Forget GDP. Meet GDI: The new economic scorecard for AI power

In the rapidly evolving landscape of artificial intelligence, a new economic metric is gaining traction: Gross Domestic Intelligence, or GDI. Analysts at Morgan Stanley suggest that this innovative measurement could reshape how investors assess national and industrial competitiveness in the AI sector. GDI focuses on the availability of AI resources, particularly the access to vital compute infrastructure such as GPUs, networking equipment, and data centers. A research initiative named Epoch AI has been at the forefront of quantifying this phenomenon, providing insights into global AI computational power distribution. According to their estimates, the United States holds a commanding lead with approximately 75% of the world’s AI compute capacity. Trailing behind, China accounts for around 10%, while the European Union, Norway, and Japan follow in the rankings. Recent data from Q4 2025 corroborates these findings, showcasing the AI compute capabilities of various nations and companies. The analysis focuses on AI chips, particularly those equivalent to Nvidia's H100 GPUs, which serve as a benchmark for measuring compute capacity. In this competitive arena, Google stands out significantly, leveraging its own Tensor Processing Units (TPUs) alongside a substantial inventory of Nvidia GPUs. Notably, all leading firms in this domain are American, with Epoch AI’s estimates indicating that China’s AI compute power is comparable to that of Oracle. As the world enters an era where the economic power of nations may increasingly hinge on their AI capabilities, GDI could become a critical factor in investment decisions and strategic planning across industries.