The Global Computing Power Game: NVIDIA’s Profit Structure and China’s Domestic Substitution Under the ‘Chip Tax’

Abstract

The intensifying geopolitical frictions between the United States and China, exemplified by the recent implementation of the ‘Chip Tax’ and advanced semiconductor export controls, are fundamentally transforming the global artificial intelligence computing power landscape. This paper examines the dual impacts of these measures through an analysis of their influence on NVIDIA’s profit structure and the simultaneous acceleration of China’s domestic substitution strategy. By adopting a mixed-methods approach that integrates quantitative financial analysis of NVIDIA’s filings with the U.S. Securities and Exchange Commission (SEC) and qualitative policy evaluation, this research reveals that U.S. restrictions have imposed substantial yet temporary financial burdens on NVIDIA. These include an inventory charge of $4.5 billion and a compression of the gross margin by 1,250 basis points in the first quarter of the fiscal year 2026. Concurrently, these policies have served as catalysts, expediting China’s development of a parallel semiconductor ecosystem. This is manifested by Cambricon’s 4,300% surge in revenue and Huawei’s Ascend 910B achieving performance close to that of H2O at 70% of the cost. The study concludes that while the ‘Chip Tax’ effectively exerts short-term pressure on established players, it inadvertently contributes to the formation of a bifurcated global technology landscape, resulting in a fragmented ‘one world, two systems’ scenario that may ultimately enhance China’s determination for technological self-sufficiency.