Will Data Centers Become the #1 Targets in the Next Global War?

The breathtaking display of autonomous systems at the 2025 𝘊𝘩𝘪𝘯𝘢 𝘝𝘪𝘤𝘵𝘰𝘳𝘺 𝘋𝘢𝘺 𝘔𝘪𝘭𝘪𝘵𝘢𝘳𝘺 𝘗𝘢𝘳𝘢𝘥𝘦 in China, followed swiftly by a closed-door 𝘈𝘐 𝘴𝘶𝘮𝘮𝘪𝘵 at the White House led by President Trump, signals more than just national pride or commercial ambition. It signals the start of the East-West competition for Artificial General Intelligence (AGI), a new and significant chapter in world politics. I see a crucial question coming from the blinking server racks of hyperscale data centers, as someone who spent most of his career building and working in them: Will these establishments (data centers) be the main targets in a future conflict?

The answer is not a simple yes or no. It is a strategic dilemma that demands we understand various competing, yet equally valid, realities.

𝗪𝗵𝘆 𝘁𝗵𝗲 𝗗𝗮𝘁𝗮 𝗖𝗲𝗻𝘁𝗲𝗿 𝗦𝗵𝗼𝘂𝗹𝗱 𝗕𝗲 𝘁𝗵𝗲 𝗧𝗼𝗽 𝗧𝗮𝗿𝗴𝗲𝘁?

According to one perspective, the reasoning is unstoppable and stems from traditional military doctrine: locate and destroy the enemy’s center of gravity.

In addition to providing the US with a new weapon, the Manhattan Project established a new class of strategic target: nuclear production facilities. The fear that one’s nuclear capability could be destroyed in a first strike was the foundation of the terrifying “use it or lose it” doctrine of the Cold War. Today, the intelligence that directs the bomb is the equivalent, not the bomb itself.

Data centers are where the AGI is created and raised. These facilities contain the algorithms, massive computing power, and carefully selected datasets that make up a country’s AI capability. Destroying this infrastructure would be a destruction of a country’s future economic and defense capabilities, not just an assault on a military asset. The destruction of these facilities would halt progress, cripple autonomous logistics, blind AI-powered surveillance, and degrade machine-speed decision-making systems.

According to this perspective, a cyber or direct attack on a significant AI training facility is the contemporary equivalent of attacking a missile silo or a nuclear reactor in the 20th century. To firmly establish technological dominance, it is a proactive, decisive move. The country with the highest AGI will not only prevail in a conflict but also change the rules of world power.

Accordingly, data centers are not merely targets but rather the most valuable strategic assets on the board, necessitating a level of defense that matches their worth.

𝗧𝗵𝗲 𝗣𝗿𝗼𝗯𝗹𝗲𝗺 𝘄𝗶𝘁𝗵 𝘁𝗵𝗲 𝗦𝗶𝗻𝗴𝗹𝗲-𝗧𝗮𝗿𝗴𝗲𝘁 𝗔𝗽𝗽𝗿𝗼𝗮𝗰𝗵

But a leader who prepares for the last war will lose the one that follows. Although the above reasoning makes sense, it is dangerously lacking some details. It assumes the application of a 20th century conflict model to 21st century technology.

Survivability and second-strike capability were central to the Soviet Union’s Cold War strategy, which went beyond simply striking first. They distributed their resources, created redundant systems, and hardened silos. AI warfare is more likely to occur in this scenario.

No single, recognizable data center will house the most important AI systems. They will be dispersed throughout resilient, decentralized networks on mobile land-based platforms, underwater cables, airborne servers!!, and within defended satellite constellations. The network itself is the target, not a structure.

Moreover, a missile will not be the first weapon used in any future conflict. It will be a covert, ongoing cyber-infiltration campaign with the much more sinister objective of poisoning data rather than destroying it. A technique called “data poisoning” that involves tainting an adversary’s AI’s training datasets could provide a strategic edge that is more significant than any physical attack.

Instead of destroying their intelligence, you would turn it into a liability by causing it to make disastrous mistakes at a critical juncture.

This development is indicative of a larger paradigm shift: the battlefield is now cognitive rather than merely physical or digital. The objective is to control the narrative, take over the information environment, and influence both human and machine decision-making.

𝗧𝗵𝗲 𝗪𝗮𝘆 𝗔𝗵𝗲𝗮𝗱: 𝗔 𝗥𝗲𝘀𝗶𝗹𝗶𝗲𝗻𝘁 𝗦𝗼𝘃𝗲𝗿𝗲𝗶𝗴𝗻𝘁𝘆 𝗗𝗼𝗰𝘁𝗿𝗶𝗻𝗲

The question of whether data centers are targets cannot be the focus of discussion as we move forward. They are. Developing a new, resilient, distributed, and adaptable notion of digital sovereignty is essential.

Future decision-makers need to support:

𝟭. 𝗗𝗲𝗰𝗲𝗻𝘁𝗿𝗮𝗹𝗶𝘇𝗲𝗱 𝗔𝗿𝗰𝗵𝗶𝘁𝗲𝗰𝘁𝘂𝗿𝗲

Putting money into research and development for AI systems that can work and train on distributed, resilient networks, removing single points of failure.

𝟮. 𝗧𝗿𝘂𝘀𝘁𝗲𝗱 𝗗𝗮𝘁𝗮 𝗖𝗼𝗿𝗿𝗶𝗱𝗼𝗿𝘀:

Creating safe, validated global data-sharing protocols in collaboration with allies to guarantee AI systems are trained on correct, untainted data.

𝟯. 𝗡𝗲𝘄 𝗗𝗲𝘁𝗲𝗿𝗿𝗲𝗻𝗰𝗲 𝗙𝗿𝗮𝗺𝗲𝘄𝗼𝗿𝗸𝘀:

establishing precise guidelines for what, in the digital sphere, qualifies as an act of war, and establishing deterrence by threatening disproportionate and crippling retaliation

Another high-stakes meeting took place in Washington last week as the world watched the imposing 2025 𝘊𝘩𝘪𝘯𝘢 𝘝𝘪𝘤𝘵𝘰𝘳𝘺 𝘋𝘢𝘺 𝘗𝘢𝘳𝘢𝘥𝘦, which featured autonomous drone squadrons, robotized infantry formations, and AI-guided missile systems. President Trump invited top executives from OpenAI, NVIDIA, Google DeepMind, Palantir, and Amazon to discuss America’s path to AI leadership.

These two events, which took place thousands of miles apart but had one thing in common, reflected a rapidly changing geopolitical reality: the start of the East-West AI race.

The East-West AI race is positioned to be the pivotal international struggle of the twenty-first century, much like the nuclear arms race of the Cold War. However, the stakes are higher than just weapon stockpiles this time. They center on who is responsible for machine intelligence, the most potent force humanity has ever produced.

The crucial question is still: will this be a race to dominate or to empower humanity, regardless of whether we refer to it as AGI, ASI, or just “AI supremacy”? How long will it be before cloud regions replace front lines and data centers replace battlefields if the aim is dominance?

Data infrastructure might transition away from civilian use, so we must prepare for that possibility. The next ten years will decide whether the East-West AI race ends in mutual gain or irreversible conflict.

The defining leadership of the next decade will belong to whoever best protects AGI, governs it ethically, and integrates it into a stable global security framework, not just to those who create it first. Those who prepare for every aspect of conflict, from the server room to the cognitive domain, will have a bright future. Building that resilient future is my main goal. As AI becomes deeply embedded in business, government, defense, and society, leaders must think beyond adoption. They must ask whether the digital foundations behind AI are resilient, secure, sovereign, and strategically governed.

If this topic resonates with you, I’d love to hear what you think.

In my book Life in the Digital Bubble, I explore how AI and digital systems will reshape not only technology but also work, families, and society in the decades ahead.

And for organizations navigating these changes today, my CIO advisory focus on helping leaders move beyond scattered initiatives and build clear operating models that turn emerging technologies into real business value.