Trump opens the door for Nvidia to sell H200 chips to China with a 25% tariff

Last update: 16/12/2025

  • Trump authorizes Nvidia to export H200 AI chips to Chinese and other customers under strict security controls.
  • The United States reserves 25% of the revenue from these sales and plans to extend the model to AMD, Intel, and other manufacturers.
  • China will have to approve and filter buyers, while accelerating the development of its own chips to reduce its dependence.
  • The move boosts Nvidia's stock price, but creates political division in Washington and maintains geopolitical pressure on the technology sector.
Trump sales of Chinese Nvidia chips

US President Donald Trump's decision to Partially open exports of Nvidia's H200 chips to China It has abruptly reshaped the landscape of artificial intelligence technology. The White House has opted for a middle ground: allow sales, but in exchange for a high tax toll, comprehensive security filter and a regulatory framework which makes it clear that the priority remains the strategic advantage of the United States.

This move, communicated directly to Xi Jinping and disseminated through Truth Social, combines economic interests, geopolitical rivalry, and electoral calculationsNvidia, AMD, and Intel will once again have access to one of their largest markets, but under close supervision and with the It remains to be seen to what extent Beijing will allow its companies to buy these processors. after having promoted a policy of technological substitution towards national suppliers.

Conditional authorization: 25% toll and security screening

Nvidia H200

Trump has announced that Nvidia will be able to sell its H200 chip to approved customers in China and other countriesprovided they pass strict national security checks. The transaction will not be a simple commercial exchange: each buyer must be vetted by US authorities, who will review the potential military, strategic, or sensitive use of these high-performance processors.

In his message, the president explained that The United States will keep 25% of the revenue generated by these salesThis is well above the 15% that Nvidia had previously agreed to with Washington for the export of the H20O model. The White House is considering extending this "license plus commission" scheme to other manufacturers such as AMD and Intelso that any access to advanced AI chips by China will inevitably have to go through the US regulatory filter.

Spokespeople like Karoline LeavittThe White House press secretary emphasized that the licenses will not be automatic and that only companies that meet a certain standard will have access. comprehensive evaluation processThe stated objective is to minimize any risk of diversion towards military programs, offensive cybersecurity, or mass surveillance systems contrary to Washington's interests.

A partial relief from the veto: the role of the H200 chip

The heart of the measure focuses on the H200, one of the most powerful AI chips in Nvidia's Hopper familyThis processor, intended for data centers and the training of large-scale artificial intelligence models, had been subject to severe export restrictions under the Biden administration and in the early stages of the current term.

To overcome the previous limitations, Nvidia went so far as to design scaled-down versions such as the H800 and H20adapted to the limits set by Washington. However, China responded coldly: the authorities recommended that its companies They will not use these degraded productsThis stance was interpreted by many analysts as a pressure tactic to gain access to more powerful hardware like the H200 itself.

Exclusive content - Click Here  How to overclock from an unlocked processor in WinZip?

The new authorization represents a change of course: Washington will allow the sale of H200, but is keeping the Blackwell and Rubin families completely out of the agreement.The next generation of Nvidia chips is designed for even more demanding AI applications. Trump has explicitly emphasized this, making it clear that these next-generation processors will remain reserved for the United States and its allies, and will not be part of the shipments to China.

Nvidia, between business and geopolitics

Nvidia revenue

For Nvidia, the decision opens a window of opportunity in one of its key markets for high-performance chipsChina accounts for a very significant portion of global demand for processors for data centers and artificial intelligence projects, so recovering some of that flow could translate into billions of additional dollars per quarter.

The company's chief financial officer, Colette KressHe even estimated that chip sales to the Chinese market could add between $2.000 billion and $5.000 billion in quarterly revenue if the restrictions were lifted. Other analysts, such as Gene Munster, estimate that the partial reopening with H200 could push Nvidia's annual revenue growth to 65% year-over-year, compared to the 51% forecast before the regulatory change.

The company's CEO, Jensen HuangShe has been one of the most active voices in Washington calling for a relaxation of the veto. According to sources close to her, cited in the American press, Huang warned the government of the risk of giving up a market worth tens of billions of dollars to emerging Chinese competitors if a total lockdown was maintained. Their pressure would have been key to forging an intermediate solution: selling some, but under very controlled conditions.

Immediate reaction on the stock market and a ripple effect on the sector

Trump's announcement had an almost instantaneous impact on financial markets. Nvidia shares rose about 1,7% in pre-market trading. from the US market and closed the previous session with a gain of approximately 1,73%. So far this year, the stock has accumulated a rise of around 28%-40% depending on the benchmark index used, well above the average performance of the S&P 500.

The movement also dragged down the rest of the semiconductor sector. AMD gained around 1,1%-1,5% in early tradingWhile Intel advanced approximately between 0,5% and 0,8%., pending further details on whether they will receive similar licenses to export their own artificial intelligence chips under the same conditions.

Analysts from firms like Morningstar believe that, despite the regulatory volatility of recent years, The new policy opens at least one clear path to significant AI revenues from ChinaHowever, they warn that the continuity of this framework is not guaranteed: Washington has gone back and forth with the restrictions and could tighten them again if the political or security situation changes.

China, between negotiation and technological autonomy

On the other side of the Pacific, the Chinese reaction has been calculatedly cold. Beijing's Ministry of Commerce has called the decision “a positive but insufficient step”insisting that US vetoes and controls remain in place distorting competitionThe H200 authorization also comes after the Asian country boosted new subsidies for its semiconductor industry with the aim of Double the national capacity for high-end chips by 2026.

Exclusive content - Click Here  How to return a product to Apple?

Chinese regulators are now considering allowing access limited and highly regulated Regarding the H200 series, according to sources cited by international media, Chinese companies wishing to acquire these processors will have to undergo their own approval process and justify why local manufacturers cannot meet their needs with domestically produced chips. In other words, Beijing also intends to set the rules and reduce its exposure to unilateral decisions by Washington.

In parallel, US restrictions have accelerated the strategy of Chinese technological autonomyThe country has intensified investment in research, manufacturing capacity, and partnerships with suppliers not subject to the same level of control. In the medium term, this move could lead to a scenario of more fragmented technological mapwith standards and supply chains that run in parallel between rival blocs.

Political clash in Washington over sales to China

Chinese AI microchips and the US

The green light for Nvidia's sales has not been unanimously received on Capitol Hill. US lawmakers are deeply divided on whether it is a risky concession or a smart move to strengthen the country's leadership in AI and semiconductors.

Some members of Congress warn of the danger of putting One of the most valuable technological assets in the United States is in the hands of its main strategic competitor.Representative Andrew Garbarino, chairman of the House Homeland Security Committee, has expressed concern that these chips could end up strengthening capabilities in fields such as quantum computing or cyber espionage, areas in which a Chinese advance could have direct consequences for Western security.

Others, such as Congressman Brian Mast, chairman of the House Foreign Affairs Committee, argue that the measure fits within a broader strategy to “master” artificial intelligence and advanced computingAs he explained, the administration is trying to avoid a system in which export bureaucracy stifles the competitiveness of American industry against competitors who operate with fewer obstacles.

Senator John Fetterman, for his part, has expressed skepticism about the necessity of these sales, recalling that Nvidia is now the most valuable company in the world by market capitalizationFrom their perspective, it is not clear that the chip giant needs to further boost its revenue at the cost of increasing interdependence with China in such a sensitive area.

National security versus technological competitiveness

Beyond the political tension, the White House insists that the priority remains maintain control over strategic technologyLimiting the export of the most advanced chips—such as Blackwell or Rubin—and subjecting H200 chips to case-by-case licensing is part of a technological containment policy aimed at preventing China from closing the gap simply by buying American hardware.

This logic puts companies like Nvidia in a delicate position: the company must to scrupulously comply with national security criteria If it wants to retain its licenses, it effectively acts as a technical extension of Washington's export control regime. Each mismanaged transaction could result in sanctions, investigations, or the revocation of permits.

Exclusive content - Click Here  The best USB-C hub: buying guide

For the industry as a whole—including cloud providers, systems integrators, and AI companies in Europe—this environment implies navigating a sea of ​​overlapping technological and political bordersIt's no longer just about evaluating price and performance: the location of data centers, applicable jurisdiction, and geopolitical risk are factors that increasingly weigh heavily when designing global artificial intelligence projects.

Impact and reading from Europe and Spain

From a European perspective, and especially for EU countries like Spain, this shift by Washington has several relevant implications. First, It reinforces Europe's dependence on US technological decisions.This is because much of the advanced computing power used by companies, universities, and research centers across the continent continues to rely on Nvidia chips and cloud services based on North American hardware.

The United States' European partners, including governments driving large AI and supercomputing projects, are being pushed to align its export policy and use of advanced chips with the US framework if they want to maintain preferential access to these technologies. This This could mean giving up part of the business with China or other destinations considered sensitive., in exchange for strengthening transatlantic security ties.

For Spain, which aspires to to position itself as a hub for data, supercomputing centers and AI development in southern EuropeThis scenario presents a mix of challenges and opportunities. On the one hand, regulatory uncertainty complicates long-term plans for companies and governments when it comes to investing in computing infrastructure based on US technologies. On the other hand, Washington's desire to ensure Western leadership in semiconductors and AI hardware could translate into New industrial alliances, investments and European projects for the manufacture and design of next-generation chips.

The H200 as a symbol of the new technological rivalry

Nvidia h200

The battle for control of the H200 illustrates the extent to which technology has become a central playing field of global competitionThese chips are not only used to train language models or image recognition systems; they are also critical components for complex simulations, massive data analysis, and next-generation military applications.

By restricting and regulating its export, the United States intends to slow down certain critical projects in the hands of their rivals And, at the same time, maintain its lead in the race for advanced artificial intelligence. China, for its part, is responding by accelerating the development of its own solutions and building an alternative supply chain less exposed to sanctions or vetoes.

The H200 chips have been transformed into something more than a cutting-edge technological productThey are a barometer of the balance of power between major powers and a reminder that economic and military dominance in the coming decades will largely be determined in the realm of advanced computing and AI infrastructure. For Europe and Spain, the challenge lies in not remaining mere spectators but finding their place in a race where every license, every tariff, and every regulatory decision can change the course of the sector.