US faces critical juncture in AI infrastructure as global rivals surge, risking national security
The rapid global expansion of artificial intelligence (AI) data centers is sparking a crucial debate in the United States, with experts warning that lagging in domestic buildout could pose significant national security risks. Amidst a renewed push for American reindustrialization, the ongoing hesitation and local opposition to constructing these vital facilities domestically are drawing parallels to past strategic errors of offshoring critical industries.
This reluctance, proponents argue, risks creating dangerous dependencies on foreign nations for advanced technological capabilities, potentially ceding a critical advantage in the global AI race. Other major powers are aggressively investing, highlighting the urgency for the US to proactively shape the trajectory of AI technology rather than fall behind.
The strategic imperative of domestic AI infrastructure
While the concept of reindustrialization gains traction across America, the pushback against AI data center development reveals a failure to learn from historical offshoring mistakes. These past decisions led to dependencies that are now difficult to reverse. Regardless of where they are built, AI technology is here to stay and will profoundly influence the future. The choice facing the United States is clear: either embrace the construction of AI data centers within its borders to steer AI’s development, or surrender this pivotal advantage to other nations eagerly vying for leadership.
The global landscape underscores this urgency. China, for instance, has committed nearly $300 billion to data center investments through 2030, a figure that excludes private sector spending. India is luring foreign cloud operators with a 20-year tax holiday, while the Republic of Korea offers significant regulatory relief. In stark contrast, many American communities and policymakers are advocating for moratoriums, potentially jeopardizing the nation’s competitive edge and long-term security interests. This stark divergence in strategy highlights a critical moment for US policy.
Data, compute, and intellectual property: Core pillars at risk
The national security implications of where AI data centers are hosted extend far beyond mere economic competition. Failure to establish robust AI data center infrastructure within the United States jeopardizes the very assets these centers contain: the vast amounts of data, the computational power (compute), and the intellectual property (IP) that collectively drive innovation, safeguard privacy, and underpin national sovereignty. Furthermore, by not building these facilities domestically, the US forfeits its ability to establish and enforce security standards for this critical infrastructure, opening doors to vulnerabilities.
Data, often referred to as the “fuel” of AI innovation, requires immense quantities to make AI useful, all of which must pass through these data centers. Consequently, data must be handled with extreme care, particularly due to privacy concerns, and treated as a national asset. Adversaries view data routed through foreign-hosted AI data centers as a prime opportunity to advance their own AI ambitions. This data can be collected for espionage or to accelerate rival AI model development, a tactic China has repeatedly employed in cyberspace and through adversarial distillation. Moreover, sophisticated adversarial attacks during model training could, unbeknownst to us, degrade and manipulate data, thereby compromising the performance and reliability of AI models. Insecure or compromised data impacts everyone, from large corporations to individual citizens, exposing potentially sensitive information across the board.
Beyond data, compute capacity represents another strategic advantage for the United States, which currently hosts 75% of the world’s high-performance compute. Strong export controls on advanced compute hardware are vital for widening America’s AI lead against adversaries and collaborating with allies to shape AI’s trajectory. The commitment to increased domestic semiconductor production, bolstered by the CHIPS and Science Act, reinforces this lead with new US manufacturing capabilities. However, the destination of these chips is crucial; many are destined for AI data centers. If these data centers are sent overseas, the chips will inevitably follow, potentially along with the businesses that supply them. This “re-offshoring” would severely impact US resilience and the economy, which has spent the last four years preparing for the reindustrialization of semiconductor manufacturing.
Today, China’s semiconductor technology lags several generations behind the leading edge, largely due to US export controls on high-end chips and equipment. However, if these critical chips are increasingly housed overseas where the US cannot enforce its security standards, it could inadvertently ease remote access for adversaries, further accelerating China’s AI development. Relying on overseas AI data centers also creates a significant strategic vulnerability during global crises, as other nations, not the United States, would control the allocation of surge compute capacity. There is no guarantee America would receive any of that capacity, and what it might receive would be significantly delayed due to geographic distance. Compute capacity is already severely constrained and will only become more limited as demand for AI tools grows, impacting the military—which already faces potential compute bottlenecks—and innovators powering new breakthroughs, including researchers, students, and startups. Establishing AI data centers in America ensures the ability to prioritize access during a crisis while simultaneously fostering the innovation economy and technical workforce.
Tying data and compute together is AI intellectual property (IP), encompassing the AI models and model weights that have kept America at the forefront of this technology. IP has long been a target for the Chinese Communist Party (CCP), and this extends directly to AI. China has already engaged in IP theft through methods like adversarial distillation to compete with proprietary American frontier models. Sending AI data centers overseas is akin to handing over this valuable IP to the CCP, which will undoubtedly leverage it for its own market dominance and, presumably, military advantage. Fortunately, America currently hosts the most AI data centers, but this lead is precarious. Approximately 60% of announced projects for this year have yet to begin construction, meaning this advantage cannot be taken for granted.
Charting a secure path forward for AI development
As companies and policymakers seize this moment to reframe the narrative about resource use and the benefits to local communities, several immediate actions can help build public confidence and safeguard America’s security interests. These measures aim to accelerate domestic AI infrastructure development while ensuring robust security and strategic control.
- The United States should leverage federal land to initiate rapid AI data center construction. An Executive Order signed in July 2025 by President Trump opened up federal land and streamlined permitting for commercial AI data centers. By utilizing this authority, America can ensure AI is perceived as a common good rather than a local burden. This approach also shifts the responsibility to the US government, rather than local communities, to collaborate with data center developers in scaling AI using modest resources.
- The Commerce Department should utilize the AI Exports Program to promote secure, global AI data center buildout in partnership with allies. Allies and partners require significant AI capacity, and by working together, the US can help establish a network of secure data centers, enhancing collective security and technological alignment.
These proactive steps are essential not only for maintaining America’s technological leadership but also for establishing global security standards for AI. By prioritizing domestic infrastructure and strategic international partnerships, the United States can mitigate critical vulnerabilities and ensure AI development aligns with its national interests and values, securing a competitive edge in an increasingly complex geopolitical landscape.
