Nvidia’s graphics processing units (GPUs)—specialized computer chips used for high-performance computing—are helping the United Arab Emirates (UAE) rise as a leader in artificial intelligence (AI). Recently, Nvidia invested $7 billion to acquire a stake in a U.S. startup called Poolside and to license the tools used to build AI models. Poolside’s new model, expected to be called Nemotron 4, will be available for free by the end of the year. According to Marc Einstein, an analyst at Counterpoint Research, the model is intended for users who already have Nvidia hardware. The UAE is constructing a flagship data center that will use 400,000 Nvidia chips. The country has been positioning itself as a neutral ground between the U.S. and China, purchasing hardware from the former and, until recently, technology from the latter. To reduce dependence on foreign software, the UAE developed its own AI model called Falcon. However, this neutrality has been shifting as G42, a state-backed AI company, severed ties with Chinese firms to secure access to Nvidia’s chips. The UAE’s closer alignment with Nvidia may deepen if it adopts the new Nemotron 4 model as well. “The last missing piece was the model layer,” said Mohammed Soliman, a senior fellow at the Middle East Institute. “Nvidia already owns the chips. Now it is scaling the factory that produces the models that run on those chips, and it is releasing them as open weights.” This means the models’ internal structures, or “weights,” will be available for others to use and modify. Two major institutions driving the UAE’s AI development have different approaches. The Technology Innovation Institute (TII), the government lab behind Falcon, insists on building AI models from scratch to ensure full technological control. In contrast, G42 has aligned with Nvidia’s push for open, freely available models. Hakim Hacid, TII’s chief researcher, emphasized the importance of mastering AI technology independently. “The value lies in mastering the underlying technology and retaining control across the entire AI stack,” he said. G42, which has not commented on whether it will use Nvidia’s model, has been an early supporter of Nvidia’s open model initiative. In July, Nvidia formed an alliance promoting the use of AI models that countries can inspect and control. G42 was the only Gulf company to join this effort. Peng Xiao, G42’s CEO, stated at the time that open models are essential for the future of AI, designed with security and sovereignty in mind from the start. Sovereignty in AI, according to Soliman, is not an all-or-nothing choice. It involves factors like how models are fine-tuned for local data and the ability to serve them to users in the UAE and beyond. The U.S. government approved the sale of Nvidia’s advanced chips to the UAE after G42 sold its stakes in Chinese technology companies. The software running on these chips is mostly rented from U.S. companies, which control how the models are used. Falcon, developed by TII, was meant to be an exception, offering a free alternative to models from the U.S. and China. Meanwhile, Saudi Arabia, a competitor to the UAE in the AI space, has taken a different approach. The state AI company Humain has been authorized to purchase tens of thousands of Nvidia chips and recently introduced an Arabic AI model developed by a Chinese lab. A year earlier, it had launched a model it claimed was built entirely in Saudi Arabia. Sam Winter-Levy, a senior fellow at the Carnegie Endowment for International Peace, noted that the Saudis are using Chinese AI models to build their own systems while relying on U.S. hardware for development and deployment. Nvidia has its own motivations for promoting the new Nemotron 4 model. Earlier versions of U.S. AI models have struggled to compete with free models from China, which were downloaded around 2 billion times this year, according to Hugging Face, a platform for sharing AI models. To address this, Nvidia has agreed to pay nearly $13 billion to acquire Hugging Face, aiming to position Nemotron as the preferred base for governments building their own AI models. Einstein noted that the Poolside deal is more about increasing future adoption of Nemotron than displacing existing models. TII believes it can pursue both strategies: building its own AI foundation while using the best available technology from around the world. “Sovereignty does not mean isolation,” said Hacid. “We build our foundation at home, while remaining open to the best of the world.”