The reform of Germany’s intelligence agencies

By Matthew Parish, Associate Editor

Thursday 13 August 2026

Germany is contemplating something close to a revolution in the constitutional architecture of its intelligence services. On 12 August 2026 the federal cabinet approved draft legislation intended substantially to enlarge the powers of the Bundesnachrichtendienst (BND), Germany’s foreign intelligence service, and the Bundesamt für Verfassungsschutz (BfV), the federal domestic intelligence service. The proposals must still pass through the Bundestag and may change substantially in that process. Nevertheless, their direction is unmistakable. Germany wants intelligence agencies that do more than listen, watch, analyse and report. She increasingly wants agencies capable of interfering with hostile operations before they cause damage.

This is an altogether more significant development than an administrative reorganisation. Germany has traditionally maintained unusually strict legal boundaries around her intelligence agencies. Those boundaries are not accidents of bureaucratic history. They represent constitutional scar tissue — institutional precautions designed in the aftermath of two German dictatorships in which secret intelligence, political surveillance and coercive governmental authority became instruments of repression.

The question now confronting Berlin is whether the Federal Republic has constructed its intelligence laws so carefully against the dangers of the twentieth century that they are inadequate against those of the twenty-first.

The three German intelligence services

Germany’s federal intelligence architecture is divided principally between three organisations. The BND collects foreign intelligence and reports ultimately to the Federal Chancellery. The BfV, under the Federal Ministry of the Interior, is concerned with threats to the constitutional order inside Germany — including political extremism, terrorism, espionage and foreign interference. The Bundesamt für den Militärischen Abschirmdienst (BAMAD), or Military Counterintelligence Service, performs the corresponding counterintelligence and constitutional-protection functions within the Bundeswehr.

This fragmented arrangement is deliberate. German constitutional culture has traditionally been suspicious of concentrating intelligence, police and coercive powers in a single institution.

The origins of this philosophy can be traced directly to 1949. The Western Allied military governors authorised the emerging Federal Republic to establish an organisation collecting information about subversive activities, but expressly withheld police powers from it. The BfV itself describes the resulting Trennungsgebot — the principle of separation between police and intelligence services — as having been created because a new Gestapo was to be prevented by all means. Intelligence officers could gather and evaluate information; police officers could arrest people and exercise coercive powers. The two functions were deliberately institutionally separated.

The historical reasoning is difficult to dispute. The Gestapo was not merely an intelligence-gathering institution. It combined surveillance with coercive authority within a dictatorship in which meaningful judicial and democratic controls had disappeared. East Germany subsequently supplied Germans with a second lesson. The Ministry for State Security — the Stasi — constructed an extraordinarily intrusive system of surveillance precisely because intelligence gathering was subordinated not to an independent constitutional order but to preservation of the ruling regime. Both the Gestapo and the Stasi illustrate the dangers that arise when secret surveillance, coercion and political power become fused without meaningful external control.

Hence modern Germany developed a paradoxical conception of intelligence. The democratic state unquestionably needed secret services — not least because the Federal Republic was situated on the central front of the Cold War — but those services should remain constrained, divided and subject to elaborate legal limitations.

For much of the Federal Republic’s history, this worked tolerably well.

Then the world changed.

From intelligence gathering to intelligence operations

The reforms now being advanced by Chancellor Friedrich Merz’s government challenge one of the assumptions underlying the traditional model: that an intelligence agency should primarily collect information while someone else acts upon it.

The proposed BND reforms would move the service towards a more operational conception of intelligence. The government has been discussing allowing substantially more intrusive access to electronic devices and communications, longer retention of acquired information and, most strikingly, active measures intended to disrupt hostile operations. Earlier proposals publicly discussed in February included interference with hostile drone operations and disabling cryptocurrency wallets associated with threatening activities.

The legislation approved by the cabinet goes considerably further in principle. The BND would acquire expanded authority to penetrate computer systems and interfere actively with hostile infrastructure. Cyber operations would therefore cease to be exclusively about discovering what an adversary is doing. In appropriate circumstances Germany could disrupt what the adversary is doing.

This distinction is enormous.

Imagine that German intelligence discovers a foreign intelligence network conducting a cyberattack against German infrastructure. Under the traditional intelligence philosophy, the BND’s essential task is discovering the network, identifying its controllers, determining its intentions and informing the German government and other competent authorities.

The emerging philosophy asks another question: why merely watch?

An intelligence service that has penetrated hostile infrastructure might be authorised to interfere with it — disabling systems or otherwise frustrating the operation. According to reports on the cabinet proposals, the contemplated concept of active countermeasures can extend beyond conventional electronic surveillance into technical disruption of hostile operations.

Germany would thereby move somewhat closer to the operational traditions of intelligence services maintained by several of its principal allies.

The digital intelligence state

There is a second dimension to the reforms, and potentially an even more consequential one. Modern espionage takes place inside computers, smartphones, communications platforms, databases, algorithms and interconnected domestic electronic devices.

A legal regime built around telephones, letters, human agents and physical surveillance inevitably encounters difficulties when intelligence targets inhabit encrypted and distributed digital environments.

The proposed reforms therefore contemplate substantially expanded powers to obtain and process electronically stored information. They also envisage increased use of automated analysis and artificial intelligence. The argument is straightforward: intelligence agencies confronted by immense quantities of digital information cannot reasonably be expected to analyse everything manually.

Yet precisely here the constitutional problems become most acute.

Traditional surveillance is expensive. Following somebody requires personnel. Listening to conversations consumes analysts’ time. Opening and examining documents takes effort. Digital surveillance changes the economics of government observation. Machines can collect, compare, classify and search quantities of information vastly greater than any human intelligence bureaucracy could previously process.

Artificial intelligence magnifies that change again.

Hence an intelligence power that appears superficially equivalent to an older investigative technique may be qualitatively different when exercised automatically against enormous datasets. The question is no longer merely whether the state may examine information. It is how much information, concerning how many people, using what computational methods, for how long — and according to what threshold of suspicion.

That explains why German intelligence reform cannot sensibly be reduced to the proposition that spies need better computers.

It concerns the constitutional limits of machine-assisted state surveillance.

The courts enter the intelligence world

Germany’s constitutional jurisprudence has already forced legislators repeatedly to confront these questions. Intelligence legislation exists beneath the Basic Law rather than outside it, and German constitutional law has developed particularly sophisticated doctrines concerning privacy, telecommunications and informational self-determination.

This has made reform legally complicated. Germany cannot simply copy the operational practices of another intelligence service and transplant them into German law.

Nor does German constitutional protection stop automatically at the frontier. The Federal Constitutional Court’s important 2020 judgment concerning BND strategic telecommunications surveillance held that the Basic Law’s fundamental-rights guarantees constrain German state authority even when the BND conducts telecommunications surveillance involving foreigners abroad. The judgment required the legislature to construct a constitutionally adequate framework around foreign surveillance, including safeguards concerning proportionality and independent oversight.

That principle carries profound implications. Foreign intelligence cannot simply be declared a constitutional wilderness in which ordinary German legal principles cease to operate.

The dilemma is therefore structural. Every time Parliament makes the BND more capable, it must simultaneously make the legal architecture governing that capability more sophisticated.

The price of history

Critics of the current proposals are consequently not irrational privacy absolutists. They are invoking one of the organising ideas of the post-war German state.

If an intelligence service can secretly enter computer systems, conduct intrusive electronic surveillance and actively interfere with perceived threats, the distinction between intelligence collection and executive action becomes less obvious.

That is precisely why the Trennungsgebot matters.

The BfV explains the traditional principle in remarkably simple terms: intelligence agencies collect and analyse information while law-enforcement measures belong to the police. Germany nevertheless already operates elaborate structures through which intelligence and law-enforcement organisations exchange information, particularly in counter-terrorism. The system has therefore never required officials to inhabit hermetically sealed bureaucratic boxes. Rather it attempts to preserve a legal distinction between knowing something secretly and exercising coercive state power because of that knowledge.

The proposed reforms place pressure upon that distinction.

The government’s response is essentially that effective oversight must evolve alongside effective intelligence. The reforms envisage strengthened or consolidated independent supervision, including an important role for the Independent Control Council. The political bargain is therefore becoming clearer: more intrusive and operational intelligence powers must be accompanied by more credible legal authorisation and scrutiny.

Whether the eventual legislation gets that balance right is likely to determine not merely its political legitimacy but its prospects before the constitutional courts.

Why now?

The immediate explanation for the reform is Russia.

Russia’s full-scale invasion of Ukraine in February 2022 produced Germany’s celebrated Zeitenwende in defence policy. The consequences for intelligence policy took longer to emerge. Germany discovered that military rearmament without corresponding intelligence capabilities addresses only part of the European security problem.

The contemporary confrontation between Russia and Europe does not fit neatly into the distinction between war and peace upon which much twentieth-century security law was implicitly based. Cyberattacks, espionage, disinformation, political influence operations, suspicious activity around critical infrastructure and sabotage can occur without armies crossing Germany’s borders.

This is the domain now generally described as hybrid warfare.

German officials have repeatedly argued that the security environment has changed fundamentally. In April 2026 the government said that profound changes to the BND were considered important precisely because the security situation had altered so substantially. In July it reiterated that intelligence law was to be comprehensively reformed to strengthen national sovereignty and operational capabilities and enable German agencies to keep pace with important European partners.

That last point reveals another uncomfortable feature of the debate.

Germany depends heavily upon allied intelligence.

Intelligence sharing is inevitable among allies, and no serious western state gathers everything itself. But dependence is different from cooperation. The Merz government increasingly appears to believe that Europe’s largest economy cannot indefinitely outsource important elements of its understanding of security threats to Washington, London, Paris and other partners. German officials have explicitly identified excessive reliance upon foreign intelligence as a weakness.

The changing transatlantic relationship only makes that calculation more important. Strategic sovereignty requires the ability to know things independently.

Drones, sabotage and the grey zone

Recent events have accelerated the argument.

The discovery of a drone carrying explosives at Leipzig/Halle airport has become an especially potent symbol of the sort of threat German officials believe the existing system handles inadequately. Reports surrounding the reform have also emphasised continuing cyberattacks, suspected foreign interference and surveillance of sensitive locations.

The significance of such incidents lies partly in their ambiguity.

A conventional military attack has an obvious character. A missile is launched, a border is crossed or soldiers open fire. Hybrid operations are designed precisely to make attribution, classification and response more difficult. A mysterious drone, an unexplained cyber intrusion, a damaged cable or an intelligence operative working through intermediaries may generate uncertainty about who is responsible and what legal response is permissible.

That uncertainty favours the aggressor.

An intelligence organisation confined principally to observation may therefore discover hostile activities without possessing sufficient authority to frustrate them quickly. By the time intelligence has travelled through bureaucratic channels to another authority possessing executive powers, the opportunity may have disappeared.

The proposed reforms are an attempt to close this gap.

Russia is not the only problem

China has likewise transformed German thinking about espionage. German security authorities increasingly describe foreign intelligence activity as involving not merely conventional theft of state secrets but technology acquisition, political influence, economic espionage and penetration of strategically important institutions.

Iran presents yet another category of challenge, involving espionage, intimidation and security threats associated with Tehran’s foreign-policy objectives. The government therefore describes the problem more broadly in terms of hostile foreign powers rather than Russia alone.

Domestic extremism further complicates the picture.

Germany’s constitutional order incorporates the concept of wehrhafte Demokratie — a democracy capable of defending itself against political movements seeking to destroy the democratic constitutional system from within. That principle gives the BfV a role quite different from that of an ordinary criminal investigative body. It watches constitutional threats before they necessarily become conventional crimes.

This creates an unavoidable democratic tension. Waiting until violent extremists commit offences may be waiting too long. Yet allowing secret agencies excessively broad powers to monitor political activity risks damaging the very democratic freedoms they exist to protect.

Germany has encountered this dilemma before. The discovery of the National Socialist Underground exposed serious failures in the domestic intelligence apparatus and produced intense controversy about informants, information sharing and institutional competence. The post-9/11 period likewise generated significant changes to Germany’s security architecture, while the later NSA surveillance controversy pushed political opinion in the opposite direction by demonstrating the potential scale of modern electronic intelligence collection.

The BfV itself identifies 9/11, the NSU affair, the NSA revelations and digital transformation as defining developments in its recent institutional history.

Germany has therefore spent the past quarter-century oscillating between two anxieties: that its intelligence services know too little and that they may know too much.

The AfD problem

There is also an unusually sensitive political dimension. The rise of the Alternative für Deutschland has required the German state to consider how constitutional intelligence powers should apply to a political movement with substantial electoral support. At the same time German authorities have become increasingly concerned about Russian attempts to influence German public opinion and political life, including activity directed towards political extremes.

This is exceptionally dangerous constitutional territory.

A democracy must be able to investigate genuine foreign espionage and clandestine influence. Yet an intelligence service must never become an instrument through which the government secretly disadvantages lawful political opposition. The distinction must depend upon evidence and legally defined threats rather than political convenience.

That is another reason why stronger intelligence powers require stronger independent supervision.

The ultimate test of a democratic intelligence system is not whether one approves of the people being watched today. It is whether one would trust the same legal powers in the hands of a government one profoundly disliked tomorrow.

Germany discovers the intelligence paradox

The Merz government’s reforms consequently expose an intelligence paradox.

A secret service weak enough never to threaten civil liberties may also be too weak to protect them.

But an intelligence service powerful enough to protect a democracy can acquire capabilities that, improperly controlled, could themselves threaten democratic government.

There is no statutory formula capable of making this contradiction disappear. The solution lies instead in institutional design — precise legal thresholds, judicial or genuinely independent authorisation where appropriate, parliamentary scrutiny, auditability, restrictions upon retention and dissemination of information and enforceable remedies when the state exceeds its authority.

Germany’s historical instincts therefore remain valuable. What may have become obsolete is not suspicion of secret governmental power but the assumption that restricting intelligence agencies primarily to passive collection necessarily represents the safest constitutional solution.

Sometimes failing to act carries constitutional consequences as well.

If a foreign state can disable infrastructure, penetrate government networks, conduct sabotage or manipulate democratic institutions while German agencies merely document what is happening, an impeccably restrained intelligence service becomes of limited comfort.

A second Zeitenwende

The first German Zeitenwende after Russia’s invasion of Ukraine concerned tanks, artillery, ammunition, defence expenditure and the rehabilitation of military power as a legitimate instrument of German statecraft.

The second may concern secrets.

For decades Germany could inhabit an unusually comfortable strategic environment. NATO provided military security, the United States supplied extraordinary intelligence capabilities, European integration reduced continental conflict and post-Cold War optimism made the restrictive traditions of German security law appear comparatively inexpensive.

That environment has disappeared.

Russia is fighting the largest European war since 1945. Cyber operations ignore national borders. Cheap unmanned systems make sophisticated disruption available to states and non-state actors alike. China possesses intelligence capabilities commensurate with her emergence as a superpower. Iran conducts operations far beyond its immediate neighbourhood. Artificial intelligence is transforming both surveillance and counter-surveillance. Germany simultaneously faces extremist challenges within its own political system.

The extraordinary feature of the proposed reforms is therefore not that Germany is abandoning the lessons of its past.

It is that two lessons of German history have collided.

The first says that secret state power is dangerous and must be fragmented, legally constrained and independently supervised.

The second says that democratic states unable or unwilling to defend themselves can perish.

Both propositions are true.

Germany’s task is consequently not to choose between the Gestapo and impotence — however tempting such melodramatic alternatives may be to participants on either side of the political debate. It is to construct a modern intelligence service capable of operating aggressively against hostile foreign powers while remaining institutionally incapable of becoming a secret police.

That requires something considerably harder than granting the BND and BfV additional powers. It requires designing those powers with sufficient precision that intelligence officers can act rapidly against genuine threats while judges, parliamentarians and independent supervisors can subsequently determine exactly what they did, upon what evidence and pursuant to which legal authority.

The proposed legislation is therefore best understood as part of Germany’s continuing Zeitenwende. The Federal Republic is discovering that strategic adulthood carries uncomfortable consequences. One of them is that a state wishing to remain sovereign must sometimes operate secretly against people who mean it harm.

Germany spent the decades after 1945 learning how to distrust its spies.

The challenge of the 2020s is learning how to make them effective without forgetting why that distrust existed in the first place.