Maximizing international competitiveness in artificial intelligence (AI) is a top priority for the United States, but the path forward is anything but simple. As policymakers debate how best to secure America’s technological edge, one proposal—imposing broad tariffs on AI hardware imports—has sparked intense discussion. Let’s unpack why this approach could backfire, and what smarter strategies might look like.
Imagine a thriving US AI sector, fueled by cutting-edge research and robust investment. Now, picture that same sector suddenly facing higher costs for the very hardware it needs to innovate—GPUs, CPUs, and printed circuit assemblies (PCAs). This is the risk posed by sweeping tariffs on semiconductor imports, even from friendly partners like Taiwan and Mexico. While the intention is to protect national security and encourage domestic manufacturing, the reality is that such tariffs could slow AI progress at home and give rivals like China an unexpected boost.
The Hardware Behind the AI Boom
AI breakthroughs don’t happen in a vacuum—they rely on powerful hardware. GPUs and CPUs are the engines that train massive AI models, while PCAs connect these components in servers and data centers. Demand for these parts has surged as AI adoption accelerates, with imports of PCAs alone more than quintupled since 2021. For US firms, access to affordable, high-quality hardware is essential to stay ahead in the global AI race.
But here’s the catch: hardware is expensive. Research shows that chips and server components account for about half the cost of training advanced AI models. If tariffs drive up prices, it’s not just tech giants who feel the pinch—startups and new entrants could be priced out, stifling innovation and competition.
Global Partnerships Matter
The US doesn’t operate in isolation. Taiwan and Mexico are key suppliers of dual-use technology—hardware with both civilian and military applications. Taiwan’s semiconductor industry, for example, is investing billions in the US, while Mexico is the largest supplier of GPUs, CPUs, and PCAs by value. Tariffs on these imports would not only strain relationships with vital partners but also limit US firms’ access to the global market.
Unintended Consequences: Helping the Competition
Ironically, tariffs meant to strengthen US security could end up helping China. By making it harder for US companies to source affordable hardware, tariffs could push other countries to deepen ties with Beijing. There’s already evidence of technology transfer and transshipment through Southeast Asia, with Chinese firms leveraging regional data centers and supply chains. If US policy drives partners away, it risks undermining export controls and making Chinese competitors more formidable.
Smarter Strategies for US Leadership
So, what’s the alternative? Rather than blanket tariffs, targeted export controls have proven effective in limiting China’s access to high-end semiconductors. At the same time, the US can double down on incentives for domestic chip manufacturing—think the CHIPS and Science Act—and foster international collaboration with trusted allies. These approaches support both security and innovation, without the collateral damage of broad tariffs.
Actionable Takeaways for Policymakers and Industry
- Prioritize targeted export controls over broad tariffs to avoid harming US firms and partners.
- Invest in domestic semiconductor manufacturing and R&D to reduce reliance on imports.
- Strengthen alliances with key suppliers like Taiwan and Mexico through trade and investment.
- Monitor global supply chains for signs of technology leakage and adapt policies accordingly.
- Support startups and new entrants to keep the US AI ecosystem dynamic and competitive.
Summary of Key Points
- Broad tariffs on AI hardware could raise costs for US firms and slow innovation.
- Key partners like Taiwan and Mexico are essential to the US AI supply chain.
- Tariffs may inadvertently benefit China by encouraging tech transfer and regional alliances.
- Targeted export controls and domestic investment are more effective tools.
- Maintaining US leadership in AI requires smart, collaborative policy—not isolationism.