Investment Rating - The report upgrades ADNOC Gas to Buy with a target price of 4 AED, reflecting a positive outlook on gas-to-power demand [1][5][12]. Core Insights - The GCC is positioned to capture a 5% market share in the global AI sector this decade, driven by low energy costs and favorable capital conditions [1][2]. - The region's AI capacity is expected to grow tenfold to 6 GW by 2030, contributing to 20% of ex-US growth during this period [2][9]. - ADNOC Gas is identified as a key beneficiary of rising gas-to-power needs, with a projected domestic sales growth of 5% CAGR by 2030 [5][12]. Summary by Sections AI and Data Center Growth - The Middle East's current market share in AI is only 1%, but it is anticipated to grow at four times the broader market rate, reaching 5% by 2030 [9]. - Significant investment of approximately $100 billion is required for data center infrastructure and advanced GPUs to support this growth [10][19]. Energy Demand and Supply - The report forecasts a 0.5-2.0% uplift in power demand across the GCC this decade, with the UAE likely at the upper end [3][28]. - The region benefits from low-cost gas and renewable energy sources, with gas prices at $2.15/mmbtu compared to $3/mmbtu in the US and $7/mmbtu in Europe [25][27]. Geopolitical and Regulatory Environment - US-led investments in Middle East data centers and favorable regulations, such as Saudi Arabia's draft 'Global AI Hub' law, enhance the region's attractiveness for AI processing [4][42]. - The report highlights the need for robust regulatory frameworks to manage data protection and digital jurisdiction as the region positions itself as a global AI hub [41][43]. Cooling and Infrastructure Challenges - The shift towards hyperscale data centers necessitates advancements in cooling technologies, with liquid cooling emerging as a more efficient solution compared to traditional air cooling [38][40]. - The report notes that the Middle East's solar energy costs are among the lowest globally, which supports the region's energy needs for data centers [29][33].