The UAE is preparing to extend its successful nation-building and investment strategy into the artificial intelligence sector, with ambitions to emerge as a leading exporter of advanced AI technologies and innovation.
How the UAE plans to apply its long-standing growth model to become a major player in AI exports.
The United Arab Emirates has emerged as one of the global leaders in artificial intelligence adoption, with the highest per capita usage of AI technologies worldwide, according to Microsoft’s latest AI diffusion report. This reflects a broader national transformation that has been unfolding over several decades, shifting the country’s identity from a resource-based economy to a diversified, technology-driven hub.
For much of its modern history, the UAE built its economic strength on the efficient production and movement of physical goods and resources. Oil and gas exports formed the backbone of early development, while industrial expansion introduced aluminium production, large-scale manufacturing, and infrastructure development. Over time, the nation also established itself as a critical global transit point, leveraging its geographic position to become a central node for international aviation, maritime trade, and logistics.
Airports, ports, and free zones became symbols of this strategic approach. Rather than waiting for demand to emerge organically, the UAE invested heavily in infrastructure ahead of time, creating capacity that could support future global growth. This forward-looking model allowed the country to position itself as a bridge between East and West, facilitating the movement of people, goods, and capital on a massive scale.
Today, that same philosophy is being applied to an entirely different domain: artificial intelligence. According to industry leaders, the UAE is now extending its long-established strategy of early investment and infrastructure-led development into the digital and AI economy.
Al Kaissi, speaking about this transformation, explained that the country is effectively following the same playbook that once powered its success in traditional industries. However, the focus has now shifted from physical commodities to data, computation, and machine intelligence.
He described this evolution as a shift from exporting tangible goods such as oil, metals, and industrial materials to producing and exporting intelligence itself. Over time, the UAE diversified its economy into financial services, aviation, logistics, aluminium production, and advanced composites. Now, he said, the next stage of this progression is centered on “manufacturing intelligence” that can serve both domestic needs and international markets.
This transition reflects a broader global trend where information and computation are becoming the most valuable economic resources. Unlike traditional goods, artificial intelligence does not have a physical form, yet its impact spans across industries, from healthcare and education to transportation, finance, and governance.
Al Kaissi emphasized that the scale of future demand for AI systems is expected to grow exponentially. In his view, the world is heading toward a future where billions—and eventually trillions—of AI-generated “tokens” will be processed daily. These tokens represent the fundamental units of machine-generated language and reasoning used in AI systems. As more applications, tools, and autonomous agents are developed, the demand for computational intelligence continues to expand without clear limits.
He noted that this creates a fundamentally different kind of market compared to traditional industries. Unlike commodities that have physical constraints, AI systems generate continuous demand through usage. Every new model or application does not reduce demand; instead, it increases it by creating new possibilities, new questions, and new interactions.
In this context, he described demand for artificial intelligence as “structurally insatiable,” meaning it is built into the nature of the technology itself. The more AI is used, the more it is required, as each system feeds into further innovation and complexity.
To prepare for this future, he explained that organizations like G42 have built their strategy around two interconnected systems. The first is what he referred to as a “token factory,” which focuses on scaling the production of machine intelligence at an industrial level. This involves building large-scale computing infrastructure capable of generating vast amounts of AI output efficiently and reliably.
The second component is an “agent factory,” which focuses on deploying that intelligence into real-world applications. These agents are designed to operate autonomously or semi-autonomously, performing tasks, making decisions, and interacting with users or systems in meaningful ways.
Together, these two systems form a loop: one produces intelligence at scale, while the other applies it in practical environments, generating new data and demand in return. This model reflects a deliberate effort to treat artificial intelligence not as a single product, but as a continuous, evolving industrial ecosystem.
The UAE’s approach to AI development therefore mirrors its earlier economic strategy in important ways. Just as the country invested in ports, airports, and industrial zones before demand fully materialized, it is now investing in computing infrastructure, data ecosystems, and AI research capabilities ahead of global saturation.
This proactive stance is designed to position the country as a central hub in the global AI economy, much like it once did in aviation and logistics. By building capacity early, the UAE aims to attract innovation, talent, and enterprise activity as global demand for AI continues to accelerate.
At its core, this strategy reflects a consistent national philosophy: anticipate future global needs, invest in enabling infrastructure, and integrate deeply into international systems of exchange. Whether in physical trade or digital intelligence, the goal remains the same—become an essential node in the world’s most important flows.
What a token factory actually is
The idea being described can be understood through a straightforward but powerful shift in perspective. In this framing, an artificial intelligence data centre is not just a storage or processing facility, but effectively a modern industrial factory. Instead of manufacturing physical goods, it produces digital output. Energy and computing power are fed into the system as raw inputs, and what emerges on the other side are “tokens”—the fundamental units of output generated by AI systems whenever they process information, generate responses, or carry out tasks.
These tokens represent the functional language of artificial intelligence. They are the building blocks of reasoning, communication, and decision-making within AI models. As usage scales up, the production of tokens becomes continuous and industrial in nature. When systems are optimized to produce them efficiently at massive scale, tokens begin to resemble a commodity—something standardized, measurable, and tradable.
In this model, ownership of the infrastructure becomes strategically critical. Controlling the facilities that generate these outputs means controlling the supply of intelligence itself. Just as industrial economies historically depended on ownership of factories producing steel, energy, or manufactured goods, the emerging AI economy depends on who owns and operates the infrastructure that produces machine intelligence.
Within this ecosystem, Core42 plays a central role as the digital infrastructure arm of G42’s broader “token factory” concept. Working alongside Khazna, which is responsible for developing and managing the physical data centre infrastructure, Core42 provides the computing backbone that enables large-scale AI operations. Together, these entities form an integrated system that connects physical infrastructure with advanced digital capabilities.
This infrastructure is designed to serve as a sovereign AI and cloud foundation for the UAE. It supports a wide range of critical sectors, including government institutions, healthcare systems, energy providers, and highly regulated industries. A key feature of this model is data sovereignty, ensuring that sensitive information remains within national infrastructure rather than being processed through external or foreign platforms.
By maintaining control over both the physical and digital layers of AI infrastructure, the system is designed to reduce external dependencies while strengthening national technological autonomy. This approach ensures that essential services and data-intensive operations can function securely within a controlled environment tailored to domestic regulatory and security requirements.
Al Kaissi highlighted this principle by emphasizing the importance of capacity as a form of strategic power. In his view, technological leadership is not determined solely by innovation or ideas, but by the ability to support and scale those ideas through infrastructure. He noted that organizations and countries face a fundamental choice: either rely on external providers for computing resources and risk long-term dependency, or invest in building their own integrated capabilities across both infrastructure and intelligence systems.
Without such investment, he argued, entities are effectively renting either computational power or intelligence itself from external providers. This creates structural reliance on others for access to essential AI capabilities. By contrast, building indigenous systems allows a country or organization to maintain control over both production and deployment of artificial intelligence.
A major physical expression of this strategy is the large-scale five-gigawatt AI campus currently under development in Abu Dhabi. This facility is envisioned as a central hub for AI computation and data processing, forming the core of the country’s long-term ambition in artificial intelligence infrastructure.
The output generated by this campus is expected to serve both domestic and international needs. On the domestic side, a significant portion of its capacity will be consumed locally. The UAE government is actively pursuing a strategy of becoming “AI-native,” integrating artificial intelligence deeply into public services, governance systems, and economic planning. This aligns with a broader national direction toward digital transformation across sectors.
In addition, the country’s population already demonstrates exceptionally high levels of AI usage. According to Microsoft’s AI diffusion report, the UAE ranks among the highest globally in per capita AI adoption. This indicates strong local demand for AI-powered tools, applications, and services, which will naturally absorb a portion of the system’s output.
However, the scale of the infrastructure being built far exceeds domestic consumption alone. As a result, any surplus capacity generated by the AI campus is expected to be directed toward international markets. This positions the UAE not only as a consumer of AI technology but also as a potential exporter of machine intelligence.
In this way, the system reflects a dual-purpose model: serving national requirements while also participating in the global AI economy. By producing intelligence at industrial scale, the infrastructure becomes a platform for both domestic transformation and international trade.
Ultimately, the strategy combines physical infrastructure, sovereign computing capability, and large-scale AI deployment into a unified vision. It reflects a broader ambition to establish long-term leadership in artificial intelligence by controlling the full stack—from energy and data centres to models and applications—within a single integrated ecosystem.
Geography and the distribution model
Al Kaissi explained that the UAE’s geographic position gives it a major strategic advantage for digital infrastructure. He noted that from within a 3,200-kilometre radius, it is possible to reach around 3.9 billion people—nearly half of the global population—while still maintaining strong connectivity, low latency, and high data transfer speeds. In his view, this makes the UAE an ideal hub for large-scale computing and AI services.
He compared this advantage to the factors that helped the UAE establish dominance in sectors like aviation and maritime logistics. Just as the success of airlines and ports was shaped by geography, early investment, and long-term infrastructure planning, he said the same principles are now being applied to artificial intelligence. He pointed to examples such as the development of Jebel Ali Port, which was built ahead of clear demand at the time but later became a critical global trade hub. According to him, this “build ahead of demand” approach proved highly successful in the past and continues to define the country’s strategy.
However, he emphasized that the current opportunity is even larger in scale. Unlike physical trade, demand for computing power and artificial intelligence is expanding rapidly and continuously across the world, creating exponential growth potential. In this context, he suggested that the UAE’s early investment in AI infrastructure positions it to serve as a central provider of global digital capacity in the future.
He also described how this infrastructure is being connected through what Core42 refers to as an “intelligence grid.” This system is designed as a distributed network of sovereign AI nodes that can be deployed across different countries while still maintaining strict data sovereignty standards. These nodes operate under formal government-to-government arrangements, often structured as “Digital Embassies,” which ensure that sensitive information remains protected according to the rules of the originating country, regardless of where the physical infrastructure is located.
Core42 already manages and operates computing clusters in several regions, including the UAE, the United States, and parts of Europe. These existing deployments form the early foundation of the broader intelligence grid architecture. Looking ahead, further expansion is planned over the coming years, with additional capacity and new geographic locations expected to be added by 2027 as part of the company’s long-term global growth strategy.
The data no one else has
A key factor strengthening the UAE’s position is not only its computing capacity, but also the depth and uniqueness of the datasets it has accumulated over time through a combination of policy decisions and national-scale programs. These datasets, according to experts, are difficult for other countries to replicate due to their specific origins and long-term consistency.
For example, one distinctive case is the routine medical screening process for expatriates entering the country, where chest X-rays are commonly required for tuberculosis checks. In many other nations, such scans are typically conducted only when there is a clinical suspicion of illness. This difference in approach has unintentionally created a highly valuable medical dataset.
Al Kaissi explained that this has resulted in what may be one of the few large-scale collections of chest X-rays representing healthy individuals. He noted that in most healthcare systems, imaging data is skewed toward patients already showing symptoms, which limits its usefulness for certain types of artificial intelligence applications. By contrast, the UAE’s dataset allows AI systems to better understand what normal health conditions look like, improving the ability to detect anomalies more accurately and at an earlier stage. This can help support doctors by enhancing both diagnostic speed and precision.
He further pointed out that this medical data is complemented by other advanced national datasets. These include genomic information collected through the Emirati Genome Programme, which provides a detailed genetic map of the local population. Such data is particularly valuable for areas like population health management, personalized or precision medicine, and the development of new pharmaceuticals.
In addition to healthcare-related information, the UAE also holds extensive geospatial and environmental datasets. Satellite imagery and analytics provided through initiatives such as Space42 contribute to large-scale mapping and earth observation capabilities. Meanwhile, decades of geological data gathered from the country’s oil and gas exploration activities offer deep insights into subsurface structures, which are essential for applications such as reservoir modeling and seismic analysis.
When combined, these diverse datasets create a powerful foundation for innovation across multiple industries. However, fully utilizing them requires significant computing infrastructure, as well as secure sovereign cloud environments capable of processing sensitive information at scale while maintaining national control over data.
Al Kaissi noted that regulatory environments in many other countries often impose stricter limitations on how such datasets can be accessed or used. Privacy concerns, legal restrictions, and fragmented data ownership can make it difficult to build similarly integrated systems elsewhere.
In contrast, he explained that the UAE has adopted a more enabling approach, where government institutions actively seek to support technological innovation while still maintaining appropriate safeguards around privacy, security, and ethical use. This balance, he said, is designed to encourage the development of solutions that improve outcomes in sectors such as healthcare, infrastructure, and public services.
He added that the initial phase of the large AI campus project is already progressing, with the first 200 megawatts of capacity nearing completion. Plans are already in place to expand the facility significantly, with a long-term target of reaching five gigawatts of total capacity.
Although the commercial export model for this AI-driven infrastructure is still in its early stages, the underlying strategy follows a familiar pattern. It mirrors the same development philosophy that transformed the UAE’s geography into a global aviation and logistics hub—turning limited natural advantages into world-leading infrastructure through early investment, long-term planning, and large-scale execution.
