Por Renan Filho · 2026-10-14**Lead**
A recent wave of Ukrainian cyberattacks targeting Russian energy grids and logistics networks has inadvertently provided a stark warning to the United States: the same tactics could be used to paralyze the vast data-center ecosystems that underpin America’s artificial intelligence industry. According to multiple cybersecurity analysts and declassified briefings shared with allied governments, the attacks demonstrated how a determined adversary could sever the power, cooling, and connectivity that keep AI supercomputers running — potentially costing the U.S. economy tens of billions of dollars in lost productivity and undermining its global technological leadership. The findings have been presented to President-elect Donald Trump’s transition team as part of a broader economic security assessment, sources familiar with the matter told FORTIS NEWS.
**Context**
The Ukrainian campaign, conducted primarily by the country’s military intelligence unit GUR and volunteer hacker collectives, has focused on disrupting critical infrastructure inside Russia since the full-scale invasion began in 2022. One notable operation involved a coordinated assault on the power substations feeding a cluster of data centers in the Moscow region, causing intermittent outages that forced several cloud-service providers to throttle operations for hours. Another attack targeted the cooling systems of a major server farm in Siberia, exploiting a vulnerability in industrial control software to trigger overheating and automatic shutdowns. While the immediate impact on Russian AI development was limited — the country’s AI sector is relatively small — the methods themselves are highly transferable.
U.S. intelligence agencies have long warned that America’s AI infrastructure is dangerously concentrated. The majority of the nation’s large-scale AI training clusters are located in a handful of regions — Northern Virginia, the Pacific Northwest, and parts of the Southwest — where they draw enormous amounts of electricity and rely on just-in-time supply chains for specialized hardware. A 2024 report from the Department of Homeland Security’s Cybersecurity and Infrastructure Security Agency (CISA) identified data centers as “critical infrastructure of the highest order,” noting that a simultaneous attack on just three major facilities could disrupt financial markets, healthcare systems, and defense networks. The Ukrainian operations, though aimed at Russia, have provided a real-world stress test of those vulnerabilities.
**Análise**
The economic stakes are immense. The U.S. AI industry is projected to contribute more than $1 trillion to GDP by 2030, according to industry estimates, with data-center construction alone accounting for billions in annual investment. A coordinated cyber-physical attack that knocked out a significant portion of that capacity — even for a few days — could trigger cascading failures across sectors that depend on AI-driven analytics, from logistics and manufacturing to finance and national security. The cost of a single major outage at a hyperscale data center has been estimated in the hundreds of millions of dollars, and a multi-site event could easily exceed the $10 billion mark when factoring in lost revenue, recovery expenses, and reputational damage.
Moreover, the demonstration of such capabilities by Ukraine — a country with limited resources compared to major state actors — underscores the low barrier to entry for this kind of warfare. Adversaries like China, Russia, or Iran could replicate or surpass these tactics with far more sophisticated tools. The incoming Trump administration faces a choice: either accelerate efforts to harden AI infrastructure through redundancy, distributed siting, and hardened supply chains, or risk a catastrophic economic blow that could set back U.S. technological dominance by years. The Ukrainian example has made that choice urgent, not theoretical.
**Fontes**
- Cybersecurity and Infrastructure Security Agency (CISA), “Critical Infrastructure Resilience: Data Centers and AI,” 2024. - Reports from the Ukrainian Main Intelligence Directorate (GUR) on cyber operations, as cited in open-source intelligence briefings by the Atlantic Council’s Cyber Statecraft Initiative. - Analysis by the Center for Strategic and International Studies (CSIS) on the economic impact of data-center outages, 2023. - Briefings to the Trump transition team, as described by two U.S. officials who spoke on condition of anonymity because they were not authorized to discuss the matter publicly.