US policy pivot on chip sales in China. What does it mean for global tech?

Last month, US chip giant Nvidia said the US government would allow the sale of certain artificial intelligence processors in China, reversing an export ban imposed in April. Major US firms also agreed to give the government a 15% cut of revenue earned from Chinese sales, a concession that was reportedly part of a deal to secure export licenses.

Semiconductors have long been a core pillar of both US and Chinese industrial strategy given their use in everything from smartphones and electric vehicles to energy grids and AI systems. Moreover, experts and officials on both sides maintain that control over advanced chip design and production must be treated as a national security priority.

The sale of certain US chips in China could have a major impact on the global technology industry and influence the geopolitical jockeying for chip supremacy. So to better understand the latest developments, we spoke with Srabanti Chowdhury, a Professor of Electrical Engineering and a Senior Fellow at the Precourt Institute at Stanford University in the United States.

Here’s what she had to say.

How would a rollback of US export curbs affect the global value chain and capital allocation?

If the US rolls back some export restrictions, companies here might see a short-term boost in business—particularly from China—for things like mature-node tools and services. But I don’t think that changes the bigger picture: the most advanced manufacturing will still stay in places like Taiwan, Korea and Japan because the real bottlenecks (like extreme ultraviolet lithography and advanced packaging) are still off limits.

Even with relaxed rules, most companies have learned that policies can shift quickly, so they’ll likely keep spreading out risk—across geographies, customers and supply chains. And from China’s side, I’d expect even more investment in building up its own ecosystem, just in case the restrictions come back. So in the long run, capital won’t just chase new openings—it’ll still be guided by the need to stay resilient.

What should policymakers keep front-of-mind with regards to security vs. commercial viability?

Policymakers need to keep both sides in view—security and commercial reality. If controls are too broad or constantly shifting, we risk hurting US companies more than helping national interests.

These firms depend on scale, global customers and predictable roadmaps to keep investing in next-gen tech. At the same time, I fully understand the need to block sensitive capabilities from falling into the wrong hands. So, what’s needed is a smarter middle ground—clear, narrowly focused restrictions that address real risks, not blanket bans. And just as importantly, a strong, stable foundation here at home—investment in R&D, manufacturing, packaging and workforce—so companies have the confidence to build the future here, even in a more fragmented world.

Could restrictions hurt US companies more than they slow China?

Yes, definitely—it may already be happening in some ways. When US companies lose access to a high-volume, high-growth customer base, especially for mature-node tools or services, it cuts into their revenue and scale, which makes it harder to reinvest in future R&D.

Meanwhile, China would double down into building their own supply chain and reducing reliance. So blanket restrictions can backfire by pushing them to become more self-sufficient, faster. We must be careful not to weaken our own industry in the process. Smart, focused controls plus strong domestic support—that’s the path that protects both security and competitiveness.

If the US allows certain chip exports, how does that affect China’s long-term plans?

If the US allows certain chip exports, China will most likely treat it as a temporary opening—not a long-term shift—and continue pushing hard on self-reliance. The last few years have shown them how fragile access can be.

So even if exports resume in some areas, I don’t think it slows down their investment in domestic fabs, foundries, packaging and electronic design automation tools. If anything, it buys them a bit more breathing room to scale up while still working toward tech independence. In the long run, they’re unlikely to count on stable access to US technology—and that mindset is already baked into their national strategy.

How might heightened US–China competition shape innovation and standards?

This kind of US–China tech race will probably speed up innovation in some ways—especially in areas like AI and quantum—but it’ll also split the global ecosystem. We’re already seeing signs of that, with different countries leaning into their own standards, supply chains and software stacks. And it’s not just AI and quantum—things like sensing systems, radio frequency, edge computing, even thermal architectures could all get shaped by this divergence.

If each side is building for its own ecosystem, interoperability takes a hit, and collaboration becomes harder. So while the pace of invention might rise, we could lose out on scale, efficiency and shared progress. For the US, leading in innovation isn’t just about who builds faster—it’s also about who sets the standards and policies. So, staying open to alliances, standard-setting bodies and partnering—where possible—should be part of the long game.

Is this a zero-sum game or can it be mutually beneficial?

I hope it’s not a zero-sum game—but it could turn into one if we’re not careful. There’s real potential for mutual gain here: competition can spark innovation, grow markets and raise the bar for everyone. But if we go all-in on decoupling and control, we lose out on scale, efficiency and the kind of collaboration that drives long-term breakthroughs.

That’s not just bad for the US—it slows global progress. The smart move is to be targeted: safeguard what’s sensitive, for sure, but still stay open to coordination where it makes sense. The semiconductor industry was built to be interdependent. Building resilience should be achieved through healthy competition and collaboration but not isolation.