The latest chapter in the US-China technology conflict features an unexpected plot twist: after years of tightening export restrictions on advanced semiconductors, the Trump administration loosened controls on AI chips bound for China on January 13—and within 24 hours, Beijing responded by instructing customs agents to ban Nvidia's H200 chips from entering the country.

The moves highlight the increasingly complex dynamics of technological decoupling, where both sides are simultaneously pursuing access to advanced technology while protecting their domestic industries.

The US Policy Shift

On January 13, 2026, the Department of Commerce published a new regulation that marked a significant departure from the trajectory of export controls:

  • License review change: Advanced AI chips moved from "presumption of denial" to "case-by-case review"
  • Affected products: Nvidia H200, AMD MI325X, and equivalent chips
  • Effective date: January 15, 2026
  • Requirements: Exporters must make specific technical, business, and end-user certifications

The policy change codified an announcement President Trump made on December 8, 2025, signaling a more flexible approach to chip exports that had been expected to arrive only after his inauguration.

"The regulation acknowledges that exporting advanced AI chips to China poses serious national security risks, while simultaneously creating a pathway to permit their sale. The result has been described by critics as strategically incoherent."

— Council on Foreign Relations analysis of the export policy

Beijing's Response

China's reaction came swiftly. On January 14, Reuters reported that Chinese authorities had instructed customs agents to ban Nvidia's H200 chips from entering the country. Additionally:

  • Government officials told domestic tech firms not to purchase H200 chips unless necessary
  • Chinese regulators signaled increased scrutiny of foreign technology acquisitions
  • The Ministry of Commerce announced reviews of foreign AI-related investments

Nvidia shares fell more than 1% on the news, though the decline was modest compared to the potential revenue implications.

The Tariff Layer

Adding complexity to the situation, the White House announced an immediate 25% tariff on semiconductors with advanced performance thresholds on January 14. This action, taken under Section 232 of the Trade Expansion Act, determined that imported semiconductors threatened to impair US national security.

The tariff creates a paradoxical situation:

  • The US is loosening export controls on AI chips to China
  • While simultaneously imposing tariffs on chip imports
  • And China is restricting imports of the very chips the US is now willing to sell

Why Both Sides Are Playing Defense

The seemingly contradictory moves from both Washington and Beijing reflect underlying strategic concerns:

US Motivations

  • Revenue considerations: American chip companies have lobbied heavily against export restrictions that cost them billions in sales
  • Controlled access: Some officials believe that selling chips under strict conditions is better than driving China to develop indigenous alternatives
  • Complexity reduction: The previous regime of restrictions had become unwieldy to enforce

China's Calculations

  • Self-sufficiency push: Beijing wants to accelerate domestic chip development, which benefits from reduced access to foreign alternatives
  • Leverage: Restricting imports gives China bargaining power in future negotiations
  • Face-saving: Being seen as choosing to restrict imports is preferable to being denied access

Impact on Nvidia

For Nvidia, the world's leading AI chip designer, the policy whiplash creates both opportunities and challenges:

  • Potential upside: Loosened US export controls theoretically open a $50+ billion market
  • Reality check: China's import restrictions may negate much of that potential
  • Stock reaction: Shares have been volatile as investors parse the implications
  • CEO diplomacy: Jensen Huang recently traveled to China to explore opportunities

The company continues to develop China-specific chips that comply with export restrictions, hedging against policy uncertainty.

The Entity List Expands

Even as it loosened some restrictions, the Bureau of Industry and Security added 80 entities to the Entity List from China, UAE, South Africa, Iran, Taiwan, and other countries. These additions reflect continued efforts to restrict access to advanced technology by parties deemed to pose national security concerns.

Industry Reactions

The semiconductor industry has responded with a mix of relief and confusion:

  • Relief: Some sales that were previously blocked may now proceed
  • Confusion: The policy environment remains unpredictable
  • Planning challenges: Long-term investment decisions are complicated by shifting rules

What Comes Next

The current situation is unlikely to be stable. Several developments could reshape the landscape:

  • Further policy changes from either government
  • Escalation of the tariff conflict
  • Progress in China's domestic chip development efforts
  • Congressional action on semiconductor policy

The Bigger Picture

The back-and-forth on AI chip exports illustrates a broader truth about US-China technology relations: complete decoupling is neither feasible nor desired by either side, but complete integration is no longer politically possible. The result is a messy middle ground where both countries pursue their interests through a combination of restrictions, exceptions, and strategic ambiguity.

For investors in semiconductor companies, the takeaway is clear: policy risk is now a permanent feature of the landscape, and companies with diversified geographic exposure may be better positioned than those heavily dependent on any single market.

The technology cold war continues—and the rules of engagement are still being written.