As the race for artificial intelligence (AI) technology intensifies, the United States is at a critical juncture to maximize its international competitiveness in this domain.
To begin with, Washington can strengthen its domestic chip fabrication capacity and friend-shore supply chains, ensuring that the critical components needed for AI development are produced through reliable and trusted mechanisms.
One aspect of this strategy may involve tightening export controls on semiconductors and other essential technologies, particularly those with dual-use capabilities.
However, a cautionary note is warranted when considering the imposition of new tariffs on dual-use AI components from allied nations.
Implementing such tariffs risks counterproductive results, potentially raising hardware costs for U.S. AI firms and ultimately benefiting Chinese competitors.
The Trump administration’s investigation under Section 232 of the Trade Expansion Act into semiconductor imports has set the stage for potential tariffs, but the consequences of such actions could be far-reaching.
A blanket tariff on semiconductor imports may raise hardware costs for U.S. AI firms, penalize crucial allies like Mexico and Taiwan, and inadvertently lower prices for Chinese rivals.
While tariffs on Chinese imports have their place in the ongoing U.S.-China tech competition, applying them indiscriminately to U.S. partners could backfire significantly.
The competition in AI has emerged between the U.S. and like-minded allies against China, which has been labeled by entrepreneurs and national security experts as possibly “the most powerful and strategic technology in history.”
Access to AI-related imports—key components such as graphics processing units (GPUs), central processing units (CPUs), and printed circuit assemblies (PCAs)—remains crucial for U.S. firms to maintain their competitive edge.
In the face of the AI technology landscape’s rapid evolution, allowing China to set the rules of engagement would be indisputably detrimental.
With the demand for AI continuously growing, understanding the role that these components play is paramount.
The costs associated with training AI models are significantly impacted by hardware prices.
The surge in PCA unit imports observed recently indicates a rising demand for the hardware needed to support ambitious AI projects, and the imposition of tariffs could hamper the U.S. AI buildout.
Data shows that PCA imports have more than quintupled since 2021, although no corresponding productivity increases can account for this spike—pointing toward expanding hardware needs among U.S. firms.
If PCA prices increase due to tariffs, the result could stifle U.S. AI innovation and adoption efforts.
The interconnectedness of the tech sector with vital international partners underscores the risks associated with indiscriminate tariffs.
Domestically, Taiwan emerges as a key supplier, shipping leading-edge GPUs to the United States and committing substantial investments to bolster the U.S. tech landscape.
Mexico also plays a pivotal role, serving as the largest supplier of GPUs and CPUs to the United States.
By imposing tariffs on semiconductor imports, the U.S. may inadvertently stunt its own firms’ growth while limiting access to the global market.
Accelerator chips and other critical server components account for around half of all costs tied to training machine learning models.
As hyperscalers—firms with vast data centers and cloud services—focus on AI development, their projected capital expenditures are projected to surge, stressing the importance of affordable hardware.
Therefore, rising hardware acquisition costs due to tariffs could pose a deterrent to the emergence of new entrants in the AI field, dulling the competitive spirit fostered in the tech ecosystem.
Interestingly, imposing heavy tariffs may not only hurt the U.S. tech sector but also fail to stifle China’s advancements in AI.
Such tariffs could, in fact, drive other nations, particularly in Southeast Asia, closer to China due to revised trade dynamics.
After the announcement of proposed global tariffs by President Trump, Chinese President Xi Jinping visited key Southeast Asian nations, suggesting a strategy to uphold multilateral trade cooperation.
These discussions included AI collaboration, which raises significant alarms regarding tech transfer to China amid U.S. export control measures.
Recent reports indicate that Chinese engineers are increasingly utilizing Malaysian data centers to train their AI models, highlighting potential loopholes in U.S. export controls.
Furthermore, Taiwan’s semiconductor exports to ASEAN nations have reached new highs—surpassing exports to the United States—potentially reflecting a shift in allegiance towards Beijing.
This increase does not solely indicate transshipment activities; however, the growing presence of Chinese firms within the Malaysian data center scene raises concerns about U.S. export controls compliance.
This scenario presents a compelling argument against tariffs, as they might drive countries to seek collaboration with China, thus exacerbating the risk of U.S. technology leakage.
Given these considerations, it becomes critical for U.S. policymakers to explore alternatives to tariffs.
The United States has effectively challenged China’s tech ambitions through bipartisan export controls that limit access to advanced semiconductors.
Although the Bureau of Industry and Security has recently rescinded the AI Diffusion Rule—a measure designed to tighten chip-related exports—many experts agree that export controls remain a crucial instrument for economic statecraft and safeguarding national security.
These controls, while not comprehensive solutions, serve as necessary tools for maintaining a competitive edge over China.
Nevertheless, strengthening domestic capabilities, including incentivizing semiconductor manufacturing and expertise, is equally vital in this strategy.
The bipartisan CHIPS and Science Act is central to this objective, aiming to bolster U.S. manufacturing initiatives and facilitate domestic growth.
While there may be opportunities for adjustments to this framework, the overarching goals should remain intact as the U.S. seeks to amplify its competitiveness in AI development.
In this pivotal moment within the global race toward artificial general intelligence (AGI), the importance of proactive policymaking cannot be overstated.
Whether the journey toward AGI resembles a sprint or a marathon, its potential for economic productivity remains substantial.
At the same time, navigating the associated risks necessitates that the United States takes the lead in shaping standards and regulatory frameworks.
Thus, securing access to critical, high-end semiconductors and technologies is paramount for the U.S. to retain its competitive edge over China in the race for AI advancements.
image source from：https://www.atlanticcouncil.org/blogs/new-atlanticist/why-tariffs-on-ai-hardware-could-undermine-us-competitiveness/
