NATO’s Theory of Victory: From Ambition to Warfighting Reality. PART 2- The Battlefield Is Trying to Disconnect You

Eva Sula, Defence & Security Leader | Strategic Advisor | NATO & EU Innovation | True deterrence is collaboration (Tallinn, Harjumaa, Estonia)

1. The battlefield is trying to disconnect you

Part 1 ended with a basic dependency in NATO’s emerging Theory of Victory. Sensors have to find and identify. Information has to reach decision-makers. Decisions have to become tasks. Tasks have to reach effectors, and the results have to return into the system quickly enough for the next decision. The architecture becomes powerful because those elements can operate as a coherent force rather than as isolated platforms. That also creates something for an adversary to attack.

Russia does not have to destroy every NATO sensor, headquarters, aircraft, launcher or uncrewed system individually. It can seek to weaken the relationships between them. Electronic warfare can interfere with communications, radar and navigation. Cyber operations can disrupt services or compromise confidence in them. Physical attack can damage telecommunications, power generation, data centres, satellite ground infrastructure and command posts. Undersea cables and terrestrial fibre can be damaged. Commercial services on which military users depend can become unavailable. Spectrum can become congested even without enemy action, while deception can make functioning sensors and networks carry information that is technically valid but operationally wrong.

These are not hypothetical categories assembled for a future war. NATO describes electromagnetic warfare as extending from attacks on radar to jamming communications and navigation, electronic reconnaissance, masking and intelligence collection across every operational domain. In November 2025, NATO’s Electronic Warfare Advisory Committee pointed directly to the widespread use of EW in Ukraine and the Middle East as evidence of its growing operational importance.

Cyber attack has already intersected directly with battlefield communications. Approximately one hour before Russia launched its full-scale invasion of Ukraine on 24 February 2022, a cyberattack disrupted Viasat’s KA-SAT satellite communications network. The UK, EU and United States subsequently attributed the operation to Russia. The primary target was assessed to be the Ukrainian military, but the consequences spread beyond it: commercial internet users were affected and thousands of wind turbines in Central Europe lost remote communications. Viasat reported tens of thousands of terminals damaged or rendered inoperable. The case is useful because military, commercial and civilian dependencies were not neatly separated when the attack occurred.

Physical infrastructure creates similar dependencies. NATO launched Baltic Sentry in January 2025 after a series of incidents involving critical undersea infrastructure in the Baltic Sea, including energy and communications cables. The Alliance has since expanded surveillance, integrated national assets and increased cooperation with private-sector infrastructure operators. NATO’s response itself illustrates the underlying problem: connectivity ultimately depends on physical assets located somewhere, owned by somebody, maintained by somebody and vulnerable to disruption.

Power belongs in the same discussion. Communications nodes, sensors, data processing, charging infrastructure, repair facilities and defence-industrial production all require energy. NATO has explicitly linked reliable energy supplies to both military forces and increased defence production, while warning about cyberattacks and attacks against critical infrastructure. A highly distributed digital force may reduce dependence on individual headquarters or platforms while simultaneously creating thousands of smaller dependencies on electricity, batteries, generators, charging, networking and computing.

The important distinction is that disconnection is not a binary condition. A system does not move neatly from connected to disconnected. Bandwidth can fall while the link remains active. Latency can make information arrive too late for the decision it was intended to support. GNSS can be jammed or, more dangerously, spoofed. One communications path can disappear while another remains available at lower capacity. A cyber compromise can leave a service running while creating uncertainty about whether its information should be trusted. Sensors can continue reporting while deception, interference or poor timing makes the resulting picture incomplete or misleading.

An obvious outage at least tells the commander that something has failed. A degraded system may continue to present itself as functional. The headquarters still receives tracks. The map still updates. Messages still arrive. The operational question becomes whether they are complete, current and trustworthy enough to support action. A force that assumes connectivity because an interface remains green can therefore be in a worse position than one that knows it has lost the network and shifts deliberately to a degraded mode of operation.

NATO is already testing this problem. Allied Command Transformation’s Digital Backbone Experimentation 2026 is examining network access when connections are weak, disrupted or unavailable, including in environments affected by jamming, electronic warfare and attacks on communications infrastructure. Significantly, ACT describes success in terms of whether forces can continue to command, coordinate and exchange priority information under degraded conditions, rather than simply whether engineers can preserve a connection.

That is the more useful standard for the Theory of Victory. The requirement cannot be permanent connectivity, because Russia will actively contest it and ordinary battlefield friction will degrade it even without successful enemy action. The requirement is a force that understands which connections are essential, which functions can tolerate delay or reduced capacity, which can continue locally, and when a degraded information environment requires a different decision process.

The first dependency to examine is one so deeply embedded in modern military and civilian systems that its absence can affect far more than navigation. Position, navigation and timing have become infrastructure. On NATO’s eastern flank, Russia is already demonstrating what happens when that infrastructure becomes contested.

2. GNSS: when position itself becomes contested

Position, navigation and timing are easy to reduce to a discussion about whether a drone can still find its target without GPS. The dependency is considerably broader. NATO’s space policy describes PNT as essential to all NATO missions because it provides precise positioning and synchronisation across military operations. GNSS consequently supports not only navigation, but also timing, communications, intelligence collection, precision fires, air and maritime operations, logistics and the coordination of geographically dispersed forces.

The first distinction is between jamming and spoofing. Jamming interferes with reception of satellite signals strongly enough that the receiver can no longer obtain reliable GNSS information. The operational symptom may therefore be relatively obvious: accuracy deteriorates, satellites disappear or the receiver reports that navigation is unavailable. Spoofing is different. False signals cause the receiver to calculate an incorrect position or time while it may continue appearing to function normally. NATO identifies both communications and navigation jamming as electromagnetic-warfare activities, while its civil aviation work separately warns about deliberate transmission of false GNSS signals.

That distinction matters because losing information and trusting false information create different command problems. A unit that knows GNSS is unavailable can switch procedures, use inertial navigation, terrain references or other sources, increase separation margins and reduce reliance on precision. A unit receiving plausible but false PNT may not immediately know that anything is wrong. Position errors can contaminate reports, tracks and targeting information. Incorrect timing can affect systems that depend on synchronisation. The danger is therefore not merely that navigation stops. It is that apparently functioning systems begin operating from a false reference.

Europe already has substantial operational evidence of both problems. NATO reported in 2025 that intentional GNSS interference was becoming more frequent and sophisticated across the Baltic, Black Sea and Mediterranean regions. EASA updated its GNSS safety bulletin again in July 2026 following analysis of recent jamming and spoofing incidents, adding new operational and training requirements and procedures for coordination between pilots and air traffic control. This is no longer an exceptional wartime phenomenon confined to Ukrainian airspace. European civil aviation is having to build routine procedures around the possibility that satellite navigation may be unreliable.

The Baltic and Nordic environment is particularly relevant to NATO. In July 2026, NCIA, the Finnish Defence Forces and Finnish Border Guard conducted operational trials of GANDALF, a NATO-developed system intended to detect, classify and locate sources of GNSS interference. Testing took place aboard the Finnish Border Guard offshore patrol vessel Turva and an H215 Super Puma helicopter. NCIA simultaneously evaluated controlled-reception-pattern antenna technology intended to allow navigation systems to continue operating during jamming. The exercise matters because NATO is not treating interference simply as something receivers should endure. Detecting what is happening, distinguishing the interference and locating its source are themselves becoming operational requirements.

The dependency extends well beyond aircraft and uncrewed systems. Precision fires need accurate coordinates and timing throughout the targeting chain. Maritime forces use satellite navigation alongside other navigation sources. Logistics systems depend upon location and timing to coordinate movement across large distances. Digital communications and networks can depend on precise timing for synchronisation. Civil transportation, telecommunications, energy and emergency services create further dependencies that matter to military operations because NATO relies extensively on civilian infrastructure for reinforcement and sustainment. NATO’s 2026 resilience requirements explicitly identify GNSS among the technological vulnerabilities that civil transportation systems must be prepared to withstand, including during prolonged and complex crises.

This is why the phrase “GNSS-denied capable” deserves considerably more interrogation than it usually receives. Denied how? Was one GPS frequency jammed, or were multiple frequencies affected? Was only GPS unavailable, or were Galileo and other constellations also denied? Was the receiver exposed to broadband interference, directional jamming or spoofing? For how long? Was the platform stationary, moving slowly or manoeuvring dynamically? Did inertial navigation bridge the outage for seconds, minutes or hours, and how rapidly did its positional error accumulate? What accuracy remained at the end of the test, and was that accuracy sufficient for the mission?

The duration question alone can radically change the meaning of a test. EASA’s current operational guidance recognises inertial coasting and other sensors as potential mitigations during GNSS jamming, while requiring operators to plan explicitly for periods in which GNSS navigation data is lost. Alternative navigation therefore does not make the problem disappear. It changes the rate at which performance degrades and the period for which the mission remains viable.

The hard question for NATO is consequently not whether the Alliance can operate without GPS in some abstract sense. It is which functions merely become less accurate or slower when assured PNT disappears, and which functions eventually cease to produce a usable military effect. That assessment needs to include the dependencies hidden inside other systems, including civilian infrastructure supporting military movement and sustainment.

It also needs to be exercised over meaningful periods. Surviving a ten-minute GNSS outage demonstrates something useful. It does not demonstrate that a formation can conduct operations for twelve hours, three days or several weeks while interference moves geographically, spoofing is introduced and alternative systems themselves become targets. Resilient PNT therefore requires more than an alternative navigation device. It requires forces to recognise degradation, determine what information remains trustworthy, understand how quickly errors are accumulating and know when the resulting uncertainty changes what they are authorised or able to do.

GNSS is only one part of that problem. Once military forces begin transmitting, sensing and communicating in order to compensate for lost information, they enter another contested environment. The electromagnetic spectrum is not simply the medium through which the connected force operates. It is terrain over which both sides are already fighting.

3. The electromagnetic spectrum is terrain

Modern forces use the electromagnetic spectrum to see, communicate, navigate, identify, target and protect themselves. Tactical radios, radar, satellite communications, drone control links, datalinks, electronic protection and many sensors depend upon it. NATO consequently treats the electromagnetic environment as part of the battlespace and electromagnetic warfare as a means both to enable friendly operations and to prevent an adversary from doing the same. It spans reconnaissance and intelligence gathering, attacks against radar, communications and navigation, electronic masking and defensive measures across every operational domain.

For commanders, the important problem is simpler than the engineering underneath it. Every transmission can create capability and vulnerability at the same time. A radio transmission allows a unit to coordinate. A datalink moves targeting information. A drone control signal allows an operator to manoeuvre a weapon or sensor. Radar reveals objects that cannot be seen visually. Yet electromagnetic energy can also be detected, classified, direction-found, disrupted or used to help locate the equipment and people producing it.

Ukraine has made the consequences unusually visible. Lessons collected by the US Army from Ukrainian Territorial Defence Forces recommend separating antennas from command posts, restricting Wi-Fi and cellular use and controlling radio emissions because Russian electronic warfare can contribute to locating command nodes. Ukrainian guidance cited by the Army warns particularly about high-power communications and stresses the need to balance reliable communications against the electromagnetic signature they create. The operational problem is therefore not simply whether a headquarters can communicate. It is whether it can communicate enough to command while remaining difficult enough to find and target.

That trade-off becomes harder as NATO pursues a more connected force. High-bandwidth communications can move imagery, sensor data and targeting information quickly enough to compress decision cycles. Persistent connections can improve shared situational awareness and allow geographically dispersed capabilities to cooperate. But the communications architecture also produces patterns. Transmissions have frequencies, power levels, locations, durations and behaviours. US Army work on signature management describes electronic equipment as producing invisible signatures that specialised sensors can detect, with greater emissions increasing the risk of targeting.

The response cannot simply be to stop transmitting. Silence protects a formation from one form of detection while depriving it of capabilities it may need to fight. Nor can commanders assume that one technically sophisticated waveform solves the problem permanently. The electromagnetic environment is interactive. Friendly forces alter frequencies, power, antennas, transmission schedules and waveforms. The adversary observes those changes and modifies its collection and jamming. The result is an adaptation cycle rather than a fixed technical contest.

Ukraine provides repeated examples. When Russian counter-UAS systems jammed commonly used commercial frequencies and navigation signals, Ukrainian operators shifted towards less common frequencies and altered power. Russian forces adapted their spectrum analysis and jammers to identify the new frequencies. Ukrainian forces then employed frequency-hopping and other techniques intended to make suppression more difficult. Russian forces have similarly modified their systems, producing an environment in which equipment configured around a fixed set of frequencies can become obsolete quickly.

Drone warfare has pushed the cycle further. Greater onboard autonomy can reduce dependence on a continuous operator link during parts of a mission. Alternative navigation can reduce dependence on GNSS. Neither makes a system immune to electronic warfare, but both change which link the adversary must attack. CSIS found that electronic warfare remains a persistent constraint on Ukrainian drone operations and that the pressure is contributing to development of functions such as terminal navigation and other forms of partial autonomy when communications become unreliable.

Fibre-optic FPV drones represent another adaptation rather than the end of the contest. A physical fibre connection between operator and aircraft removes the radio-frequency command link that conventional jamming would otherwise attack. That can make the drone particularly useful in heavily jammed areas, but it exchanges one set of constraints for another. Fibre length limits range, the spool adds weight and the physical cable affects how the system can be employed. Meanwhile, Russian forces have increasingly used fibre-optic FPVs alongside artillery, glide bombs and other systems to attack Ukrainian UAV operators, electronic-warfare positions, radar and command posts. RUSI documented this increasingly systematic effort during 2025.

There is another complication for NATO: friendly use of the spectrum can interfere with friendly operations. Jamming intended to suppress hostile drones can also affect friendly uncrewed systems, communications or navigation if spectrum management and coordination are inadequate. NATO-Ukraine lessons published in 2026 explicitly identify friendly interference as a risk when EW is used as part of air defence. A multinational force operating large numbers of radars, radios, drones, jammers and datalinks therefore has a spectrum-management problem before the adversary transmits anything at all.

This changes what commanders need to understand about communications. The question cannot be reduced to whether a unit has coverage. They need to know what the transmission reveals, how long it can safely persist, what capacity is actually necessary for the mission, what alternatives exist and what happens when the preferred waveform becomes unusable. They also need enough awareness of friendly electromagnetic activity to understand whether apparent enemy interference is actually congestion or fratricide.

The resulting trade-off is uncomfortable. The communications method that gives NATO the greatest speed, bandwidth and connectivity may also create something Russia can detect, classify, locate, jam or eventually target. Reducing the signature may mean transmitting less often, using less bandwidth, changing routes, distributing nodes or accepting slower information flow. US Army analysis of Ukraine reaches precisely this problem: low-probability-of-detection techniques improve survivability, but they can restrict the amount of information that can be passed.

That makes electromagnetic discipline a command problem rather than an EW-specialist problem. Commanders have to decide what information genuinely needs to move, when it needs to move and whether the operational benefit of transmitting it outweighs the exposure created by doing so. NATO’s connected force therefore cannot be designed around maximum connectivity at all times. It needs the ability to vary its electromagnetic behaviour according to threat, mission and available alternatives while continuing to command and fight.

And even when the signal survives, it still has somewhere physical to go. Radios connect to networks, networks depend on infrastructure, satellite services require ground equipment, data processing consumes power, and high-capacity communications eventually encounter cables, servers and facilities. The electromagnetic spectrum may be invisible terrain, but much of the connected force that depends upon it is very physical indeed.

4. There is no cloud in the sky

“Cloud” is useful computing terminology, but poor physical intuition. NATO’s digital architecture does not exist somewhere above the battlefield. Every cloud service ultimately runs on processors inside physical facilities. Every connection eventually depends on radios, fibre, satellites, ground stations, terminals or terrestrial telecommunications. Those systems require electricity, cooling, maintenance, software, credentials and people. The more military capability depends on digital services, the more important it becomes to understand where those services physically live and what else they depend upon.

NATO’s Digital Transformation Implementation Strategy describes the Digital Backbone as a federation of networks and systems providing a resilient and secure continuum of digital services, including cloud and edge computing, connecting sensors, decision-makers, actors and effectors across national, organisational, operational and security boundaries. The Alliance Digital Strategy deliberately adopts a hybrid approach, combining federated cloud services with tactical edge computing and next-generation networks so that operations can continue in degraded, contested and denied environments. That architecture is intended to avoid dependence on one centralised system. Federation, however, distributes dependencies rather than abolishing them.

Consider satellite communications. SATCOM gives deployed forces connectivity where terrestrial infrastructure is unavailable or contested and is essential to NATO command and control. NATO does not depend upon a single satellite system: its current arrangements combine national military capabilities across several frequency bands, while some NATO systems also use commercial services. NCIA operates seven satellite ground stations and a satellite centre supporting these services. The architecture therefore extends from spacecraft in orbit to terrestrial antennas, terminals, cryptographic equipment, terrestrial backhaul, network operations and the personnel maintaining them. Losing access to the satellite is only one way to interrupt the service. Ground infrastructure and user terminals matter just as much.

Commercial SATCOM can substantially increase capacity, geographic coverage and diversity. NATO is actively expanding its relationship with commercial providers, and NCIA is establishing arrangements that allow additional commercial capacity to be acquired quickly when operational demand increases. But NATO’s own Commercial Space Strategy explicitly warns against overreliance on any single provider or solution and calls for arrangements that ensure access during peace, crisis and conflict. That distinction matters. A service that can be purchased during an exercise is not automatically a service whose availability, priority, cybersecurity, terminals, support and contractual conditions have been secured for a major war.

Terrestrial communications create another layer. Fibre offers enormous bandwidth and is unaffected by radio-frequency jamming along the cable itself, which makes it highly valuable in a contested electromagnetic environment. It remains physical infrastructure. Cables cross borders, enter landing stations and exchange points, connect data centres and can be cut accidentally or deliberately. NATO launched Baltic Sentry in January 2025 following disruption to critical undersea infrastructure in the Baltic Sea, precisely because energy and communications links beneath the sea have become a security problem rather than merely a commercial one.

The same physical reality applies inside data centres. NCIA describes NATO data centres as hosting information essential to daily operations and providing operational networks across the NATO Command Structure. Moving a service into a cloud environment can make computing more flexible and allow workloads to be distributed, but the underlying facilities still require power, cooling, connectivity, hardware replacement, physical protection and cyber defence. Edge computing moves some processing closer to tactical users and can reduce dependence on constant reachback. It does not eliminate infrastructure either. The edge device still needs power, storage, software, security credentials and eventually some mechanism for synchronisation and resupply.

Civil infrastructure is inseparable from this problem. NATO’s 2026 Resilience Baseline Requirements require Allies to plan for reliable civil communications during crisis and conflict, including physical and digital risks, supply-chain dependencies, foreign ownership and cascading disruption across energy, transport, health care, emergency services and military support. That is an important acknowledgement of how modern collective defence actually works. Military networks do not sit beside civilian society in a sealed architecture. They use infrastructure, services, energy, transport and industrial capacity drawn from the same societies NATO is defending.

This is where apparently independent backup systems need closer examination. A military headquarters may have terrestrial fibre, commercial mobile connectivity and SATCOM and reasonably describe them as three communications paths. But resilience depends upon what sits underneath them. Two routes may enter the same facility, depend upon the same electricity distribution, use the same identity service, pass through the same terrestrial exchange or rely on terminals whose software is supported through the same supply chain. Three visible alternatives can therefore conceal one shared point of failure.

The same applies at Alliance scale. A cloud service may be geographically distributed while authentication remains concentrated. Communications may have several carriers while electrical power depends upon vulnerable local infrastructure. Commercial satellite constellations may provide alternative routes while their gateways, network-management systems or user terminals create different concentrations of risk. There is no inherently resilient medium. Fibre, radio, SATCOM, commercial telecoms, cloud and edge computing offer different strengths and different failure modes.

The operational question is consequently not whether NATO has sufficient connectivity options. It is where the Digital Backbone physically lives, which dependencies those options share, and what happens when several are attacked at the same time. NATO’s current experimentation reflects that problem. Digital Backbone eXperimentation 2026 is testing resilient communications across multiple networks and how information can continue moving when individual paths degrade, rather than assuming one preferred network remains continuously available.

That is the standard a wartime architecture needs to survive. Redundancy should be measured by independent failure paths, not by the number of communications logos on an architecture diagram. Cloud and edge services should be assessed together with the facilities, energy, networks, credentials and personnel that keep them functioning. Commercial capacity should be incorporated deliberately, with its wartime availability and dependencies understood in advance.

Even then, maintaining the connection solves only part of the problem. A network can remain powered, reachable and apparently functional while the information moving through it has been altered, stolen, delayed or made deliberately difficult to trust. The next question is therefore not simply whether NATO remains connected, but whether commanders can still believe what the connected system is telling them.

5. Connected, compromised, or confidently wrong?

Availability is only one measure of whether a military network is functioning. A communications system can remain online while the information moving through it has become incomplete, altered, stolen or untrustworthy. For a force increasingly designed around rapid data exchange, sensor fusion and digitally supported decisions, that distinction matters. Losing a network creates an obvious operational problem. Continuing to use a compromised network can create a less visible one because the force may continue making decisions from information whose integrity it can no longer establish.

Cyber operations illustrate the difference. An attacker does not necessarily need to destroy a system to obtain military effect. Stolen credentials can provide legitimate-looking access. Compromised administrator accounts can expose multiple systems. Software or configuration changes can alter how information is processed. Intelligence can be extracted without interrupting the service at all. An attacker with persistent access can also choose when disruption becomes most useful rather than announcing its presence immediately.

Russian operations against Ukraine provide documented examples of several parts of this spectrum. Microsoft attributed the January 2022 WhisperGate attacks against Ukrainian organisations to a Russian GRU-associated actor and later documented the same actor using stolen credentials, legitimate administrative commands and web shells to maintain access inside Ukrainian government and IT-provider networks. WhisperGate itself was destructive malware disguised as ransomware, but the broader campaign demonstrates why cyber compromise cannot be reduced to systems being visibly switched off. Access, persistence, intelligence collection and eventual disruption can occur at different stages of the same operation.

For NATO, this becomes particularly important because the Digital Backbone is explicitly intended to connect data sources, services, decision-makers and effectors across domains and levels of command. NATO’s 2026 Alliance Digital Strategy consequently places Zero Trust at the centre of digital resilience, requiring continuous verification and least-privilege access rather than assuming that a user, device or service should remain trusted because it has already entered the network. The strategy also requires mission-critical services to be identified, protected and supported through diversified alternatives.

Zero Trust helps address access and identity. It does not answer every question about whether the information itself remains reliable. NATO’s Data Strategy makes that distinction explicit by defining trusted data as data for which users and systems can assess the integrity of the data and the processes surrounding it. Its Data Quality Framework goes further, calling for monitoring, anomaly detection, conformance testing and remediation throughout the data lifecycle. That is significant because information can become operationally dangerous without being fabricated by an adversary. It can also be stale, duplicated, incorrectly labelled, stripped of context or combined with information generated at different times and confidence levels.

The command problem becomes harder when deliberate manipulation and ordinary data-quality failures produce similar symptoms. Imagine a headquarters receiving a track from several sensors. One source places it several kilometres from another. A third has stopped updating. An intelligence feed identifies the object differently. The network itself is available, authentication succeeds and every system is technically operating. The commander still has to decide whether the discrepancy results from sensor limitations, latency, configuration error, electronic deception, cyber compromise or a real change on the battlefield.

The answer cannot simply be to distrust everything whenever cyber activity is suspected. That would allow an adversary to create operational paralysis merely by generating uncertainty. Nor can the force continue treating all authenticated information as reliable. Authentication can establish where a message apparently came from and whether particular security conditions were satisfied. It does not automatically establish that the originating sensor was correct, its software uncompromised, its coordinates current or its interpretation accurate.

This creates a requirement for provenance, although commanders do not need to use the technical vocabulary to understand the operational need. They need to know where important information originated, how old it is, whether it has been transformed, what corroborates it and what confidence should reasonably be placed in it. When those conditions change, the system must communicate uncertainty rather than presenting increasingly questionable information with the same visual authority as well-supported information.

The consequences vary with the decision. An uncertain track may be perfectly adequate for directing another sensor to investigate. It may be inadequate for firing a long-range weapon. A degraded logistics picture might justify dispatching a reconnaissance team but not rerouting an entire reinforcement column. Trust therefore cannot be treated as a universal percentage attached to a dataset. The required confidence depends on what the commander intends to do with the information and the consequences of being wrong.

This also changes cyber incident response. The immediate instinct after discovering compromise may be to isolate systems, revoke credentials or shut down services. Those actions can protect the network while simultaneously removing capabilities from an ongoing operation. NATO’s Digital Transformation Implementation Strategy describes cybersecurity as a foundation for systems that are not only secure but also trusted and mission-ready. In war, those requirements can conflict temporarily. A commander may need to continue operating a degraded service while cyber specialists investigate it, restrict certain functions while preserving others, or shift decisions to alternative information sources until confidence is restored.

The difficult question is therefore not simply whether NATO can keep its networks running under cyber attack. It is what a commander does when the network is functioning but its information may no longer be trustworthy. That requires predefined procedures for changing confidence, corroborating critical information, restricting particular actions, identifying authoritative sources and restoring trust. It also requires exercising the uncomfortable middle ground between normal operation and complete network loss.

A system that fails obviously can trigger a contingency plan. A system that remains fast, connected and visually convincing while some of its inputs have become unreliable presents a different risk: it can preserve the appearance of information superiority after the underlying evidence has deteriorated. Part 3 will examine that problem at the level of sensors, data and machine-supported classification. For the connectivity problem here, the requirement is narrower but fundamental. NATO must be able to continue operating not only when information is scarce, but when the degree to which it can be trusted is itself contested.

6. Graceful degradation is a warfighting requirement

Resilience is sometimes described as the ability to keep a system available. That is too narrow for a force operating against an adversary deliberately attacking communications, navigation, data, infrastructure and trust. Some services will disappear. Others will deteriorate. Bandwidth will change, information will become older, particular routes will fail and units will periodically lose access to capabilities normally provided from elsewhere. The operational requirement is therefore graceful degradation: losing parts of the architecture without losing the ability to continue the mission.

NATO’s own digital strategy increasingly reflects this requirement. The 2026 Alliance Digital Strategy calls for mission-critical services to be identified, prioritised and protected, with redundant and diversified solutions based on the PACE principle: Primary, Alternate, Contingency and Emergency. It also requires cloud and tactical-edge computing to support operations in degraded, contested and denied environments. The important word is not simply “redundant.” It is “mission-critical.” NATO cannot protect every digital service equally, and commanders need to know which ones must survive for the formation to continue fighting.

That requires understanding what happens when each service disappears. Some functions can continue locally. A unit may retain its own sensor picture after losing access to the wider operational picture. An edge system may continue processing information already available locally when cloud services are unreachable. A weapon system may operate using preloaded mission data. A headquarters may retain enough information to continue executing an existing plan while losing the ability to receive rapid updates from higher command. Other functions may depend upon reachback for intelligence, specialised processing, authorities, targeting support or access to information held elsewhere. Graceful degradation begins by identifying these differences before combat rather than discovering them during it.

Time matters as much as availability. Information does not immediately become useless when a connection disappears. A terrain database may remain valid for months. A logistics report might remain useful for hours. A track of a moving missile launcher can become operationally irrelevant within minutes. Weather, airspace control measures, friendly positions, threat information and targeting data all decay at different rates. Designing for disconnected operations therefore requires commanders and systems to understand not merely whether information is available, but how old it is and how long it remains good enough for the decision being made. NATO’s Data Strategy already requires data to be discoverable, accessible, trusted, interoperable, curated and secure. In a degraded environment, the temporal dimension of that trust becomes unavoidable.

The same principle applies to PNT. An inertial navigation system may bridge a period of GNSS denial, but its error generally grows with time unless corrected by another reference. A formation can therefore be resilient to a short interruption without being resilient to sustained denial. The operational question is not whether an alternative exists, but how long it provides sufficient accuracy for the mission and what happens after that threshold is crossed. A navigation solution adequate for movement may not be adequate for precision targeting.

Command authority has to degrade gracefully as well. A force that can technically operate without higher headquarters but must request permission for decisions over a communications link that no longer exists is not operationally independent. Mission command becomes relevant here not as a slogan about initiative, but as a practical mechanism for maintaining action when information and communications are incomplete. Commanders need clear intent, understood boundaries and authorities delegated far enough in advance that subordinate formations can continue when higher-level coordination becomes intermittent.

Delegation does not mean pushing every decision downward. Some actions carry strategic, legal or escalation consequences that require higher authority. Others depend upon information unavailable locally. The design problem is to identify those boundaries beforehand. Which fires can continue under existing authorities? What happens when identification confidence deteriorates? Which cyber or electromagnetic activities require coordination because they may affect friendly or civilian systems? At what point does loss of the recognised operational picture require changing the mission rather than continuing it locally? Those are command questions embedded inside the communications architecture.

This is also where redundancy and resilience diverge. Three communications systems do not necessarily provide three independent alternatives. A tactical radio network, commercial mobile service and satellite terminal may appear diverse while sharing an electrical supply, authentication infrastructure, terrestrial backhaul, software-management system or remote support dependency. NATO’s 2026 Resilience Baseline Requirements explicitly call for national risk assessments to consider cascading effects, supply chains, foreign ownership and dependencies among communications, energy, transport and other sectors. Resilience depends on understanding common failure paths, not counting interfaces.

The same scrutiny should be applied to the Digital Backbone. NATO defines it as a federation of networks and systems connecting sensors, decision-makers, actors and effectors across organisational, national, operational and security boundaries. Federation can increase resilience because capability is distributed, but only if the constituent systems can continue useful operations when parts of that federation become unreachable. Otherwise, federation can reproduce dependency at greater scale.

None of this means romanticising analogue warfare. Paper maps, runners and voice radio remain useful contingencies in particular circumstances, but they cannot reproduce the volume, speed or complexity of information required by modern air defence, long-range fires, multinational logistics or multi-domain operations. A headquarters cannot simply print the common operational picture and declare itself resilient. The purpose of fallback procedures is to preserve the minimum operational function required for the mission, not to pretend that losing digital capability has no cost.

That minimum needs to be designed deliberately. A formation should know what information it must possess locally, which services require reachback, which datasets need caching, which decisions can be delegated, which communications paths are genuinely independent and how long each degraded mode remains operationally acceptable. It should also know what capability it is losing at every step. Graceful degradation is not successful because the unit continues doing something. It is successful when commanders understand what remains possible, what has become slower or less certain, and what must stop because the conditions for responsible execution no longer exist.

That makes degradation something NATO must exercise rather than merely engineer against. If a formation has never operated after losing its preferred network, identity service, GNSS source or reachback capability, its fallback architecture remains an assumption. The next task is therefore to identify those assumptions deliberately, break them under controlled conditions and record what happens when they fail.

7. Which assumption breaks first?

Exercises are often described in terms of what they test: a headquarters, a concept, a communications architecture, interoperability between nations or the ability to execute a particular mission. There is another question that deserves equal attention. What assumptions had to remain true for the thing being tested to work at all?

That distinction changes how failure is treated. If a tactical system cannot exchange data with another system during an exercise and technicians restore the connection two hours later, the immediate problem has been solved. But the operational lesson may be much larger. Perhaps the architecture assumed that a particular gateway would always be available, that two nations used compatible data standards, that credentials could be issued quickly enough, that bandwidth would remain sufficient or that a specialist support team would be present. Repairing the connection does not establish whether the underlying assumption was corrected.

NATO already has a formal mechanism intended to prevent observations from ending as anecdotes. The Joint Analysis and Lessons Learned Centre defines an observation as an issue that may require improvement. After analysis identifies the root cause, a recommended remedial action and a responsible tasking authority, it can become a Lesson Identified. NATO considers it a Lesson Learned only after the remedial action has actually been approved and implemented. The NATO Lessons Learned Portal exists to manage and track that process. This distinction is important because NATO itself recognises that identifying a problem and learning from it are not the same thing.

The same discipline should be applied specifically to the assumptions beneath connected warfare. One practical mechanism would be an Assumption Failure Register maintained alongside technical and operational exercise reporting:

Assumption → test condition → observed failure → operational consequence → decision authority → immediate workaround → corrective action → owner → retest

The first entry matters because assumptions need to be stated before the exercise where possible. If a formation assumes continuous access to a particular network, sufficient bandwidth for full-motion video, reliable GNSS timing, reachback to a cloud service or availability of a national gateway, that dependency should be visible. The exercise can then deliberately stress it. Otherwise, the assumption often becomes visible only after something breaks.

The test condition records what actually challenged it. The observed failure records what happened technically. The operational consequence asks the more important question: what could the force no longer do? A lost datalink that prevents one display from updating is different from a lost datalink that stops a targeting chain. A failed identity service affecting administrative access is different from one preventing operators from entering mission-critical systems. The same technical fault can therefore carry radically different military significance depending on where it sits in the architecture.

Decision authority and immediate workaround capture another issue that technical reporting can miss. Who recognised that the assumption had failed, and who had authority to change the way the force operated? If a workaround requires approval from a headquarters that cannot be reached through the degraded network, the contingency exists on paper but not necessarily in practice. If technicians can restore a service only through specialist reachback unavailable in combat, an exercise may have demonstrated repairability without demonstrating wartime resilience.

NATO has substantial machinery for learning from exercises, and recent work shows that it is being used against exactly these kinds of operational problems. JALLC’s analysis of STEADFAST DETERRENCE 2025 and STEADFAST DUEL 2025 reviewed exercise lessons across campaign synchronisation, joint targeting, logistics, interoperability, cyberspace and space to inform the 2026 exercise campaign. Separate JALLC analysis of STEADFAST DETERRENCE examined interoperability between SHAPE and US European Command, including information sharing, situational awareness and the decision-making continuum.

The question is therefore not whether NATO possesses a lessons process. It does. The harder issue is whether exercises are designed aggressively enough to expose assumptions before war does, and whether recurring technical failures are analysed as possible architectural failures rather than isolated incidents.

A connection that fails in one exercise may appear local. The same dependency embedded across several headquarters, nations or weapon systems may represent a systemic vulnerability. Repeatedly repairing it during exercises can actually conceal the scale of the problem if each occurrence is treated as a separate technical ticket. The useful question is whether the failures share a root cause and how much of the force depends on the same assumption.

NATO’s own lessons methodology provides an important standard here. Leadership is expected to decide on remedial action, assign an action body, commit resources and track implementation. JALLC describes true organisational learning as requiring leadership follow-through, not simply collection and dissemination of observations.

That leads directly to the question Keegan Pearl raised in response to Part 1: which assumption breaks first, and who makes the call when it does? An exercise should be able to answer both parts. It should identify the dependency that failed, show the operational consequences, establish who had authority to respond and record whether the workaround preserved the mission.

The final column in the register is therefore as important as the first: retest. A corrective action is not evidence that the problem has disappeared. The changed architecture, procedure, authority or training needs to be exposed again to the condition that caused the original failure, preferably under greater operational pressure. Only then is there evidence that resilience has improved.

That is the difference between documenting failure and learning from it. NATO already has the institutional framework for the latter. The challenge for connected warfare is to make assumption failure a deliberate object of experimentation, then carry what breaks through ownership, correction and retesting. Otherwise the Alliance can become very good at briefing lessons that its next exercise, or an adversary, is able to discover again.

8. Exercise the failure, not the PowerPoint

The Assumption Failure Register is useful only if exercises create conditions in which important assumptions can actually fail. NATO already conducts some of the largest and most sophisticated interoperability testing in the world. The next step is to ensure that proving systems can connect is consistently complemented by testing what the force can still accomplish when those connections deteriorate.

CWIX demonstrates the scale of the first task. At CWIX 2026 in Bydgoszcz, more than 4,000 participants from 46 Allied and Partner nations conducted over 30,000 interoperability tests across 19 capability areas. NATO describes CWIX as its premier digital interoperability event, designed to discover integration problems before operations depend on the systems involved. That is necessary work. A multinational force cannot fight digitally if national systems cannot exchange information, interpret each other’s data or connect to the mission network in the first place.

Steadfast Cobalt addresses another part of the problem. During Steadfast Cobalt 26 Bravo, more than 1,200 participants from 19 Allies and 37 NATO entities operated across 41 locations, testing whether national and NATO communications and information systems could connect through a shared Mission Network. The purpose was explicitly operational: participating forces need to exchange information and coordinate securely while retaining their national systems.

Neither example should be dismissed as merely technical. Interoperability is a prerequisite for multinational warfighting. But interoperability testing and operational stress testing answer different questions. The first asks whether systems can work together under defined conditions. The second should ask what happens to the mission when some of those conditions are deliberately removed.

That means designing failure into the scenario rather than treating it as an inconvenience for technicians to resolve. Remove GNSS from a formation that normally depends upon it and leave it unavailable long enough for alternative navigation errors to become operationally significant. Reduce bandwidth until headquarters can no longer move every data product it normally expects. Cut reachback to a service located outside the theatre. Remove a commercial communications provider without warning. Make an identity-management service unavailable while operators are changing shifts or moving between systems. Introduce contradictory tracks from apparently legitimate sources. Degrade one data feed rather than simply switching it off. Force a national information-release restriction into the middle of a multinational targeting process.

Then leave the failure in place.

The purpose is not to make an exercise artificially difficult. It is to discover what the force actually does when the architecture stops behaving as designed. Does a headquarters know which information should receive scarce bandwidth? Can a subordinate unit continue operating when it cannot reach the preferred cloud service? Can users authenticate through an alternative mechanism? Does the staff recognise that two tracks conflict, or does one automatically overwrite the other? Can a multinational force continue a task when one nation cannot release the data on which the original process depended?

NATO is already moving towards this kind of experimentation. Digital Backbone eXperimentation 2026 is explicitly stress-testing communications under denied, degraded, intermittent and limited conditions. ACT says success will be measured by whether forces can continue to command, coordinate and share priority information when individual networks are weak, disrupted or unavailable. That is a materially different metric from simply restoring connectivity.

The same logic needs to extend beyond communications. If GNSS disappears, measure navigation accuracy and mission completion over time. If bandwidth collapses, measure which information reaches whom and whether decisions still arrive before they become irrelevant. If reachback is removed, measure what can be processed locally. If a data source becomes questionable, measure whether the staff detects the problem and how decision thresholds change. If a national restriction interrupts information sharing, measure whether the multinational process has an authorised alternative rather than quietly assuming the restriction away.

This is where exercise metrics become important. Network availability, packet loss, latency, restoration time and interoperability test results remain valuable engineering measures. They do not by themselves describe operational resilience. A network restored in twenty minutes may still represent mission failure if an engagement window lasted five. A service unavailable for six hours may have little operational consequence if the formation possessed sufficient local information and delegated authority to continue.

The question after a failure should therefore not stop at “How quickly did technicians restore the network?” It should also ask “What could the force still accomplish while it was unavailable?” That means measuring missions completed, decisions made, information lost, authorities exercised, targeting opportunities missed, uncertainty accumulated and functions that had to stop.

ACT now describes its experimentation model in similar terms, arguing that networks, sensors, software, operators and procedures need to be tested as a chain against operational problems rather than as isolated technologies. DiBaX, Layered Counter-UAS Initiative-X and Next Generation Targeting are increasingly being brought together so that data feeds and operational complexity stress the Digital Backbone itself.

That is the direction connected warfare requires. CWIX can establish that systems interoperate. Steadfast Cobalt can establish that multinational mission networks can be formed and operated. Operational experimentation can then attack the assumptions underneath them, while larger exercises establish whether headquarters and formations can still execute their missions when those assumptions fail.

The useful output is therefore not a successful demonstration followed by a slide of technical observations. It is evidence about how capability degrades: what survived, what stopped, what decisions moved elsewhere, what workarounds proved viable and which dependencies turned out to be more important than the architecture had assumed. Combined with the Assumption Failure Register, that evidence gives NATO something considerably more valuable than proof that its systems can connect. It shows what the Alliance can still do when an adversary is trying to make sure they cannot.

9. Designing for the battlefield we actually have

None of this is an argument against connectivity. NATO should pursue a more connected, data-enabled force because the operational advantages are substantial. Better information sharing can improve situational awareness across domains, connect sensors with effectors that previously operated separately, coordinate dispersed formations, reduce duplication and shorten the time between detecting something and deciding what to do about it. NATO’s Digital Transformation Implementation Strategy is built around precisely that ambition: creating a federated Digital Backbone capable of linking sensors, decision-makers, actors and effectors across the Alliance.

The mistake would be treating connectivity as an environmental constant rather than a contested military resource. Russia has already demonstrated the ability and willingness to interfere with GNSS, conduct electronic warfare, attack communications infrastructure and combine cyber activity with conventional operations. Ordinary battlefield friction adds congestion, equipment failure, power loss, damaged infrastructure and human error before deliberate adversary action is considered. The realistic operating environment is therefore neither permanently connected nor completely disconnected. It moves between different degrees of availability, bandwidth, accuracy and trust.

That changes what good architecture looks like. NATO needs to connect when possible, degrade intelligently, operate locally when necessary and reconnect without losing trust, but those requirements have to exist in the design rather than being improvised after failure. Systems need useful local functionality when reachback disappears. Command arrangements need sufficient delegated authority for periods of isolation. Alternative communications and PNT need genuinely different failure paths. Information needs provenance and age indicators so that reconnecting does not simply merge stale or questionable data into the operational picture. Exercises need to establish how long degraded modes remain militarily useful and which missions become impossible as individual dependencies disappear.

This also changes how resilience should be measured. A resilient formation is not one whose network never goes down. Against a capable adversary, that is an unrealistic standard. The better measure is whether loss or compromise of individual components produces understood and bounded operational consequences. Commanders should know what they have lost, what remains available, how much confidence to place in it and what they are still authorised to do. Recovery matters, but so does the ability to fight during the period before recovery occurs.

That brings us to a different problem. Assume NATO succeeds in keeping enough information moving through this contested environment. The radios work sufficiently well. Alternative PNT holds. The Digital Backbone degrades without collapsing. Data continues reaching commanders and effectors.

There is still no guarantee that the system understands what it is seeing.

Sensors produce observations. Radar returns, radio-frequency detections, imagery, acoustic signatures and electronic emissions are measurements of the environment, not conclusions about it. Algorithms can classify those observations, correlate them and assign probabilities. Human analysts and commanders can add context and judgement. None of those steps automatically converts incomplete and adversarially contested information into truth.

That is where Part 3 begins. AI Does Not Know What It Is Looking At will move from connectivity to identification: sensor fusion, classification, training data, labelling, data poisoning, deception, confidence, provenance and human judgement. The question is no longer whether NATO can move information quickly enough. It is whether the information being moved is good enough to justify the decisions made from it.

Materials and sources used for Part 2

NATO digital transformation, Digital Backbone and data

NATO, Alliance Digital Strategy, 13 January 2026 Sets NATO’s current strategic direction for digital transformation, including resilient and interoperable infrastructure, the Digital Backbone, cloud and edge computing, Zero Trust, mission-critical services and operations in degraded, contested and denied environments. https://www.nato.int/en/about-us/official-texts-and-resources/official-texts/2026/01/13/alliance-digital-strategy

NATO, Digital Transformation Implementation Strategy 2.0, 26 May 2026 The implementation framework for NATO’s digital transformation and the development of secure, resilient and interoperable digital capabilities. Particularly relevant to the relationship between the Digital Backbone, multi-domain operations, people, processes and technology. https://www.nato.int/en/about-us/official-texts-and-resources/official-texts/2026/05/26/natos-digital-transformation-implementation-strategy

NATO, Data Strategy for the Alliance, 5 May 2025 Used for NATO’s Alliance Data Sharing Ecosystem, trusted data, interoperability, data governance and the requirement that users and systems be able to assess the integrity of data and the processes surrounding it. https://www.nato.int/en/about-us/official-texts-and-resources/official-texts/2025/05/05/data-strategy-for-the-alliance

NATO, Data Quality Framework for the Alliance, 29 August 2025 Relevant to the discussion of data quality, monitoring, validation, anomaly detection, remediation and the difference between network availability and trustworthy information. https://www.nato.int/en/about-us/official-texts-and-resources/official-texts/2025/08/29/data-quality-framework-for-the-alliance

NCIA, Digital Modernization Useful for the physical and institutional reality beneath NATO’s Digital Backbone. NCIA describes the backbone as being built on existing telecommunications and network infrastructure provided by nations and the NATO Command Structure, while NATO data centres host information and operational networks essential to Alliance activity. https://www.ncia.nato.int/about-us/cyber-defence/digital-modernization

NCIA, IT Modernization Relevant to NATO’s Operational Network, Protected Business Network, Identity and Access Management, cloud adoption and the wider dependencies behind NATO’s digital services. https://www.ncia.nato.int/about-us/service-portfolio/it-modernization

GNSS, PNT and interference

NATO, NATO addresses electromagnetic interference threats to civil aviation safety, 8 May 2025 Documents the increasing frequency and sophistication of GNSS jamming and spoofing across the Baltic, Black Sea and Mediterranean regions and NATO’s work with ICAO and EUROCONTROL. https://www.nato.int/en/news-and-events/articles/news/2025/05/08/nato-addresses-electromagnetic-interference-threats-to-civil-aviation-safety

EASA, EASA updates Safety Information Bulletin on GNSS interference, 3 July 2026 Important current European evidence. The revised bulletin followed analysis of jamming and spoofing incidents and introduced updated operational, training and pilot-ATC measures. https://www.easa.europa.eu/en/newsroom-and-events/news/easa-updates-safety-information-bulletin-gnss-interference

EASA, Protecting the backbone of aviation: EASA and EUROCONTROL host workshop on Europe’s GNSS interference Action Plan, 9 July 2026 Documents GNSS radio-frequency interference as a persistent operational problem across European and neighbouring airspace and explains the European aviation response. https://www.easa.europa.eu/en/newsroom-and-events/news/protecting-backbone-aviation-easa-and-eurocontrol-host-workshop-europes-gnss-interference-action

NCIA, NCIA and Finland strengthen resilience against satellite navigation interference, 2026 Operational trials involving NCIA, the Finnish Defence Forces and Finnish Border Guard. GANDALF is designed to detect, classify and locate GNSS interference and was evaluated in operational airborne and maritime conditions. https://www.ncia.nato.int/newsroom/news/ncia-and-finland-strengthen-resilience-against-satellite-navigation-interference

NCIA and NSPA, testing GANDALF against GNSS signal threats, 24 April 2025 Useful background on NATO’s work to detect and distinguish deliberate GNSS jamming and spoofing rather than treating GNSS loss simply as a receiver problem. https://www.ncia.nato.int/newsroom/news/ncia-and-nspa-join-forces-to-test-new-sensor-to-counter-gnss-signal-threats-and-improve-natos-situational-awareness

Electromagnetic warfare and battlefield adaptation

NATO, Electromagnetic Warfare NATO’s baseline explanation of the electromagnetic environment and electromagnetic warfare, including attacks against radar, communications and navigation systems, electronic masking, reconnaissance and intelligence collection. https://www.nato.int/en/what-we-do/deterrence-and-defence/electromagnetic-warfare

NATO, NATO Electronic Warfare Advisory Committee, Istanbul, 6 November 2025 Useful current NATO assessment of the operational importance of EW following its widespread employment in Ukraine and the Middle East. https://www.nato.int/en/news-and-events/articles/news/2025/11/06/the-nato-electronic-warfare-advisory-committee-newac-convenes-in-istanbul

US Army, Lessons Learned from the Ukrainian Territorial Defense Forces: Command Post Survivability, 6 February 2024 Particularly useful for the commander-level consequences of electromagnetic signatures, communications discipline, antenna placement, Wi-Fi and cellular use, and the relationship between emissions, detection and command-post survivability. https://www.army.mil/article-amp/273510/lessons_learned_from_the_ukrainian_territorial_defense_forces_command_post_survivability

RUSI, Emergent Approaches to Combined Arms Manoeuvre in Ukraine Used for the continuing adaptation contest in Ukraine, including Russian efforts to identify and attack Ukrainian UAV operators, EW positions, radar and command posts, and the growing employment of fibre-optic FPV systems alongside other fires. https://www.rusi.org/explore-our-research/publications/insights-papers/emergent-approaches-combined-arms-manoeuvre-ukraine

CSIS, Ukraine’s Future Vision and Current Capabilities for Waging AI-Enabled Autonomous Warfare Relevant to the relationship between EW pressure, communications loss, drone operations and the development of greater onboard autonomy and alternative approaches to navigation and terminal guidance. https://www.csis.org/analysis/ukraines-future-vision-and-current-capabilities-waging-ai-enabled-autonomous-warfare

SATCOM, commercial infrastructure and physical dependencies

NCIA, Satellite Communications Documents NATO’s SATCOM architecture, including the seven NCIA-operated satellite ground stations, NATO’s satellite centre and the provision of military satellite capacity by France, Italy, the United Kingdom and the United States. https://www.ncia.nato.int/about-us/service-portfolio/satellite-communications

NATO, Commercial Space Strategy, 13 February 2025 Important for the distinction between commercial capacity and assured wartime availability. NATO explicitly addresses continuous commercial access during peace, crisis and conflict and the need to avoid overreliance on any single provider or solution. https://www.nato.int/en/about-us/official-texts-and-resources/official-texts/2025/02/13/nato-commercial-space-strategy

NATO, Baltic Sentry, 14 January 2025 Used for the physical vulnerability of critical undersea infrastructure and NATO’s response to damage involving communications and energy cables in the Baltic Sea. https://www.nato.int/en/news-and-events/articles/news/2025/01/14/nato-launches-baltic-sentry-to-increase-critical-infrastructure-security

SHAPE, Baltic Sentry to enhance NATO’s presence in the Baltic Sea, 14 January 2025 Operational Allied Command Operations account of Baltic Sentry and the protection of critical undersea infrastructure. https://shape.nato.int/news-releases/baltic-sentry-to-enhance-natos-presence-in-the-baltic-sea

Civil infrastructure, energy and resilience

NATO, 2026 Resilience Baseline Requirements, 15 September 2026 One of the most important sources for this part. The requirements address resilient energy, civil communications, transport, supply chains, commercial providers, cascading failures, physical and cyber risks, GNSS and other technological dependencies, and civil support to military operations during crisis and conflict. https://www.nato.int/en/about-us/official-texts-and-resources/official-texts/2026/09/15/2026-resilience-baseline-requirements

NATO, Guidance for Enhancing Public-Private Cooperation for Resilience, 16 July 2026 Relevant to the dependence of collective defence on privately owned and operated infrastructure and the need to incorporate commercial actors into resilience planning before crisis. https://www.nato.int/en/about-us/official-texts-and-resources/official-texts/2026/07/16/guidance-for-enhancing-public-private-cooperation-for-resilience

Cyber compromise and trust

UK Government, Russia behind cyber-attack with Europe-wide impact an hour before Ukraine invasion, 10 May 2022 Official UK attribution of the Viasat KA-SAT cyberattack. The attack began approximately one hour before Russia’s full-scale invasion, primarily affected Ukraine but also disrupted civilian and commercial users elsewhere in Europe, including wind-farm communications. https://www.gov.uk/government/news/russia-behind-cyber-attack-with-europe-wide-impact-an-hour-before-ukraine-invasion

Microsoft, Ongoing Russian cyberattacks targeting Ukraine: Cadet Blizzard, 14 June 2023 Used for the discussion of stolen credentials, persistent access, legitimate administrative tools, web shells and the distinction between cyber compromise and obvious network destruction. https://blogs.microsoft.com/on-the-issues/2023/06/14/russian-cyberattacks-ukraine-cadet-blizzard/

The Microsoft source is vendor threat intelligence rather than an Alliance or government document, so it is best read alongside the official UK attribution and NATO material rather than treated as a standalone authority.

NATO interoperability, experimentation and stress testing

Allied Command Transformation, CWIX 26 Strengthens NATO’s Digital Interoperability and Operational Readiness, 2026 CWIX 2026 involved more than 30,000 interoperability tests across 19 capability focus areas. Particularly useful for distinguishing the essential task of proving interoperability from the different task of testing operational performance after dependencies fail. https://www.act.nato.int/article/cwix-2026-concludes/

NCIA, Mission Network expansion during Steadfast Cobalt 26 Bravo, 5 October 2026 Documents more than 1,200 participants from 37 NATO entities across 19 Allied nations and 41 locations testing national and NATO CIS within a shared Mission Network. https://www.ncia.nato.int/newsroom/news/-ncia-supports-mission-network-expansion-during-steadfast-cobalt-26-bravo-

Allied Command Transformation, Digital Backbone eXperimentation 2026 DiBaX is directly relevant to Part 2 because it is testing the Digital Backbone against operational problems, including resilient communications across multiple networks and integration of data and operational applications. https://www.act.nato.int/activities/dibax/

Allied Command Transformation, The Backbone of NATO’s Digital Transformation: DiBaX26, 7 October 2026 The current DiBaX26 account, published while this article was being written. It describes experimentation around resilient communications, 5G network slicing, situational awareness and the ability of the Digital Backbone to support operational activity under demanding conditions. https://www.act.nato.int/article/backbone-nato-digital-transformation-dibax26/

Allied Command Transformation, From Experimentation to Capability: How ACT Turns Innovation into Military Advantage, 29 September 2026 Useful for the distinction between a technology working independently and its ability to integrate, exchange information quickly enough and work within actual operational processes. It also links DiBaX, Layered Counter-UAS Initiative-X and Next Generation Targeting experimentation. https://www.act.nato.int/article/experimentation-to-capability/

NATO lessons learned and assumption failure

NATO Joint Analysis and Lessons Learned Centre, Lessons Learned The basis for Section 7. NATO distinguishes an Observation from a Lesson Identified and a Lesson Learned. A Lesson Identified requires root-cause analysis, remedial action and a responsible Tasking Authority. It becomes a Lesson Learned only after the remedial action has been approved and implemented. https://www.jallc.nato.int/activities/lessons-learned

NATO JALLC, NATO Lessons Learned Handbook, 4th Edition Detailed description of NATO’s Lessons Learned Capability and process, including observation, analysis, root causes, remedial action, tasking, implementation, monitoring and validation. https://www.jallc.nato.int/application/files/3716/5521/6526/JALLC_LL_Handbook_4th_Edition.pdf

NATO Lessons Learned Portal NATO’s central mechanism for collecting, managing, tracking and sharing Observations, Lessons Identified, Best Practices and Lessons Learned. https://nllp.jallc.nato.int/AboutUs/Pages/WhatisNatoLLPortal.aspx

JALLC, Analysis of STEADFAST DETERRENCE 2025 and STEADFAST DUEL 2025 in support of the 2026 Exercise Campaign, 10 April 2026 Relevant to the argument that exercise findings must feed subsequent adaptation. JALLC analysed lessons covering campaign synchronisation, joint targeting, logistics, interoperability, cyberspace and space for use in NATO’s 2026 exercise campaign. https://www.jallc.nato.int/articles/jallc-publishes-report-exercises-steadfast-deterrence-2025-and-steadfast-duel-2025-support-2026-exercise-campaign

JALLC, Analysis Products Includes the STEADFAST DETERRENCE 2025 interoperability work examining SHAPE-USEUCOM information sharing, situational awareness, organisational alignment and the decision-making continuum, as well as the broader STEADFAST DETERRENCE / STEADFAST DUEL lessons analysis. https://www.jallc.nato.int/organization/products/jallc-analysis-products

Part 2 | Source note

This article draws primarily on official NATO, Allied Command Transformation, NATO Communications and Information Agency, Joint Analysis and Lessons Learned Centre, EASA and Allied-government material, supplemented by operational analysis from RUSI, CSIS and documented Ukrainian battlefield lessons.

The central distinction throughout Part 2 is deliberate: connectivity, availability, integrity and operational usefulness are not the same thing. A network can remain connected while bandwidth collapses, timing becomes unreliable, information is delayed or corrupted, credentials are compromised, or the physical infrastructure beneath the service is attacked. Resilience therefore cannot be demonstrated simply by showing that systems connect.

The operational test is whether the force can continue to understand the situation, exercise command, make authorised decisions and produce military effects while parts of the architecture are degraded or unavailable, and whether it can reconnect afterwards without importing stale, corrupted or untrusted information into the wider system.

That is also why the Assumption Failure Register introduced in Part 2 matters: Assumption → test condition → observed failure → operational consequence → decision authority → immediate workaround → corrective action → owner → retest

It complements NATO’s existing Lessons Learned process rather than replacing it. Its purpose is to make the assumptions beneath connected warfighting explicit enough to attack during experimentation and exercises, and to ensure that a failure repaired during an exercise is not mistaken for an assumption that has actually been corrected.