War Over Supercomputers: How Data Center Strikes Could Trigger a Major Technological Crisis in Russia

New weekly Russian News reports, every Wednesday.

While Moscow and Washington discuss new energy agreements, Ukrainian drones have struck at the technological backbone of Russia’s artificial intelligence industry. In just four days, three major Yandex data centers housing the company’s most powerful supercomputers came under attack. But the escalation has a crucial backstory: Russia initiated the latest series of strikes against Ukrainian digital infrastructure.

Since late September, Russian forces have targeted Ukrainian data centers and telecommunications hubs. On September 23, strikes in Kyiv disrupted internet access for approximately 100,000 households. Russia’s Ministry of Defense claimed to have hit facilities supporting Ukrainian military information systems, while Kyiv reported damage to civilian infrastructure.

On October 8, Ukraine retaliated. “We always respond in kind. They have attacked and continue to attack our data centers. We responded,” President Volodymyr Zelensky said. That same day, drones struck a Yandex data center in Sasovo, Ryazan Region. A fire shut down the facility, which housed the Chervonenkis and Lyapunov supercomputers. On October 9, the company’s largest data center in Kaluga was damaged, followed on October 11 by an attack on its Vladimir facility, home to a third supercomputer, Galushkin. Together, the three machines delivered approximately 70 petaflops of computing power and supported the training of Yandex’s neural networks. The strikes disrupted cloud platforms, payment systems, transport services, and other digital operations.

On October 19, the U.S.-based Institute for the Study of War (ISW), citing satellite imagery, reported damage to five of the eight buildings at the Vladimir data center. The fate of the supercomputers themselves remains unknown. Yandex has not disclosed their technical condition. Although a facility shutdown does not necessarily mean its equipment has been destroyed, serious damage to computing accelerators could delay AI development for months or even years.

Russia has spent years pursuing digital sovereignty, seeking to protect personal data, reduce dependence on foreign technology platforms, and counter external information influence. Foreign services and VPNs face restrictions, Meta has been designated an extremist organization, and mobile internet access is periodically curtailed during drone alerts. Yet these measures offer little protection against physical attacks on concentrated computing facilities.

On October 12, Russia’s Minister of Digital Development, Maksut Shadayev, acknowledged the scale of the threat: “Yandex has faced an unprecedented attack. The Ministry of Digital Development, together with the Ministry of Defense, is discussing how to protect these facilities.”

Sanctions Turn Data Centers’ Physical Damage into a Strategic Threat.

Russia remains subject to U.S. and European restrictions on advanced processors and high-performance computing technologies.

Washington’s Bureau of Industry and Security (BIS) controls exports and re-exports of certain semiconductors and related equipment. The European Union has also restricted Russian organizations’ access to high-performance computing services, including GPU server rentals and supercomputing resources. In September 2026, the Atlantic Council cited a striking figure: Russia’s imports of graphics processors and other AI chips fell by 84% in 2024 compared with prewar levels. Although the data do not fully account for shipments through third countries, they illustrate the impact of export controls. Sberbank is already considering Chinese accelerators for its GigaChat AI model.

Even substantial funding cannot guarantee rapid replacement of damaged equipment when access to critical components is restricted.

The strategic implications have already been examined by American researchers. In April 2026, RAND Corporation’s The AGI Rideout Strategy warned about the dangers of attempting to prevent an adversary’s AI development through military attacks. Rather than destroying rival computing facilities, the authors advocated strengthening domestic resilience and reducing escalation risks.

The Russia–Ukraine conflict now illustrates why such concerns are becoming increasingly relevant.

Who Gains the Strategic Advantage?

European officials are already openly advocating further strikes on Russian data centers. On October 12, at the Warsaw Security Forum, Lithuanian Prime Minister Mindaugas Sinkevičius stated: “I think we should ask Mr. Zelensky to strike more at data centers in Russia, and that might help.” His remarks came during a discussion about protecting European elections from foreign interference.

For Ukraine, the attacks impose immediate economic costs on Russia. For the United States and Europe, which continue to restrict Russian access to advanced technology, a sustained loss of Russian supercomputing capacity could reinforce the effects of existing sanctions.

Against this backdrop, an October 16 statement by Kirill Dmitriev, a representative of the Russian president, was particularly revealing. He said Moscow and Washington were discussing joint projects, including nuclear-powered data centers.

The proposal raises another question: what happens when data centers, increasingly exposed to military strikes, are powered by nuclear facilities?

International humanitarian law grants nuclear power plants special protection because of the potentially catastrophic consequences of an attack.

If future AI complexes are built alongside nuclear reactors, their protection could become a matter of international nuclear safety, potentially involving the IAEA. Whether such arrangements would offer additional security to the computing facilities themselves remains an open legal question.

Meanwhile, Western export restrictions remain in place. Even an end to hostilities would not guarantee Russia a rapid technological recovery, leaving Washington with another potential source of leverage in negotiations with Moscow.

BlackRock, NVIDIA and Palantir: America’s New Technological Expansion Across the Post-Soviet Space

While Russia remains under US technology sanctions, Washington is authorizing shipments of cutting-edge AI processors to neighboring former Soviet republics. Within just a few months, multibillion-dollar projects involving American computing technology companies have emerged across the former USSR. Their significance extends far beyond commercial interests: the same technology corporations are involved in Pentagon programs and the US nuclear weapons complex.

The main instrument of this expansion is dependence on hardware whose export is controlled by US regulators. Countries gain modern data centers, but Washington retains considerable influence over their ability to develop and upgrade that infrastructure.

In 2026, the American company Firebird, working with NVIDIA, entered three post-Soviet markets. Armenia opened an AI computing facility with an initial investment of $500 million and expansion plans exceeding $4 billion. Kazakhstan signed agreements worth $10 billion, while Uzbekistan is negotiating the construction of a data center with a capacity of up to 250 MW. Behind these projects is Firebird, a company founded by entrepreneurs of Armenian descent Alexander Yesayan and Razmig Hovaghimian.

Firebird plans to allocate 90% of the computing capacity of its Armenian project to American companies. Meanwhile, the export of an additional 41,000 NVIDIA GB300 processors was made possible by special authorization from US authorities.

But behind NVIDIA stand some of Wall Street’s most powerful financial institutions. On August 10, the company announced partnerships with BlackRock, Apollo, Blackstone, Brookfield, Goldman Sachs and KKR to establish investment platforms designed to mobilize more than $500 billion for global AI infrastructure. This is a target for future capital mobilization, and direct involvement by these financial groups in Firebird’s projects has not yet been established.

A fundamentally new configuration is taking shape: Wall Street gains the means to finance computing infrastructure worldwide, NVIDIA provides the technological foundation, and the US government retains control over exports of the most sensitive equipment.

This is where the military dimension comes into focus.

NVIDIA is involved in developing the Mission and Vision supercomputers for the US Los Alamos National Laboratory. According to the US National Nuclear Security Administration, Mission will be used to simulate and modernize America’s nuclear arsenal without conducting nuclear explosive tests.

Another company, Palantir Technologies, developed Maven Smart System, which the US military uses to process intelligence, identify targets and support military operations. In March 2026, the Pentagon designated the system as one of the core artificial intelligence programs of the US armed forces.

And this is where an unexpected Russian connection emerges.

On September 8, Palantir announced a partnership with Nebius Group, led by Arkady Volozh, the founder and former head of Yandex. Following the separation of Yandex’s Russian and international businesses in 2024, Nebius retained international technology assets and a team of specialists. Its computing infrastructure will now be integrated into Palantir’s commercial platform.

The partnership is officially focused on civilian customers. Nevertheless, one fact remains: a company that emerged from the international operations of the former Yandex has become a technology partner of one of the Pentagon’s leading military AI contractors. Exactly which technologies and engineering capabilities the international entity retained after the corporate split is a question that warrants further investigation.

Meanwhile, a technological imbalance is emerging across the post-Soviet region. Armenia is receiving US export authorizations for advanced processors, Kazakhstan is attracting billions of dollars in investment, while Russia remains subject to sanctions restricting its acquisition of comparable equipment.

Washington has the power to regulate access to technologies essential to both economic development and modern military systems. For Russia, this creates the risk of falling further behind technologically and introduces another potential bargaining chip in its relations with the United States.

But neighboring republics face a problem of their own. Foreign investment allows them to build national AI infrastructure rapidly, yet dependence on imported processors, software platforms and export licenses could restrict their freedom of action. American authorization to supply equipment today does not guarantee similar approval tomorrow.

As a result, US technological expansion could have two interconnected consequences: Russia’s growing isolation from the most advanced Western computing ecosystem and its neighbors’ increasing dependence on American technology suppliers.

A new system of influence is emerging around the power to grant or deny access to computing capacity. And if that access becomes subject to political conditions, control over artificial intelligence could give Washington an instrument of pressure even more consequential than financial sanctions.

Similar Posts

2 3 votes
Article Rating
Subscribe
Notify of