Data centers emerge as strategic targets as AI reshapes modern warfare

Data infrastructure has evolved from a tool of geopolitical competition into a direct component of modern warfare as the US-Israeli conflict with Iran escalates, according to Gloria Shkurti Ozdemir, a researcher at the think tank SETA and lecturer at Azerbaijan's Khazar University. Speaking to Anadolu, Ozdemir analyzed how Tehran's retaliatory strikes against Amazon Web Services (AWS) facilities in Bahrain and the United Arab Emirates demonstrate that digital infrastructure now occupies the same strategic position as industrial-age targets like oil refineries and power plants.

Digital infrastructure as strategic target

"Large-scale data centers serve a similar function to air bases, naval bases, and logistics centers, as these structures today act as a kind of 'digital base,' expanding the US' technological presence and integrating American firms into regional state capacity," Ozdemir explained. She noted that these facilities link financial systems, public infrastructure and, in some cases, defense mechanisms to US-controlled digital infrastructure. "These facilities therefore represent strategic infrastructure, and when civilian systems, state databases and sensitive operational processes depend on them, the infrastructure effectively becomes dual-use — historically, such infrastructure has often been considered legitimate targets in warfare."

Evolution of military AI

Ozdemir traced the integration of AI into warfare from earlier applications—the US military's use of Palantir for intelligence, Google's Project Maven for drone image analysis, and Microsoft and Amazon for defense cloud infrastructure—to today's more sophisticated systems. Defense contractors like Lockheed Martin, Raytheon and Northrop Grumman have long integrated machine learning into missile and autonomous vehicle platforms. However, she emphasized that "what makes today's use of AI in warfare different is the introduction of large-scale foundation models developed by firms like OpenAI and xAI—these models not only analyze data but also demonstrate reasoning capabilities across multiple domains, while generating scenario simulations and providing structured decision-making support."

Accelerated warfare and deterrence

Ozdemir explained that operational superiority now depends on executing the OODA loop—observe, orient, decide and act—at unprecedented speeds. "Previously, the US' superiority was based on hardware capacity, but decision-making processes were still conducted at human speeds—advanced AI systems are shifting this balance, and some analysts call this trend 'accelerated warfare.'" This transformation is reshaping deterrence calculations, as "whichever side can simulate likely responses by an adversary gains a strategic advantage." The development is also transforming Silicon Valley's relationship with the Defense Department, as fundamental AI models developed for commercial purposes are increasingly integrated into national security architecture, blurring the boundary between civilian and military technology.

Israel's AI use and algorithmic risks

Ozdemir noted that Israel has historically maintained a military structure closely integrated with high-tech systems and is leveraging this edge through US-based cloud infrastructure to increase operational scale and speed. Leaks have revealed Israel's use of AI systems including Lavender, Where's Daddy and Gospel during the Gaza war to process massive intelligence volumes for target identification. While algorithmic systems provide the advantage of processing data at far greater speeds, Ozdemir warned that "this acceleration in target generation also introduced serious ethical and tactical risks beyond mere technical superiority." Scaling lethal target identification through algorithms shifts decision-making elements from humans to machines. "The military edge Israel derives from AI is not negligible—there is a clear advantage," she said. "However, this acceleration also carries the risk of errors, civilian harm, and accountability challenges."