Nvidia commands over 75% of the global AI accelerator market as of mid-2026. While China’s labs have been shipping increasingly impressive models, from DeepSeek to Moonshot AI to z.ai, the silicon those models train on remains overwhelmingly American. US entities control roughly 75% of global AI-ready GPU cluster performance.
The trillion-dollar backlog
During GTC 2026, Nvidia CEO Jensen Huang announced an estimated $1 trillion in AI chip demand projected through 2027. That’s not revenue. That’s backlog, meaning customers lining up with purchase orders that Nvidia hasn’t even fulfilled yet.
The demand is being fueled by a spending arms race among America’s largest tech companies. Microsoft, Meta, Amazon, and Google have collectively committed hundreds of billions of dollars to expanding domestic data center capacity. Private AI investment in the US hit $285.9 billion by 2025, a figure that exceeds China’s AI investment by more than 23 times.
US export restrictions on advanced Nvidia chips to China, which began in October 2022 and were adjusted with revenue-sharing arrangements in 2025, have further cemented this asymmetry.
