Central News Agency (Reporter Chang Hsin-yu, San Francisco, April 16) - Stanford University recently released its annual AI Index Report, highlighting Taiwan's increased importance as a key factor in AI hardware resilience. The core conclusion points out that the United States has 5,427 data centers, far surpassing other countries, while TSMC controls the lifeline of the global AI hardware supply chain.
This is the ninth time Stanford University has released its Artificial Intelligence Index (AI Index) Report. Compared to previous years, which mostly included Taiwan in statistical lists, this year emphasizes Taiwan's economic data and critical position in the supply chain.
In the 423-page report, Stanford University lists 10 core conclusions on its website; TSMC's critical position in the global AI hardware supply chain is one of them. The text elaborates that since 2022, AI model computing power has grown at a rate of approximately 3.3 times per year, and almost all computing demand relies on one Taiwanese wafer foundry, which also highlights the fragility of the hardware supply chain.
In 2025, the United States will have 5,427 data centers, more than 10 times the number of any other country. TSMC is critical to the AI chip supply chain. Nvidia and AMD chips are all manufactured by TSMC, and each layer of the supply chain has extremely high technical barriers, requiring decades of experience accumulation, dedicated equipment, and huge capital investment.
In addition, Taiwan's AI adoption rate in the second half of 2025 is 28.4%; in industrial automation, the report compares 2023 and 2024, pointing out that Taiwan has the highest growth in industrial robot installations globally, with an annual growth rate of 33%.
In recent years, the performance of AI models in the United States and China has been a key focus of the report. The core conclusion points out that the US lead has narrowed, and since early 2025, models from the two countries have repeatedly surpassed each other. In February, China's DeepSeek-R1 once tied with top US models; as of March, Anthropic's strongest model in the US led by a margin of 2.7%.
Furthermore, the total private AI investment in the United States in 2025 reached US$285.9 billion (approximately NT$9.2 trillion), more than 23 times China's US$12.4 billion. However, the US$12.4 billion does not include Chinese government funding, so it does not represent China's overall AI expenditure.
Notably, the number of AI researchers and developers migrating to the United States has decreased by 89% since 2017, and by 80% in the past year alone.
Regarding the impact of AI on jobs, there is a significant gap between AI experts and the general public: 73% of experts expect AI to have a positive impact, while only 23% of the general public share the same view, a difference of 50 percentage points. (Editor: Chen Hui-ping) 1150417
Choose to stand with facts, every sponsorship you make is a force to protect press freedom.
Download the Central News Agency's "First-hand News" APP to stay updated with the latest news.
The text, images, and videos on this website may not be reproduced, publicly broadcast, or publicly transmitted and used without authorization.
FACT BOX
- Source: CNA (Central News Agency)
- Category: research