UAE Reviews 5-Gigawatt AI Data Center Plan, Shifts to Distributed Sites

GeopoliticsProduct / TechIndustry Impact 4
โดย ロイター·AE·Read original
Summary · why it matters

The United Arab Emirates is quietly reviewing plans for a 5-gigawatt artificial intelligence data center, one of the largest outside the United States, according to six people familiar with the matter. The original concept called for building a giant 10-square-mile data center campus in Abu Dhabi, but after reconsideration it is likely to be changed to a plan to distribute multiple data centers across the UAE. Authorities are also considering air defense systems and building some facilities underground to protect them from drone and missile attacks, and this is the first time a possible change to the plan's basic design has been reported. The plan was announced during U.S. President Trump's visit to the UAE last year and is central to a strategy led by state-linked company G42 aimed at moving away from oil dependence and becoming an AI power. In exchange for access to advanced technology such as AI semiconductors from U.S. firm Nvidia, the UAE has agreed to scale back its AI ties with China and to invest in the data center business of the U.S. Stargate initiative. The review began shortly after Iran launched missile and drone attacks on Gulf states hosting U.S. forces in retaliation for U.S. and Israeli airstrikes, and in March two Amazon Web Services data centers in the UAE and one in Bahrain were attacked. According to two of the people, the planned construction site is near Al Dhafra Air Base on the outskirts of Abu Dhabi, a location that had not previously been disclosed.

Impact on stocks 2

Artificial Intelligence · 2 stocks
Amazon.com Inc
AMZN
▼ NegativeGeopoliticsrelevance

Two AWS data centers in the UAE were attacked in March, highlighting physical security risk to Amazon's regional infrastructure.

NVIDIA Corporation
NVDA
± MixedGeopoliticsrelevance

UAE's review of its 5GW AI data center plan could alter Nvidia chip demand, but the article only notes Nvidia as the technology supplier in the broader deal.

Theme Impact 9

Off-coverage companies 1

G42Private± Mixed
Geopoliticsrelevance

G42's flagship 5GW AI data center plan is being reviewed and likely redesigned into distributed sites amid regional attack concerns.

Related news

Digital Realty Launches ServiceFabric MCP Across 800 Data Centers

Digital Realty made ServiceFabric MCP available on September 15, a software layer that lets AI agents design, monitor and troubleshoot network connections across more than 800 data centers, including third-party sites. The launch follows second quarter results reported on July 23, when Core FFO per share, excluding net promote, rose to $2.13 from $1.87 a year earlier, while the headline figure of $2.65 included a $188 million net promote. Renewal leases in the quarter were signed at rates 25.4% higher on a cash basis, and signed leases waiting to start added up to a $1.9 billion backlog of annualized base rent at 100% share. Management lifted its 2026 Core FFO per share outlook, excluding net promote, to $8.15 to $8.20, even as the company carried about $18.6 billion of debt at June 30, 2026 and set its 2026 development spending outlook, net of partner contributions, at $4.25 billion to $4.75 billion. Digital Realty calls MCP an emerging standard still being validated, and the announcement puts no dollar figure on what it could add to revenue.
Insider Monkey·3hRead more →
impact 5

US Hyperscalers to Spend Up to $725 Billion on AI Infrastructure in 2026

The top five US hyperscalers are projecting a combined capital expenditure of $660 billion to $725 billion for 2026, nearly double their 2025 outlays, as the AI build-out shifts from software to physical infrastructure. Microsoft is guiding for roughly $175 billion in adjusted capital expenditure for both FY2026 and FY2027, with two-thirds of quarterly spend going to short-lived assets like CPUs and GPUs and the rest to long-lived data center infrastructure, and it added 1 gigawatt of capacity in Q3 FY2026, doubling its global footprint in two years. Amazon AWS has raised its 2026 capex guidance to approximately $220 billion, with CEO Andy Jassy saying AI capacity is expected to remain constrained through 2027 and contracted demand extending into 2028. Meta saw profit drop 14% in Q2 2026 despite a 28% revenue increase as its build-out, including a 1 gigawatt data center in Ohio and a Louisiana facility that could scale to 5 gigawatts, compressed margins, while Alphabet raised its 2026 capex guidance to as much as $205 billion and its Google Cloud backlog more than doubled year-over-year to $240 billion. The Stargate joint venture involving Oracle, OpenAI and others targets up to $500 billion in infrastructure investment by 2029, and Oracle's FY2026 capex reached $55.7 billion, more than doubling from the previous year.
Yahoo Finance·4hRead more →
impact 4

Citi: Data Center Opposition Has Not Weakened AI Construction Pipeline

Citi says growing political opposition to artificial intelligence infrastructure ahead of the November U.S. midterm elections has not materially weakened the data center construction pipeline. Data center development has become a bipartisan flashpoint, with local governments introducing moratoriums and at least 15 state legislatures proposing tighter regulatory restrictions, yet spending remains strong as AI infrastructure demand continues to support development. The impact has been concentrated among speculative and early-stage projects, which are increasingly delayed or cancelled during local approval processes, while late-stage developments that have already secured sites and grid connections continue to move ahead. Hyperscalers are seeking workarounds to power constraints and local restrictions, with Amazon pursuing direct investment in nuclear development with Dominion Energy and Meta securing a major nuclear power purchase agreement with Constellation Energy. Citi does not expect another market shock comparable to the emergence of DeepSeek, arguing investors have already adjusted to the prospect of highly efficient Chinese models, though it flags a potentially greater risk from governments restricting models deemed too dangerous, which could abruptly create excess computing capacity.
Investing.com·5hRead more →