Why a Saudi Nuclear Program Is Not Worth Ending US Nonproliferation Policy

Excerpt After years of intense diplomacy in the 1990s, the United States finally persuaded Russia and China to halt further transfers of sensitive uranium enrichment and plutonium reprocessing technologies to states like Iran and Pakistan. These dual-use capabilities can fuel both nuclear weapons and peaceful nuclear reactors. In so doing, Washington built a fragile consensus that they are simply too dangerous to spread. A generation later, America itself is poised to break that hard-won precedent by authorizing a uranium enrichment plant on Saudi soil. What follows is predictable: Moscow and Beijing, no longer constrained by American example, will begin competing to offer enrichment and reprocessing to prospective customers as sweeteners in their own commercial nuclear deals. The carefully constructed barrier against the spread of nuclear weapons fuel will collapse, and several new states will edge closer to obtaining the bomb. That is the risk reportedly embedded in the current US-Saudi nuclear cooperation accord—a “123 agreement” under Section 123 of the US Atomic Energy Act (AEA), which establishes terms for foreign nuclear cooperation. The agreement or accompanying side deals would reportedly require a two-year joint study to determine whether the United States should build and operate an enrichment plant in Saudi Arabia. Opening the door to enrichment in the kingdom—even under American operational control—would dismantle several decades of deliberate US policy against the proliferation of sensitive fuel-cycle technologies. President Donald Trump can secure US nonproliferation objectives by renegotiating the deal before submitting it to Congress in September. American companies can still build reactors and supply fuel fabricated in the United States, but enrichment itself should never leave US soil. Andrea Stricker is deputy director of the Nonproliferation Program at the Foundation for Defense of Democracies. Follow her on X: @StrickerNonpro.