The Kremlin's Digital Fortresses: Ukraine's War Against Russia's Data Centres

By Matthew Parish

Wednesday 4 November 2026

The modern state is increasingly an electronic construction. Behind its ministries, banks, railway systems, commercial enterprises and military headquarters lies an enormous and largely invisible architecture of computers, servers, databases and communications networks. These are the instruments through which governments exercise authority, businesses conduct transactions and ordinary people organise their lives. A country might possess thousands of tanks, hundreds of combat aircraft and a formidable nuclear arsenal, but if its digital infrastructure stops working, large parts of its economy and administration can be paralysed without a single soldier crossing its borders.

Ukraine has begun exploiting this vulnerability in Russia, with potentially far-reaching consequences. Throughout October 2026, Ukrainian long-range drones struck a succession of major Russian data centres belonging to some of the country's largest technology corporations. The attacks disrupted banking, electronic payments, commercial transactions, cloud computing and even elements of the Russian judicial system. More significantly, they threatened infrastructure essential to Russia's ambitions to become a leading developer of artificial intelligence. The Kremlin has responded with an extraordinary programme of emergency measures, culminating in a presidential decree issued by Vladimir Putin on 2 November 2026 requiring private data centres to make part of their computing capacity available to the state.

These developments represent a significant evolution in the conduct of contemporary warfare. For much of the conflict, Ukraine's long-range strikes against Russia concentrated upon oil refineries, fuel storage installations, military airfields, ammunition depots and factories producing military equipment. Such attacks had obvious military or economic purposes. The destruction of a data centre, by contrast, can have consequences that are less immediately visible but potentially much more diffuse, extending through thousands of businesses and millions of individual transactions.

The underlying strategic question is whether Ukraine has discovered a particularly effective means of imposing economic and technological costs upon Russia, or whether Moscow can adapt its digital architecture sufficiently to absorb the damage. The answer matters beyond the immediate progress of the war. It may tell us something fundamental about the vulnerability of modern societies to sustained military attacks against the technological systems upon which they have become dependent.

The October attacks

The principal Ukrainian offensive against Russian data centres began during the night of 7 to 8 October 2026, when drones struck a facility operated by Yandex in Sasovo, in Russia's Ryazan region. The attack caused a fire and forced the facility to suspend operations. Yandex subsequently acknowledged that the damage was substantial and that it could not immediately determine whether the affected equipment could be restored.

On 9 October, another Yandex data centre, this time in the Kaluga region, was attacked. A third facility, in Vladimir, was struck on 11 October. Within four days, Ukraine had damaged three of Yandex's six publicly identified major data centres, forcing two of them to suspend operations completely. Yandex was compelled to transfer workloads to alternative infrastructure and operate parts of its network under emergency conditions.

The campaign did not end there. On 14 October, Ukrainian drones attacked a major Sberbank data centre in Balakovo, Saratov region. Sberbank is not merely Russia's largest bank but also one of her principal technology companies, with substantial investments in artificial intelligence and advanced computing. Its management confirmed the attack while maintaining that the bank's services continued operating normally.

On 18 October, a further attack struck a facility associated with VK, the technology company responsible for VKontakte, Russia's leading social media platform. VK likewise reported no interruption to its core services. Nevertheless, the cumulative effect was to demonstrate that Russia's principal digital corporations could no longer assume that their most valuable physical infrastructure was beyond the reach of Ukrainian weapons.

The campaign apparently expanded beyond commercial technology companies. On 23 October, Ukrainian monitoring channels reported a drone attack on a data centre associated with Russia's Federal Tax Service in Gorodets, in the Nizhny Novgorod region. The facility reportedly formed part of the distributed infrastructure supporting federal taxation databases and government records. However, independent confirmation of the damage and its consequences remains limited.

The significance of these attacks lies not simply in the number of facilities struck but in their interconnected functions. A modern data centre is not merely an enormous electronic filing cabinet. It houses the computing machinery that enables thousands of organisations to operate simultaneously, often performing functions so integral to daily life that their dependence upon remote servers is scarcely appreciated until those servers become unavailable.

How effective have the attacks been?

There is considerable evidence that the attacks upon Yandex achieved substantial immediate disruption. Yandex operates an extraordinarily diverse ecosystem of Russian electronic services, including internet searches, digital mapping, taxi ordering, food deliveries, payment processing, cloud storage and artificial intelligence applications. Its commercial cloud infrastructure reportedly serves approximately 64,000 customers, ranging from small technology companies to major financial institutions and retailers.

When the company's data centres ceased functioning normally, the consequences cascaded through these dependent businesses. Russian consumers experienced difficulties ordering taxis, accessing digital documents, making payments and collecting goods purchased online. Several banks and major commercial enterprises reported interruptions, while Russian arbitration courts reportedly postponed hearings or temporarily lost access to electronic databases. Even some applications used for medical monitoring were affected, demonstrating how an interruption in apparently ordinary commercial infrastructure can have consequences for vulnerable individuals.

Nevertheless, we should distinguish between temporary inconvenience, prolonged interruption and permanent destruction of computing capacity. They are three quite different things, and the available evidence does not justify treating them as interchangeable.

Yandex managed to restore most of its principal cloud services by transferring workloads to alternative Russian facilities. Modern cloud infrastructure is deliberately designed with redundancy, permitting data to be replicated across multiple locations and applications to continue operating when individual servers fail. Not all customers necessarily enjoy equivalent protection, however. Some are responsible for arranging their own backups, and the survival of stored data does not mean that sufficient computing capacity remains available to process it efficiently.

Sberbank and VK both maintained that their principal services had not been disrupted by the attacks upon their facilities. These assertions cannot establish the absence of physical damage, but they demonstrate that the destruction or impairment of a data centre does not necessarily entail the collapse of the services associated with it. Russia's internet has suffered significant disruption, but reports of its imminent comprehensive breakdown would be premature.

There have also been suggestions that Ukrainian strikes caused a dramatic reduction in Russian social media bots and organised online propaganda. Such claims have considerable political appeal, because they imply that the destruction of physical computers can simultaneously disrupt the Kremlin's international information operations. However, investigations into widely circulated evidence for these claims have found serious problems, including fabricated or misinterpreted graphs. No reliable public assessment yet establishes that the attacks substantially reduced Russia's capacity to conduct online influence campaigns.

The strongest conclusion is therefore a more measured one. Ukraine has demonstrated that it can interrupt important elements of Russia's digital economy, inflict substantial repair costs and force major technology companies to undertake emergency reorganisation. Whether these interruptions become a sustained constraint upon Russian economic activity depends upon the extent of the physical damage, the availability of replacement equipment and Russia's success in reorganising its infrastructure.

The artificial intelligence dimension

The consequences for Russian artificial intelligence research may ultimately prove more consequential than the immediate disruption of ordinary electronic services. Yandex has invested heavily in developing large language models and other advanced artificial intelligence systems, aspiring to compete with the technological achievements of American and Chinese corporations.

Such ambitions require vast quantities of computing power. Training sophisticated artificial intelligence models involves coordinating thousands of specialised processors over extended periods, performing calculations on an almost unimaginable scale. The equipment is expensive, consumes substantial electricity and requires carefully managed environmental conditions. Replacing it is considerably more difficult than purchasing ordinary computers.

Yandex's facilities at Sasovo and Vladimir have housed three of the company's most significant supercomputers. Two, known as Lyapunov and Chervonenkis, were located at Sasovo, while the Galushkin system was associated with Vladimir. The computers have supported the development of Yandex's artificial intelligence technologies and represent substantial investments in specialised computing equipment.

The precise condition of these machines following the Ukrainian attacks has not been publicly established. Nevertheless, satellite imagery indicated serious damage to the facilities, and analysts have warned that high-performance processors, including Nvidia A100 chips, may have been damaged or destroyed.

This is where Western economic sanctions become especially relevant. Russia already faces substantial obstacles in acquiring the most advanced Western-designed semiconductors. Although various international trading arrangements may allow sanctioned technology to reach Russian purchasers, obtaining sophisticated processors in sufficient quantities, integrating them into working supercomputers and maintaining the necessary technical support are all considerable undertakings.

The destruction of equipment that is already installed and functioning can consequently impose costs far beyond its nominal replacement value. A computer that cannot be repaired or replaced may delay an entire research programme, while scientists and engineers accustomed to having powerful computational resources available may find themselves competing for substantially diminished capacity.

The distinction between artificial intelligence research and ordinary electronic services is fundamental. A disrupted taxi application may be restored within a few days by transferring its workload elsewhere. Replacing a damaged supercomputer may require months of expensive reconstruction, assuming the necessary equipment can be obtained at all.

The military implications are also potentially significant. Artificial intelligence increasingly contributes to intelligence analysis, autonomous systems, satellite image interpretation and other applications relevant to modern warfare. However, it would be incorrect to assume that every Russian commercial AI system has a military function, or that the destruction of Yandex's computing infrastructure has necessarily impaired Russian battlefield operations. The available public evidence does not permit such a definitive conclusion.

What can be said is that Ukraine's attacks have exposed a vulnerability in Russia's technological ambitions. There is little point in aspiring to become a world leader in artificial intelligence if the physical infrastructure necessary to train and operate sophisticated models cannot be reliably protected.

The paradox of Russia's sovereign internet

Perhaps the most interesting aspect of these events is the manner in which Russia's own political philosophy has contributed to its vulnerability.

For more than a decade, the Kremlin has pursued what it describes as digital sovereignty. The underlying objective has been to reduce Russia's dependence upon foreign technological infrastructure, particularly that operated by American companies, while enhancing the state's capacity to supervise and control electronic communications.

Legislation introduced in 2015 required the personal data of Russian citizens to be stored within Russia. Subsequent measures promoted a sovereign Russian internet, commonly known as the Runet, and expanded the powers of the state communications regulator, Roskomnadzor. Foreign cloud services became increasingly difficult for Russian businesses to use, particularly after the full-scale invasion of Ukraine in February 2022 and the imposition of extensive Western sanctions.

The consequence was an increasing concentration of Russian electronic infrastructure within the country's own territory.

This approach made political sense from the Kremlin's perspective. A state seeking extensive control over information naturally prefers the relevant computers, communications facilities and databases to remain within its jurisdiction. It also reduces certain vulnerabilities to foreign sanctions, commercial withdrawal and external political pressure.

However, it creates another category of vulnerability. Computers situated within Russia are exposed to the same military threats as other Russian infrastructure, whereas computers distributed across numerous foreign jurisdictions may be substantially more difficult for Ukraine to attack physically.

Ukraine adopted a markedly different approach after the beginning of the full-scale invasion. With the assistance of international technology companies and foreign governments, it transferred substantial elements of its critical electronic infrastructure to cloud facilities outside its borders. This geographical dispersal has contributed to the resilience of Ukrainian government services despite repeated Russian attacks upon domestic digital infrastructure.

Russia's attacks upon Ukrainian data centres during September and October 2026 form an important part of this story. Russian forces damaged several telecommunications and computing facilities, disrupting internet services and damaging the infrastructure of Ukrainian commercial providers. Ukraine's government subsequently committed additional funding to geographically separated backup systems and alternative computing facilities.

The Ukrainian attacks upon Russian data centres therefore followed a period of intensified Russian attacks upon equivalent Ukrainian infrastructure. President Volodymyr Zelensky indicated that Ukraine was responding reciprocally to Moscow's actions. Yet the contrasting consequences illuminate the significance of the two countries' different technological strategies.

The essential irony is that a policy designed to protect Russia's technological independence has increased its exposure to physical attack. The Kremlin sought to reduce dependence upon the outside world by keeping Russian information inside Russia, only to discover that this arrangement created a concentrated collection of valuable and potentially vulnerable assets.

The Kremlin's response: underground data centres

Moscow has begun responding to these developments with a programme intended to harden its remaining computing facilities against physical attack.

In late August 2026, Putin signed a decree allowing the authorities to assume temporary control of certain critical infrastructure where operators failed to provide adequate protection. The measures covered several important economic sectors and were subsequently understood to apply to data centres supporting government agencies, banking systems and other essential services. Operators were already considering stronger physical defences, additional communications redundancy and improved cybersecurity arrangements.

Following the October attacks, Russia's Ministry of Digital Development approved further recommendations addressing the physical protection of data centres. These proposals divide facilities into three categories according to their importance, distinguishing strategically essential infrastructure from facilities supporting important regional or commercial services and those whose disruption would primarily affect their own operators.

The most striking recommendation is that strategically important new data centres should be constructed underground or within reinforced bunker-type facilities. Other proposed measures include reducing their visibility from the air, strengthening protective structures and combining physical barriers with electronic systems designed to counter aerial threats.

These recommendations reflect a transformation in the way Russian authorities understand digital infrastructure. Until comparatively recently, data centre design was principally concerned with uninterrupted electrical supply, temperature regulation, communications reliability, physical security against intruders and protection against accidental fire. The possibility of military drones attacking commercial computing facilities was not generally treated as a central consideration.

Russia is now attempting to incorporate the principles of military fortification into civilian technological architecture. This represents a substantial change in engineering priorities, with corresponding increases in construction and operating costs.

There are already indications of how expensive such adaptation may become. On 2 November, reports emerged that the Russian telecommunications company Rostelecom was negotiating with the mining corporation Nornickel about constructing underground data centres. Industry estimates suggested that these facilities could cost approximately 30 to 40 per cent more than comparable above-ground installations.

Underground construction is not a universal solution. Data centres require extensive electricity supplies, cooling systems, communications connections, fire protection arrangements and physical accessibility for maintenance. Reinforced or underground facilities may offer greater resistance to some forms of attack, but they do not eliminate operational risks or the need for redundant infrastructure.

Moreover, the cost of protecting existing facilities may be especially burdensome. New buildings can be designed around demanding security requirements, whereas modifying large established data centres may involve significant engineering work and operational disruption. Russian technology companies already face the financial pressures associated with sanctions, expensive borrowing and restricted access to specialised imported equipment.

The Kremlin is therefore confronting a familiar wartime economic problem. Resources devoted to protecting existing infrastructure cannot simultaneously be devoted to expanding productive capacity or investing in technological innovation.

The decree of 2 November: emergency mobilisation of computing power

The most significant recent measure is Presidential Decree No. 802, signed by Vladimir Putin on 2 November 2026. It establishes an emergency framework for maintaining the operation of Russian information systems when ordinary computing infrastructure is disrupted.

Under the decree, the owners of non-state data centres must reserve 10 per cent of their available floor space and computing capacity within ten days of its entry into force. These resources may continue to be used commercially until the government requires them. When directed by a specially constituted government sub-commission, operators must make the designated resources available within 24 hours.

The authorities will also determine the maximum remuneration payable for the requisitioned capacity. Russia's Digital Development Ministry must inventory relevant facilities and identify the available computing resources, while the government is required to develop further arrangements for ensuring the reliability of information systems.

In practical terms, the Kremlin is attempting to establish a national emergency reserve of computing power. Rather than leaving individual companies entirely responsible for finding alternative servers after an attack, the state wishes to possess authority to reallocate resources from one operator to another.

There is an obvious logic to this arrangement. During a major disruption, the ability to identify available facilities and coordinate their use could contribute to the continuity of important government and commercial systems. The existence of a national inventory might also help the authorities understand where Russian computing capacity is concentrated and where shortages are emerging.

However, a decree cannot create additional servers, specialised processors or reliable electrical capacity merely by instructing their owners to make resources available. Nor can all computing workloads be transferred instantaneously between different hardware and software environments. The practical value of the new arrangements will depend upon whether suitable spare capacity exists, whether systems can be migrated safely and whether the operators involved possess the expertise and communications infrastructure necessary to perform such transfers.

There is also a question of economic incentives. Private companies investing substantial sums in data centres may become reluctant to expand their facilities if they believe the government can requisition capacity on comparatively unfavourable financial terms. The immediate resilience obtained through emergency state powers could therefore come at the expense of private investment over the longer term.

As of 3 November, the decree is newly issued. Its implementation and effectiveness remain to be demonstrated.

A remarkable concession to foreign computing infrastructure

Perhaps the most revealing provision of the November decree is its explicit authorisation, subject to approval by the Federal Security Service, of the temporary use of computing resources located outside Russia.

This is a significant departure from the uncompromising domestic orientation of the Kremlin's digital sovereignty programme. Although the permission is conditional and limited, it acknowledges that Russia may need access to foreign computing facilities to ensure the continuity of her own information systems.

The provision does not amount to an abandonment of Russian data localisation legislation or a general authorisation to transfer sensitive information abroad. Nor does it demonstrate that large numbers of Russian government systems have already been relocated internationally. It nevertheless creates a legal mechanism through which Moscow can contemplate arrangements that its previous political rhetoric tended to discourage.

The implications are substantial. A country seeking technological independence may discover that resilience requires interdependence, because information systems distributed across multiple jurisdictions are not exposed to the same physical risks as those concentrated within a single state.

Russia may be able to obtain assistance from foreign computing providers willing and legally able to serve its organisations. However, there will remain difficulties concerning sanctions, the security of sensitive information, network performance, commercial reliability and the willingness of foreign governments to accommodate Russian state activities.

The FSB's mandatory involvement is particularly interesting. It suggests that Moscow regards the possibility of transferring computing resources abroad as a matter of national security rather than an ordinary commercial decision. The Kremlin wants the advantages of geographical dispersal without relinquishing centralised political oversight.

There is a contradiction here that may prove difficult to resolve. Effective technological resilience often requires commercial flexibility, decentralisation and rapid independent decision-making. The Russian government's response, by contrast, places additional responsibilities in the hands of central authorities whose approval may be required before commercial operators can adopt solutions that would otherwise be technically straightforward.

The legal and humanitarian questions

The expansion of the conflict into civilian computing infrastructure also presents difficult questions of international humanitarian law.

Data centres are not automatically legitimate military targets merely because they support the economy of a country engaged in war. Under the law of armed conflict, civilian facilities may become military objectives where their nature, location, purpose or use makes an effective contribution to military action and their destruction or neutralisation offers a definite military advantage.

A facility supporting military command systems, intelligence processing or weapons-related artificial intelligence might therefore qualify as a military objective in circumstances where an ordinary commercial data centre would not. Even where a facility constitutes a military objective, the principles of proportionality and precautions in attack remain applicable.

In an analysis published on 29 October, international law scholars Marko Milanovic and Michael Schmitt examined the justifications publicly advanced by Russia and Ukraine for their respective data centre attacks. They observed that Russia had generally claimed military uses for the Ukrainian facilities it attacked, although the factual accuracy of those claims remained unestablished. By contrast, Ukraine's publicly articulated explanations had often emphasised retaliation and the disruption of Russian information operations, arguments that do not in themselves establish a lawful military objective.

The authors concluded that Ukraine might possess stronger legal justifications than those she had publicly disclosed, but that her public statements did not adequately establish them. Their analysis illustrates why reciprocal attacks against civilian infrastructure cannot be regarded as lawful merely because the opposing party has previously undertaken similar operations.

These distinctions deserve particular attention because digital infrastructure performs functions essential to civilian welfare. Medical systems, electronic payment networks, public administration, transport services and emergency communications may all depend upon facilities that simultaneously serve commercial or military customers.

A successful attack upon a computing facility may cause relatively little immediate physical injury while imposing substantial indirect consequences upon people who depend upon the services it provides. The resulting humanitarian risks must be taken seriously irrespective of which belligerent conducts the attack.

There is an additional danger that both countries will increasingly integrate their military and civilian computing requirements within shared infrastructure. Such arrangements may be commercially efficient, but they complicate the legal and practical task of protecting civilians from the consequences of military operations. The distinction between civilian and military technology cannot simply be abandoned because modern computing systems are interconnected.

Can Russia adapt?

It would be a mistake to underestimate Russia's capacity to respond to these attacks. The country possesses a substantial technology sector, experienced engineers and an extensive network of existing data centres. It also retains significant industrial resources and the ability to impose emergency regulatory requirements upon private businesses.

The evidence that Yandex restored much of its ordinary commercial infrastructure after the October attacks is itself important. Russia's digital economy did not collapse, and existing redundancy arrangements have already demonstrated their value.

Nevertheless, adaptation will come at a price. Additional physical protection, geographically dispersed computing capacity, emergency reserves, more extensive backup arrangements and the possible relocation of valuable equipment all involve considerable expense. The more resources Russia commits to protecting its digital infrastructure, the fewer resources remain available for its further development.

This is particularly significant in the context of artificial intelligence, where international competition depends upon sustained investment in increasingly sophisticated computing equipment. Russia's difficulties in accessing the most advanced semiconductors make the loss of existing high-performance systems especially problematic.

There is also the question of time. Physical defences can sometimes be strengthened relatively quickly, but new underground data centres, replacement supercomputers and geographically distributed information systems require substantial periods of planning, construction and testing. The Kremlin cannot accomplish these projects merely through administrative instruction.

The effectiveness of Ukraine's attacks should therefore be understood in terms of their cumulative consequences rather than their immediate publicity. The destruction of a computing facility may temporarily disrupt ordinary services, but it can also impose prolonged replacement costs, divert investment and force an adversary to reconsider the architecture of its entire technological economy.

The future of warfare against digital states

There is a wider lesson in the Ukrainian campaign against Russian data centres. Contemporary societies have developed extraordinary technological capabilities while becoming increasingly dependent upon a comparatively small number of physical facilities that support them.

For decades, the development of the internet was accompanied by the assumption that distributed communications networks would prove highly resilient against physical attack. To a considerable extent this remains true. The internet itself is not a single machine that can be destroyed, and the failure of individual data centres does not imply the disappearance of an entire national network.

Yet the commercial development of cloud computing has created concentrations of computational power and essential services within vast installations. These facilities generate substantial economies of scale, but they also create dependencies extending far beyond their physical boundaries.

Ukraine's attacks have demonstrated that the economic and administrative consequences of damaging such facilities may be disproportionate to their physical size. Russia's response, involving proposed underground construction, stronger physical defences, emergency requisitioning powers and a conditional relaxation of restrictions upon foreign computing resources, suggests that the Kremlin recognises the seriousness of the problem.

The paradox is striking. Russia has spent years attempting to build a technologically sovereign state whose information infrastructure remains within the reach of its political authorities. Ukraine's drones have revealed that this same infrastructure is not necessarily beyond the reach of military attack.

There is no persuasive evidence that Ukraine has brought Russia's digital economy close to comprehensive collapse. Nor can we conclude that her strikes have definitively crippled Russia's artificial intelligence programmes, because the condition of several critical computing installations remains unknown. But the documented disruption, the potentially prolonged consequences for specialised equipment and the extraordinary character of Moscow's recent regulatory response establish that the campaign has had effects extending well beyond the physical damage inflicted upon individual buildings.

The most important achievement of the Ukrainian attacks may ultimately be the strategic dilemma they impose upon Russia. Moscow must decide how much of its national wealth to commit to protecting a technological infrastructure that was designed primarily for commercial efficiency and political control rather than wartime resilience. It must also reconsider whether a digital system organised around centralisation and domestic sovereignty can provide the flexibility required to survive prolonged military pressure.

The war in Ukraine has repeatedly demonstrated that modern conflicts are not won or lost exclusively upon the battlefield. They are also contests between industrial systems, economic institutions, technological capabilities and competing forms of national organisation.

Russia has discovered that its data centres are now part of this contest. The computers upon which her modern state depends have become another vulnerability that she must devote resources to defending, and the Kremlin's willingness to contemplate underground bunkers for its most sophisticated electronic machinery illustrates how profoundly this war is changing the architecture of contemporary civilisation.

The fortress of the future may not contain soldiers or artillery. It may contain computers, and the struggle to keep them functioning may become an increasingly important feature of twenty-first-century warfare.