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

Attendees during a media event at the Google Midlothian Data Center in Midlothian, Texas.
Bloomberg/Getty Images

A version of this story originally appeared in the BI Tech Memo newsletter.
Sign up for the weekly BI Tech Memo newsletter here.

There’s an emerging new way to measure economies in the AI era: Gross Domestic Intelligence, or GDI.

“We believe investors may begin to assess ‘Gross Domestic Intelligence’ resources at the national level, and this may well become an important investment overlay when assessing the competitiveness of entire nations and industries.” Morgan Stanley analysts wrote in a note to investors recently.

GDI measures who has the most AI resources, especially access to AI compute infrastructure including GPUs, networking gear, and data centers.

Epoch AI, a research group that tracks AI and its impact, took a stab at this last year by estimating the share of AI computational power globally. The US dominated, with roughly 75%. China was second with about 10%, followed by the EU, then Norway and Japan.

This group, backed by Dustin Moskovitz’s Coefficient Giving charity, came out with similar estimates this week. The new data, from Q4 2025, measures the AI compute capacity of various companies and countries. It’s based on AI chips that include, and are equivalent to, Nvidia’s H100 GPUs. Compute capacity is how many operations each AI chip can perform at its peak.

Google dominates in a mind-blowing way, with its own TPUs and a lot of Nvidia GPUs. Every top company is American. China only has as much AI compute as Oracle, according to Epoch AI estimates.