Every decision can be blocked. Every decision gives the attacker time — which it uses to calibrate the next act more precisely below the new informal threshold.

The BND chief identified the core problem: Russia doubts that Article 5 will be triggered. Not because it underestimates Europe's capacity, but because it knows its will. As long as these doubts are justified — and they are, as long as the mechanism remains as it is — the grey zone has no floor.

IV. GRADUS — The Concept

GRADUS stands for Graduated Response Against Documented Underhand Subversion. The Latin gradus means step, degree, grade. It precisely describes the logic of the concept: every documented step of the attacker counts. The sum of the steps determines the level of response. And the levels are defined in advance — publicly, bindingly, automatically.

The concept has three elements that together form a new deterrence architecture.

First element: The public hybrid register. All proven or attributable hybrid operations against NATO or EU members are documented in a public register — with date, type, target, degree of attribution and cumulative total. The register is not secret. It is readable by everyone: by the public, by parliamentarians, by journalists, by the attacker. Not every entry requires evidence beyond reasonable doubt — it requires the threshold of overwhelming probability defined by the NATO Hybrid Analysis Panel, which is publicly documented. Russia knows at any point where it stands on the counter.

Second element: Cumulative thresholds with predefined packages of measures. This is the core of the concept. GRADUS defines three thresholds — in advance, publicly, politically binding:

Threshold 1 — Documented pattern: When more than ten attributable hybrid acts against NATO members are documented within twelve months, Package A automatically enters into force: heightened diplomatic communication, intensified intelligence cooperation, public attribution of all documented acts, activation of the EU Anti-Coercion Instrument against specific economic ties of the attacking state. No political decision at the moment — the decision was taken when the threshold was agreed.

Threshold 2 — Systematic campaign: When more than thirty attributable hybrid acts are documented within twelve months, or when individual acts claim human lives or permanently damage critical infrastructure: Package B. In addition to Package A: coordinated counter-sabotage within the mandate of NATO cyber capabilities, ship interdiction rights under the SHADOW protocol, automatic activation of sanctions under the AGORA protocol against the attacker's financial sector, reinforced presence on the territory of the most affected NATO member.

Threshold 3 — Accumulated state of war: When documentation of the hybrid campaign reaches the degree of de-facto warfare against NATO members — defined by cumulative economic damage, endangerment of human lives or demonstrated coordination with military operations — the combination is treated as an Article 5 equivalent. Not because each individual act would be — but because the sum is.

Third element: The principle of collective authorship. GRADUS does not solve the deniability problem through better attribution — that is the goal of other concepts in the series. GRADUS solves it through a paradigm shift: it is no longer sufficient to be unable to prove the immediate perpetrator. When an attacking state systematically derives political or military benefit from hybrid operations without preventing them, it is liable for the campaign — regardless of the provability of individual acts. This is the logic by which states are made liable for the actions of proxy groups — not new in international law, but never consistently applied in NATO practice.

V. The Self-Punishment Structure — the Grey Zone Shrinks with Every Act

The Kessler Principle was physical: those who generate debris endanger themselves. GRADUS creates a political-legal parallel: those who commit a hybrid act increase the counter — publicly, irrevocably, with predefined consequences.

The decisive difference from current practice: today Russia calculates per act. It asks: is this attack large enough to trigger a response? The answer is almost always no. So it carries it out.

With GRADUS the calculation is different: every act — even the small one, even the deniable one — moves the counter. The thresholds are known. The packages of measures are known. The response is not politically negotiable once the threshold is crossed. Russia no longer needs to ask: does this act trigger a response? It must ask: how close am I to the next threshold?

That shrinks the grey zone from both sides. From above: because the aggregate response becomes calculable and therefore deterrent. From below: because the costs of individual operations now accumulate instead of dissolving into ambiguity.

There is also a direct self-punishment: a public register that systematically documents Russia's hybrid warfare and makes it internationally visible is a foreign-policy burden. Countries of the Global South that regard Russia as a neutral or even friendly actor read the register. Investors read it. Institutions read it. The costs rise with every entry — even before the response threshold has been reached.

VI. What GRADUS Is Not

GRADUS is not a declaration of war. The packages of measures are deliberately constructed to remain below the conventional war threshold — economic, diplomatic, intelligence-based, cyber. The response is proportionate to the provocation, not escalating beyond it.

GRADUS is also not an instrument to treat Russia in a permanent state of exception. The register counts acts — not intentions. As long as no acts are documented, there are no consequences. That is an incentive for restraint, not a permanent punishment.

And GRADUS does not replace Article 5. It extends it into the domain Article 5 does not cover. The tank doctrine remains — for tanks. GRADUS is for everything below.

VII. What Must Be Done

First: A NATO-EU hybrid register, operated by the NATO Hybrid Fusion Cell with involvement of the EU Intelligence Centre (IntCen). Publicly accessible, daily updated, with clearly defined attribution criteria. Not every entry requires criminal standards of proof — it requires the intelligence standard of overwhelming probability, which is publicly documented. Russia should see its running total.

Second: A political decision — one-time, in the NATO Council — on the three threshold values and the predefined packages of measures. The difficult political work is done now, not at the moment of crisis. The decision is irrevocable until formal review. This removes from the attacker the hope of being able to rely on ambiguity at the moment of crossing.

Third: Integration of GRADUS with the other concepts in the series. Threshold 2 automatically activates the SHADOW protocol for suspect vessels, AGORA measures against financial infrastructure and the DEMOS protocol for disinformation operations. The concepts reinforce each other — GRADUS is the trigger, the others are the tools.

Fourth: An EU legal basis for collective authorship in the hybrid context. EU law already knows state responsibility for proxy actions in other areas. For hybrid warfare, an analogous instrument must be created: if a state demonstrably derives benefit from a coordinated hybrid campaign without preventing it, it bears liability — regardless of the attributability of individual acts.

Fifth: Media strategy for the register. The register only unfolds its deterrent effect if it is received. Regular reporting on the state of the register — by EU institutions, not propaganda organs — creates the public visibility that anchors political self-commitment. When citizens know the counter stands at 28 of 30, the political cost of inaction when the threshold is crossed becomes too high.

Russia knows: Article 5 will not be triggered.
It knows this because it is true.
GRADUS makes it no longer true.
Not through threat — through mechanism.