According to reports from the media citing informed sources, some core officials of the Trump administration in the United States are delaying a purchasing agreement that would allow UAE tech companies and government-supported giants to spend hundreds of billions of dollars to purchase NVIDIA’s (NVDA.US) most advanced Blackwell architecture AI chips due to National Security concerns.
It is understood that U.S. President Donald Trump announced this preliminary purchasing agreement during his visit to the Middle East in May, and both parties had hoped to quickly finalize the details. However, up to now, the two countries have not reached a final agreement, in part due to concerns from some U.S. officials that competitors such as China might take the opportunity to acquire these cutting-edge American high-performance chips.
The media cited informed sources reporting that unless the UAE agrees to modify the terms to address the National Security concerns of U.S. officials, the divergence is unlikely to be resolved.
One of the options being discussed by officials is to cancel the permission for the AI startup G42, headquartered in Abu Dhabi, to directly acquire or use NVIDIA AI chip clusters. According to the original plan, G42 was expected to receive about 20% of NVIDIA’s AI chip allocation.
Currently, the U.S. Department of Commerce does not intend to approve the supply of any AI chips to G42, although there is still a possibility of future clearance, the media reported.
NVIDIA, the U.S. Department of Commerce, G42, OpenAI, and Microsoft have not yet responded to any requests for comment.
A spokesperson for U.S. Commerce Secretary Howard Lutnick stated that Lutnick is “confident that the implementation plan for the agreement signed in the UAE will proceed smoothly.”
The UAE is also optimistic about the negotiations related to trade with the U.S. government and agreements for purchasing AI chips. The UAE Ambassador to the U.S., Yousef Al Otaiba, stated in a declaration that this large-scale purchasing agreement “will bring great benefits to both countries.”
On Tuesday, Lutnick met with UAE representatives at a technology and energy conference in Pittsburgh to discuss how to advance the agreement.
However, slow progress in negotiations has led to significant divisions within the U.S. government and has frustrated some tech industry executives hoping to expand their AI-related business overseas.
According to reports, NVIDIA CEO Jensen Huang emphasized the importance of the agreement during a recent meeting with President Trump.
Some government officials are also concerned that some Chinese tech companies are attempting to sell AI chips in the Middle East and may seize market opportunities while the agreement is delayed.
The core debate of the agreement centers on whether to directly acquire NVIDIA’s cutting-edge Blackwell architecture AI chips and NVL server clusters based on that architecture, which will provide the strongest computational support for large-scale AI data centers globally.
Last month, media reports indicated that the construction progress of a billion-dollar AI data center project known as “Stargate UAE,” part of the Middle Eastern version of the Stargate initiative, is indefinitely delayed due to ongoing national security concerns. The project is collaboratively built by NVIDIA, Oracle, OpenAI, Cisco, SoftBank Group, and G42 from the UAE, and was announced after Trump’s visit to the UAE.
OpenAI and G42 are both supported by Microsoft’s cloud computing services and cloud AI computational resources.
According to preliminary terms, UAE companies will receive hundreds of thousands of NVIDIA chips over the coming years and assist NVIDIA and American cloud computing giants in building more and larger AI data centers. According to the agreement, most AI chips will be used for data centers in the Middle East operated by American technology companies.
American tech giants like OpenAI and Microsoft are expected to operate multiple large AI data centers in the UAE. Insiders point out that exporting NVIDIA AI chips to the UAE, whether delivered to UAE companies or American tech giants, requires export licenses approved by multiple departments, which may intensify national security scrutiny.
In addition to being allocated about 20% of NVIDIA AI chips, UAE AI startup G42 may also deeply engage in building the data centers operated by American cloud computing giants in the UAE.
Some American officials are concerned that countries like China may access the most advanced AI chips through G42 or other entities headquartered in the UAE, thereby obtaining critical technologies for the development and iteration of large AI models.
Directly canceling supplies to G42 could anger the UAE government; the UAE views this agreement as key to maintaining a lead in the global AI race, and believes G42’s participation is core to the framework agreement made in May.
David Sachs, the White House AI lead, stated at the Pittsburgh summit: “If we don’t provide this technology, global competitors will.” He believes concerns about chip transfers have been “greatly exaggerated.”
Sachs is one of the chief architects of the US-UAE AI chip agreement, while Lutnick facilitated the signing of the preliminary agreement. Reports add that after the signing of the agreement, key officials responsible for advancing details from both countries have rarely met again.
Global demand for AI computing power continues to surge.
As the U.S. government delays NVIDIA AI chip exports to the UAE due to National Security concerns, AI computing power demand continues to show explosive growth. Pennsylvania, the most typical “swing state” in the U.S., will receive a $92 billion investment to build a “super AI center.” Meta, Facebook’s parent company, plans to invest hundreds of billions of dollars to construct several large-scale AI data centers to support its artificial intelligence technology updates and iterations, with the goal of achieving artificial general intelligence (AGI), and the first data center is expected to be operational next year.
Global cloud computing giant Amazon AWS announced a large-scale collaboration agreement with South Korea’s SK Group, under which AWS will build the largest data center in South Korea, with a capacity of up to 100 megawatts (MW). Additionally, just this week, “AI chip leader” NVIDIA announced the resumption of H20 AI GPU sales in the Chinese market.
Meta CEO Mark Zuckerberg stated on Monday that the technology giant will invest hundreds of billions of dollars to build several large data centers to support its development of artificial intelligence, aiming to achieve artificial general intelligence (AGI), with the first super AI data center expected to be operational next year. Wall Street Analysts generally interpret Zuckerberg’s statement as showcasing confidence in a substantial increase in mid-term cash flow, while also signaling that the demand for AI computing power centered around NVIDIA AI GPUs remains incredibly strong and that the AI capital expenditure cycle is far from over.
The generative AI leader, dubbed the ‘rival of OpenAI,’ Anthropic predicts that by 2027, AI large models will have the capability to automate nearly all white-collar jobs, thus the demand for AI computing power coming from inference will be vast, expected to drive exponential growth in the AI computing infrastructure market. ‘AI inference systems’ is also considered by Jensen Huang to be the largest source of future revenue for NVIDIA.
Taiwan Semiconductor’s latest performance report released on Thursday indicates that the surge in AI computing power demand has driven Taiwan Semiconductor’s Q2 net income to soar by 61%. Taiwan Semiconductor expects a roughly 30% increase in sales calculated in U.S. dollars by 2025, significantly higher than the previous growth forecast of “close to the mid-20% range,” mainly benefiting from the surging orders for AI chips based on 3nm and 5nm advanced process technologies. As the demand for AI computing power remains incredibly strong, Taiwan Semiconductor is actively expanding its backend capacity to enhance the actual output of CoWoS advanced packaging, primarily for NVIDIA AI GPU capacity, which also demonstrates the company’s confidence that the incredibly strong demand for AI chips will continue until 2026.