Ping Xiao, the CEO of G42, addressed the rapid implementation of artificial intelligence projects and the increasing global demand during sessions at the World Economic Forum in Davos. He highlighted that G42 is operating a specialized business unit focused on building what they term “token factories.” This initiative encompasses the development of fundamental infrastructure and the management of graphical processing units (GPUs), which are crucial for enabling large-scale production and operation of AI technologies, particularly amidst the ongoing expansion of global demand for generative AI.
During his remarks, Xiao mentioned that the company is currently progressing on a massive AI campus project, boasting a total capacity of 5 gigawatts, which was announced last May during a visit from former U.S. President Donald Trump. He firmly believes that, from a business perspective, the cost of AI will eventually align with energy costs, viewing this correlation as a strategic advantage for the UAE due to its robust national energy production capabilities.
He explained that providing energy at this scale, directly linked to the grid and without complications related to permits or regulatory barriers, has been a key factor in the group’s deep involvement in AI infrastructure. Additionally, this supportive environment has allowed the group to establish strategic partnerships with leading global tech firms such as OpenAI and Microsoft, aimed at delivering extensive computing capabilities.
Xiao noted that construction is advancing rapidly, with over 7,000 workers on site and supported by more than 100 cranes. They are witnessing the achievement of approximately 250 megawatts of capacity every quarter, within a phased plan to reach the ultimate goal of 5 gigawatts.
He expressed strong optimism about the project’s future, anticipating that a significant portion of this capacity will be utilized domestically, while the excess capabilities will allow the UAE to export “tokens,” or what he described as AI packages, to global markets. He added that despite G42 operating from a relatively small nation in terms of population, the company’s focus remains on leveraging AI to serve the broader community.
As part of its key performance objectives for this year, G42 aims to produce over a billion AI agents to bolster GDP growth. Xiao stated that these agents will encompass a wide array of applications, including software development, petroleum engineering, and cybersecurity. He estimated that operating one billion AI agents, even if each worked just 12 hours a day, would consume around 1 gigawatt of AI infrastructure capacity.
In conclusion, Xiao emphasized that this level of ambition underscores the significant investments currently being funneled into AI infrastructure in the UAE, describing large-scale AI technology production as a crucial lesson for governments worldwide, especially given the accelerating global demand for these technologies and their transformative impact on economic models.