When President Donald Trump warned on November 20, 2026, about imposing a heavy tax on American corporations constructing foreign data centers, he framed the pitch around domestic economic growth. Speaking to reporters at the White House, he emphasized that if companies bypass states like Texas, California, or West Virginia, plenty of domestic regions are ready to step in because they possess abundant energy and resources.

While Trump specifically singled out Google's €13 billion clean energy and data center initiative in Finland—noting his displeasure with the overseas investment—his rhetoric reverberates profoundly across the Gulf. Saudi Arabia, the United Arab Emirates, and Nordic countries pitch themselves to the artificial intelligence sector similarly, but with the added advantage of direct fiber-optic cable routes to European clientele. Because the Gulf's monumental AI strategies rely entirely on the American technology firms currently under scrutiny, Trump's vague threat introduces immense uncertainty.

The Deep Reliance of Gulf AI Infrastructure on US Firms and Hardware

Consider the region's two cornerstone initiatives. Amazon Web Services (AWS) and HUMAIN—the artificial intelligence enterprise managed by Saudi Arabia's Public Investment Fund—announced plans to launch up to 50 megawatts within a designated Saudi AI Zone by 2028. This infrastructure will utilize both Amazon's proprietary Trainium processors and Nvidia hardware. Meanwhile, OpenAI unveiled Stargate UAE in May 2025 in collaboration with G42, Oracle, Nvidia, Cisco, and SoftBank, targeting a 200-megawatt rollout. OpenAI emphasized that this venture was coordinated closely with Washington, expressing gratitude for Trump's backing during high-profile diplomatic tours.

Emirati champions have also met rigorous compliance prerequisites. G42 committed to eliminating Chinese hardware from its systems in 2023, prompting a $1.5 billion investment from Microsoft the following year. Microsoft has since expanded its operations significantly in Abu Dhabi. Similarly, HUMAIN CEO Tareq Amin assured markets in October 2025 that the firm would avoid Huawei equipment entirely. Shortly thereafter, the U.S. Commerce Department authorized both G42 and HUMAIN to procure up to 35,000 Nvidia GB300 Blackwell chips each, subject to stringent security and reporting protocols.

Consequently, the November 20 penalty warning carries severe implications. Should financial penalties target American corporate expenditures abroad, the AWS zone in Saudi Arabia and the Stargate facilities in Abu Dhabi risk being penalized alongside Google's Finnish operations. The lack of precise definitions regarding the proposed tax creates an immediate financial burden for entities financing these Gulf campuses.

Why Export Controls Overshadow Potential Data Center Taxes

While congressional approval is required to enact new tax laws—though the White House could pursue alternative tariff or sanction mechanisms—Washington's primary instrument of influence throughout the year has been the export license.

On July 10, the Commerce Department reclassified the UAE into an advantageous export-control tier, permitting the Emirati government and select entities to acquire advanced computing components license-free. The directive explicitly named the UAE government, G42, Core42, alongside major U.S. players like Microsoft, Oracle, and OpenAI. However, this approval carries an expiration date of April 6, 2027, requiring Emirati firms to transition into U.S.-aligned structures or reapply. Conversely, Saudi Arabia received no equivalent concession, leaving HUMAIN dependent on individual U.S. authorizations.

Furthermore, a Commerce Department directive enacted on November 19 mandated export licenses for advanced chips destined for foreign data centers supplying massive computing power to entities headquartered in China and other restricted nations. This ruling superseded the previous administration's AI Diffusion Rule, prompting warnings from Alibaba regarding potential disruptions to its international cloud expansion across Türkiye, Finland, and the Netherlands.

These mechanisms demonstrate Washington's overarching authority in dictating where advanced AI hardware operates and under what conditions. While tax proposals would target the developers, export controls regulate the silicon itself, blending trade policy with geopolitical strategy.

Serving European markets from Riyadh or Frankfurt involves distinct hurdles. Neither Saudi Arabia nor the UAE features on the European Commission's adequacy list under the General Data Protection Regulation (GDPR). Consequently, transferring European personal data requires complex legal frameworks, standard contractual clauses, and local legal assessments, adding financial and operational friction.

Consequently, Gulf facilities are largely restricted to model training, batch inference, and non-regulated tasks that tolerate latency. The market demand for such capacity remains unproven, as no major European enterprise has publicly committed to utilizing Gulf infrastructure.

A notable partnership to monitor is a smaller initiative at NEOM. In October 2025, HUMAIN and DataVolt established a framework partnering with Amsterdam-headquartered Nebius as an anchor client for up to 50 megawatts at Oxagon on the Red Sea, slated for phased deployment starting in late 2027. While NEOM highlights the site's digital connectivity, the project still depends on American hardware licenses and remains subject to regulations concerning Chinese end-users.

Evaluating Exposure to Future Tax and Regulatory Shifts

The Gulf commands the energy, land, and financial capital necessary for massive AI expansion, yet regulatory permission remains firmly in Washington's hands. As policy evolves, the viability of these international joint ventures will hinge on whether a formal tax materializes, whether Riyadh secures licensing parity with Abu Dhabi, and whether non-American enterprises ultimately sign commercial contracts for Gulf-based capacity.