The UAE has set its sights on mass-scale AI production. Through the Stargate data centre project, it aims to generate 60 trillion tokens, seeking to establish itself as the world’s emerging “factory of intelligence” .
Speaking at the Milken Middle East and Africa Summit in Abu Dhabi, UAE Minister of State for AI, Digital Economy and Remote Work Applications Omar Al Olama said that reaching the 60-trillion-token mark would represent roughly 60% of global AI production, signalling the Emirates’ intention to occupy a commanding spot in the world’s AI infrastructure map.
Why Tokens Matter
In artificial intelligence, tokens are the basic building blocks that AI models use to process information. Every word, punctuation mark, space, sound clip, image fragment, or video element is broken down and converted into tokens through a process known as tokenisation. The higher the number of tokens a system can generate or process, the greater its capacity to deliver complex AI outputs, from sophisticated text and image creation to large-scale data analysis and simulations.
Modern “AI factories” are specialized data centres purpose-built to perform this token processing at extreme speed and volume. Graphics-chip leader Nvidia has described these facilities as next-generation industrial engines, capable of turning raw computation into usable intelligence, effectively translating machine language into real-world insights and decisions.
According to Al Olama, tokens themselves are becoming a central economic unit of the AI era.
“The currency of the future is tokens,” he said, emphasizing their role in driving productivity improvements, quality-of-life gains, and smarter decision-making across industries.
Stargate and the Scale of the Buildout
At the centre of the UAE’s strategy is Stargate, a new hyperscale facility planned to reach 1 gigawatt (GW) of operating capacity. Initial operations are targeted at 200 megawatts in 2026, with Stargate forming part of a much larger 5GW campus development in Abu Dhabi. Officials describe the project as a capacity-led approach that will expand in line with real-world demand rather than attempting to overshoot immediately.
While no formal timeline has been provided for hitting the 60-trillion-token target, the government maintains that the necessary investments and partnerships are already being lined up. As Al Olama noted, execution depends on several moving parts, including timely access to equipment, infrastructure buildout, and partner coordination, but confidence remains high as long as those fundamentals remain aligned.
The UAE also acknowledges that building such massive computing capacity will be adaptive rather than fixed. Data centre output can be increased or reduced over time as market conditions evolve, even if today’s blueprint remains centered on the full 5GW build.
Is This an AI Bubble?
With global markets increasingly questioning whether the AI investment boom could resemble past technology bubbles, UAE leadership is pushing back against the narrative of irrational exuberance.
Al Olama drew a parallel to the dot-com era, arguing that time horizon matters. In short windows of a few years, early internet valuations looked inflated. Over decades, however, the technology delivered, and reshaped nearly every sector of the global economy. Today’s largest corporations, he noted, are fundamentally internet-based enterprises.
Applied to AI, the message is clear: short-term volatility may fuel bubble talk, but the long-term trajectory points toward lasting transformation rather than collapse.
Beyond its ambitions as a global AI producer, the UAE is already seeing measurable financial returns from deploying AI locally, particularly in energy production. According to Al Olama, AI tools have generated approximately Dh500 million ($136 million) in cost savings so far by improving efficiency and lowering production expenses.
Major firms such as Adnoc Distribution, the fuel retail arm of the national energy group ADNOC, are integrating AI systems as part of broader operational streamlining programs. These tools track output, optimize workflows, and reduce inefficiencies across extraction, logistics, and retail operations.
Al Olama explained the approach succinctly: energy performance is measured through clear metrics, barrels produced and the cost required to produce them and AI is now being used to bring those numbers down.
The expectation is that efficiency gains will continue compounding over time, converting early operational savings into long-term productivity growth and stronger economic returns.
The UAE’s 60-trillion-token ambition is a deeper strategy to anchor the country within the physical backbone of global AI as a mass producer of computational intelligence itself. Through large-scale data centres, international partnerships, and sector-specific AI deployment, the Emirates aims to lock in long-term relevance as AI transitions from experimental technology into core economic infrastructure.
