• China plans to implement its own approval process for Nvidia H200 chip purchases, requiring buyers to justify why domestic alternatives cannot meet their needs.
  • Former President Trump announced Nvidia can export H200 chips to "approved customers" with a 25% U.S. government surcharge, reversing Biden-era restrictions.
  • The move highlights deepening U.S.-China semiconductor tensions, with Beijing pushing technological self-sufficiency despite potential access to superior foreign hardware.

Nvidia gains theoretical access to China's approximately $50 billion market for high-end GPUs under Trump's new policy, announced Monday, but faces practical hurdles as Chinese regulators prepare countermeasures. According to people familiar with the matter, Beijing is considering a restricted approval process that would force companies to demonstrate domestic chips cannot fulfill their requirements before purchasing H200s—a chip approximately six times more powerful than the previously allowed H20 model.

"What institutional investors like us are really focused on is regulatory stability," one industry executive noted, though speaking about different markets. In this case, instability defines the landscape: the U.S. opens a door China seems determined to close partially. Trump claimed Chinese President Xi Jinping responded "positively" to the export announcement, yet Beijing's planned restrictions signal continued emphasis on reducing dependence on American semiconductors. This creates a contradiction—one government permits while the other potentially limits—reflecting deep-seated concerns about foreign technology and national security.

Market reaction has been muted, with Nvidia shares showing only modest response to Trump's announcement, suggesting investor caution about actual sales opportunities. The semiconductor market dynamics reveal a complex competitive landscape where Chinese companies like DeepSeek, Alibaba, and Huawei have made substantial progress in AI development and hardware manufacturing despite previous U.S. export restrictions. China previously rejected H20 exports as a security concern, instructing domestic companies to cease purchases—a decision that inadvertently benefited Chinese competitors by reducing direct competition.

Efforts to navigate these crosscurrents have hit a snag. Without a more aligned approach, Nvidia might see limited penetration in China despite the policy shift. The White House views its move as a middle-ground strategy—avoiding complete restrictions while ensuring U.S. technology standards remain globally dominant and preventing Chinese competitors from capturing too much market share. Meanwhile, Beijing actively encourages domestic technology adoption over reliance on U.S. exports, with regulators now crafting mechanisms to control H200 access.

Industry-specific elements like filing deadlines and approval processes will determine how this unfolds. Chinese AI companies and data centers face competing pressures: the H200 offers superior performance, but domestic patriotism and regulatory barriers make acquisition difficult. "We have a constant balance with the banks, which really we consider our partners and not only our binary competitors," another executive said in a different context, highlighting how partnerships can blur lines. Here, the tension is more stark—between foreign innovation and local substitution.

Human touches emerge in statements from those affected. DeepSeek's CEO previously stated that "bans on shipments of advanced chips are the problem," not financing, while Tencent executives highlighted that computing chip resources remain their most critical constraint. These voices underscore the practical challenges Chinese tech firms face, caught between performance needs and policy directives. Attempts to reach Nvidia and Chinese regulatory officials for comment were unsuccessful by publication time.

The situation stems from years of escalating U.S.-China semiconductor competition. The Biden administration implemented strict export controls on advanced AI chips, aiming to prevent China from developing competitive AI capabilities. However, these restrictions demonstrated limited effectiveness—Chinese firms have produced world-class AI models, and companies accelerated hardware development to fill voids. The previous compromise of allowing H20 exports failed when China rejected them, prompting the White House to reconsider its approach.

Looking ahead, short-term uncertainty surrounds actual H200 sales volumes, as regulatory approvals remain selective and Chinese companies increasingly pursue domestic alternatives. The semiconductor market continues bifurcating into U.S.-aligned and China-aligned supply chains. Some analysts within the White House believe export controls provide only temporary advantages while China pursues technological self-sufficiency regardless of U.S. policy. Yet, experts argue restricting advanced chips is critical to maintaining competitive advantage in AI development, particularly regarding computing resources for large-scale clusters.

Natural transitions between topics reveal a broader context: concerns about U.S. domestic chip manufacturing capacity and over-reliance on Taiwan Semiconductor Manufacturing Company (TSMC), alongside China's strategic control of rare earth minerals essential for batteries and critical technologies. These interdependencies create leverage points for both nations in ongoing technological competition. The fundamental tension remains—the U.S. seeks to maintain leadership and market access, while China seeks independence. China's planned restrictions suggest that even with U.S. approval, Beijing will carefully control access to prevent excessive dependence on foreign semiconductors.

Correction: An earlier version misstated the power comparison; the H200 is approximately six times more powerful than the H20, not five times.