The geopolitics of silicon just got messier. China has added NVIDIA's RTX 5090D V2—a chip specifically designed to comply with US export restrictions—to its banned imports list. The timing makes it particularly pointed: the ban was announced while NVIDIA CEO Jensen Huang was visiting the country.
This is calculated geopolitical maneuvering masquerading as trade policy. NVIDIA created the RTX 5090D V2 specifically to thread the needle of US-China technology restrictions. The United States has imposed export controls limiting advanced chip sales to China, particularly those useful for AI development. NVIDIA responded by developing variants with reduced capabilities that technically comply with those restrictions while still serving the Chinese market.
China just closed that loophole. By banning the RTX 5090D V2, Chinese authorities are signaling they won't accept deliberately neutered technology. It's a rejection of the entire premise: that China should be grateful for permission to buy limited versions of chips they can't access in full form.
The timing of the ban—added to customs checkpoint lists last Friday while Jensen Huang was in-country—isn't coincidental. According to Ars Technica, this kind of announcement during a high-profile CEO visit sends a clear message: China views these semiconductor restrictions as an ongoing negotiation, not settled policy they'll passively accept.
This reveals how quickly the semiconductor supply chain is fragmenting into competing spheres. A decade ago, the chip industry operated on genuinely global supply chains. Components, manufacturing, and sales crossed borders with relatively few restrictions. That era is over.
What we're seeing now is technology becoming inseparable from geopolitics. NVIDIA isn't just selling graphics processors—it's navigating competing demands from two superpowers treating semiconductors as strategic assets. The US wants to limit China's access to advanced chips that could accelerate AI and military capabilities. China wants technological self-sufficiency and resents restrictions it views as attempts to constrain its development.
NVIDIA is caught in the middle. The company generates substantial revenue from the Chinese market. It also must comply with US export controls or face severe penalties. Creating compliance variants like the RTX 5090D V2 was NVIDIA's attempt to maintain market access while following US law. China's ban renders that strategy moot.
The broader implication is that semiconductor supply chains are now explicitly splitting into US-aligned and China-aligned ecosystems. Companies will increasingly need to choose markets or develop completely separate product lines for different geopolitical spheres. The efficiency and innovation advantages of unified global supply chains are being sacrificed to strategic competition.
For NVIDIA, this means lost revenue and increased complexity. For the tech industry broadly, it signals an acceleration of bifurcation that will increase costs, reduce efficiencies, and slow innovation. When companies can't access global markets or must develop multiple product variants for different political blocs, everyone loses some benefit of scale.
The US-China technology rivalry isn't resolving. It's intensifying. And companies like NVIDIA that operate at the intersection of cutting-edge technology and great power competition will face increasingly difficult tradeoffs.
Jensen Huang's visit to China was likely meant to maintain relationships and signal NVIDIA's commitment to the Chinese market despite restrictions. The ban announcement during his visit suggests Chinese authorities aren't interested in symbolic gestures. They want access to unrestricted technology or they'll develop their own alternatives.
The chip wars aren't over. They're just getting started.
