The construction of American data centers in the Gulf is sparking sharp debate in Washington. Is it a giveaway of strategic assets, or a necessary step to ensure American supremacy in the global AI race and prevent China from filling the void?
Introduction: Debate in the heart of Washington
“Why are we putting data centers and research hubs in Dubai? These high-paying new tech jobs should be in the United States.” That was the question posed by U.S. Congressman Ro Khanna, echoing the concern of a broad faction in Washington known as the “AI non-proliferation camp.” Answering that question is David Sacks, President Trump’s AI advisor, who believes these partnerships are essential to maintaining America’s edge.
This debate centers around the U.S.-UAE AI acceleration partnership and similar agreements with the Kingdom of Saudi Arabia. Do these deals represent a giveaway of the 21st century’s “enriched uranium”—that is, computing power—or are they a smart strategic move to address the energy and capital challenges facing America, and more importantly, counter Chinese technological expansion?
The “Non-Proliferation” camp: Computing power as a strategic asset
Critics, such as Representative Khanna, argue that Washington is “exporting” its strategic computing power to the Middle East. For them, compute—the data processing capability that powers everything from simple calculations to advanced AI systems—is a crucial lever of national power that must be distributed very carefully. The United States currently holds approximately 75% of the world’s leading computing capacity, and this camp believes preserving that advantage is the key to winning the race toward artificial general intelligence (AGI).
The “Global Proliferation” camp: Energy is the new bottleneck
On the other side, David Sacks and Sriram Krishnan, senior White House advisors for AI, argue that the bottleneck in AI infrastructure is no longer chips, but electricity. U.S. data center demand is projected to consume up to 9% of the national grid by 2030, and America cannot build high-voltage lines and substations quickly enough to keep pace with this demand.
In contrast, Gulf partners can fund and operate gigawatt-scale data complexes at a low cost of $0.03 per kilowatt-hour and bring them online within months. From Sacks’s perspective, leveraging this excess energy is the only reliable way to expand computing capacity quickly enough to maintain U.S. superiority.
The real Chinese threat
The proliferation camp adds that the technological gap between the U.S. and China is much narrower than non-proliferation advocates believe. Huawei’s Ascend chips, paired with DeepSeek models, provide Beijing with an integrated domestic “tech stack” that has already begun spreading abroad. As Huawei nears parity with American chips and China expands into markets like Pakistan, Thailand, and Egypt, countries around the world are increasingly faced with a choice between two stacks: the American stack or the Chinese stack.
If the United States rejects willing partners like the UAE and Saudi Arabia—who have shown a clear preference for the American stack—it risks creating a vacuum that China will rush to fill.
Computing power as a commodity, not a secret
Some argue that cooperation should be limited to countries that share America’s democratic values. But this view assumes that computing power is a classified asset. In reality, it is a commodity, much like oil in the 20th century. Value lies not in secrecy, but in scale, volume, and reliable access. What matters is who can build and scale infrastructure, not who hides it behind a firewall.
Through these partnerships, Gulf states can position themselves as backend computing service providers for countries in emerging markets that lack the capacity to train and run AI systems—the very core of what is known as “sovereign AI.”
Conclusion: Leading global proliferation or watching China win
U.S.-Gulf partnerships should not be viewed as exporting computing capacity abroad, but rather as creating capacity that did not exist in the first place. These deals represent Washington’s most realistic path to expanding its computing capacity in partnership with willing and capable nations. The choice is not between trusting authoritarian regimes or upholding democratic values, but between leading the global proliferation of U.S. AI infrastructure or standing idly by while the Chinese stack becomes the default choice for the world.
