Nvidia stated it anticipates a $5.5 billion hit following US restrictions on its ability to export artificial intelligence chips to China, causing the Silicon Valley giant’s shares to decline in after-hours trading.

The group announced late Tuesday that the H20 chip, tailored for the Chinese market to comply with export controls already preventing the sale of its most powerful chips in China, will now require a specific license for sales to customers there.

Nvidia noted that the US indicated this move is necessary to address the risk of H20 chips being used “in a supercomputer in China.”

The chip manufacturer stated it would take a $5.5 billion charge related to H20 chips in the quarter ending April 27. The company’s shares fell by 6% in after-hours trading on Tuesday, while futures tracking the technology-heavy Nasdaq 100 index dropped by over 1%.

Washington’s crackdown on H20 chips is the latest instance of the US using tariffs and other trade barriers to increase pressure on Beijing. President Donald Trump raised custom duties by 145% on goods imported from China, though he granted a temporary reprieve on some consumer electronics.

White House Press Secretary Karoline Leavitt stated Tuesday that “the ball is in China’s court,” urging China to strike a new trade deal with the US.

The US Commerce Department confirmed later on Tuesday that it had issued new export license requirements for AMD’s MI308 and equivalent chips, in addition to the H20.

“The Commerce Department is committed to acting pursuant to the president’s directive to protect our national and economic security,” a spokesperson stated.

AMD is Nvidia’s closest direct competitor in the artificial intelligence data center chip market. The company did not immediately respond to a request for comment.

This move by the US underscores how Nvidia, a chip designer at the heart of the artificial intelligence boom that has seen uncontrolled growth over the past year and briefly made it the world’s most valuable company, is exposed to geopolitical tensions between Washington and Beijing.

On Monday, the Trump administration initiated a national security investigation that could lead to new tariffs on semiconductors, while avoiding immediately imposing higher taxes on chips.

The restrictions come despite Nvidia CEO Jensen Huang joining other tech executives seeking favor with Trump. Huang recently dined with Trump at the Mar-a-Lago resort and met with the president at the White House in January.

Nvidia also stated on Monday that it would spend up to half a trillion dollars on US artificial intelligence infrastructure over the next four years through partnerships with companies like Taiwan’s TSMC and Foxconn.

The company introduced the China-focused H20 processors last year after the Biden administration imposed export controls on its chips.

These processors are less powerful than the top-tier graphics processing units, or GPUs, sought after by Microsoft, OpenAI, Google, and Amazon.

Despite its lower performance, the H20 still saw solid demand in China. However, Beijing has taken steps to encourage local tech companies to use domestic chips from companies like Huawei and could freeze out Nvidia’s products with new energy efficiency rules.

Nvidia’s shares had lost approximately 16% year-to-date as of Tuesday’s close, as concerns grew about the escalating arms race between the US and China over the infrastructure powering artificial intelligence. They have also been dragged into a broader market sell-off triggered by the escalating trade war.

On Tuesday, Bernstein analysts stated that the H20 accounted for approximately $12 billion of Nvidia’s $17 billion in revenue in China, but there was no clarity at this stage on whether licenses would be granted or whether the product line would be completely “wiped out.”

The launch of Nvidia’s newest artificial intelligence chips has faltered as successive US administrations have sought ways to control the export of the technology.

The US argues that supercomputers can be used in everything from China’s development of hypersonic weapons to the modeling of nuclear weapons.

China has repeatedly accused the US of using national security tools like export controls to impede its economic development.

The “artificial intelligence proliferation” rule, implemented in the final days of the Biden administration, is set to take effect in May unless a decision is made to withdraw it by the Trump administration. This rule will impose much tighter controls on where the most powerful US chips can be exported, using a “tiered” licensing system that restricts exports to all but a handful of countries.

Last week, Republican senators wrote a letter to Commerce Secretary Howard Lutnick, urging the administration to rescind this rule, which has been met with pushback across the sector, including from Nvidia.

NATO has purchased an artificial intelligence-based military system from Palantir, the US software company headed by Donald Trump supporter Peter Thiel and with strong Pentagon connections.

The alliance’s choice comes at a time of increased concern among European members about a possible US withdrawal, following Trump’s threat to stop protecting the continent unless capitals significantly increase defense spending.

NATO is also racing to keep pace with the development of artificial intelligence military capabilities by rivals such as China. According to the alliance’s statement on Monday, Palantir’s Maven Smart System (MSS NATO) uses generative artificial intelligence, machine learning, and large language models to provide “commanders with a secure, shared operational capability” and will be used to support NATO operations.

Such “battlefield management systems” allow teams of 20-50 soldiers to do the work of reviewing battlefield data that previously required teams of hundreds or even thousands of personnel in recent conflicts such as Afghanistan and Iraq.

Noah Sylvia, an analyst at the Royal United Services Institute (RUSI), a London-based think tank, told the Financial Times (FT), “It can replace all those teams doing quite boring tasks.”

Sylvia noted that France had developed “Artemis,” which he described as “not a competitor but a domestic alternative” to Palantir’s Maven system, to avoid dependence on the US.

NATO is moving quickly to enhance its defense technology capabilities. According to the alliance, the completion of this contract, which was “one of the fastest contracts in NATO history,” took only six months, and the system is expected to be operational within the next 30 days.

Sylvia said, “To have it procured in six months is insane by defense standards. Software usually takes years to procure, certify, and then deploy, and by that time, it’s usually out of date.”

NATO announced that this acquisition, which “demonstrates a strong and lasting partnership between the North American and European technology base,” was completed last month. The financial terms of the deal were not disclosed, but it is likely to be one of Palantir’s most significant defense contracts this year.

Thiel, one of Silicon Valley’s most prominent figures, was a leading supporter of Trump’s initial presidential candidacy in 2016 and played a major role in the selection of JD Vance, his vice president, as Trump’s vice-presidential candidate. Thiel is a pioneer of the “techno-libertarian” group in Silicon Valley and is known for his anti-democratic views.

According to federal records, Palantir has won more than $2.7 billion in US government contracts since 2009, with more than $1.3 billion of that from the Department of Defense. Palantir’s market capitalization has surpassed the total of the Pentagon’s traditional top 5 contractors.

The company’s shares have increased by more than 300% in the last 12 months, as investors expect the company to benefit from the Trump administration’s defense spending, as well as commercial customers using artificial intelligence systems.

The US Army is also using its own version of Palantir’s Maven technology and signed a five-year contract for $99.8 million for this technology last September.

A similar system was also used in Ukraine. Maven is used to combine satellite imagery with other battlefield intelligence sources, scan targets, and use machine learning to accelerate attacks.

The Pentagon’s Project Maven system dates back to 2017, when Google began using its technology. Google later withdrew from the program in 2018 after thousands of its employees protested the use of artificial intelligence in warfare.

Palantir provides NATO with a customized version of Maven that provides a platform where other software applications and data sources can be integrated.

Palantir’s senior advisor Shon Manasco said, “We are proud to support NATO’s effort to enhance its deterrence by establishing an AI-backed warfighting platform. This partnership underscores the alliance’s commitment to fearlessly lead in technological innovation.”

NATO said that MSS NATO will “enhance intelligence fusion, targeting, battlefield awareness, operational planning, and decision-making processes.”

General Markus Laubenthal, Chief of Staff at NATO’s military headquarters Shape [Supreme Headquarters Allied Powers Europe] in Belgium, said, “ACO [Allied Command Operations] is at the forefront of embracing technologies that make NATO more agile, adaptable, and responsive to emerging threats.”

Laubenthal added that innovation is the foundation of NATO’s warfighting capability.

The commander also praised MSS NATO for its capacity to “leverage complex data, accelerate decision-making,” and add “real operational value.”

US President Donald Trump stated that he would continue to impose tariffs on phones, computers, and popular consumer electronics products, considering the weekend’s exemption a procedural step in his broader effort to reshape US trade.

The delay, announced late Friday and exempting a range of popular electronic products from the 125% tariff applied to China and the 10% fixed rate applied worldwide, is temporary and part of a long-standing plan to apply a different and specific tax to the sector.

Shortly after finishing a round of golf on Sunday, Trump posted on social media, “NOBODY is getting ‘off the hook.’”

According to Trump, the exempted products are “just being moved into a different Tariff ‘bucket,’” and the administration will take “a look” at semiconductors and the entire electronic supply chain.

Speaking to reporters on Air Force One, Trump said that decisions would be made soon, with details on the tariff rate for semiconductors to be announced within the next week.

However, Trump also signaled that he is open to discussions with companies regarding the scope of the sectoral tariff on semiconductors and products based on them, such as iPhones and tablets.

“We’re going to discuss it, but we’re also going to talk to the companies. You have to have a certain flexibility. Nobody should be so rigid,” Trump said.

Friday’s pause appears to be a temporary victory for Apple and other manufacturers, particularly those relying on Chinese production.

According to a report in Bloomberg, Apple has managed to avert its biggest crisis since the pandemic, at least for now.

Trump’s 125% tariffs on goods manufactured in China threatened to disrupt the supply chain as severely as the Covid-19 pandemic did five years ago.

By exempting many popular consumer electronics on Friday night, the US President handed Apple a major win. These products include iPhones, iPads, Macs, Apple Watches, and AirTags.

A new and lower sectoral tariff may also be applied to goods containing semiconductors, but a 20% tariff is still applied to electronic products shipped from China.

Until Trump reinstates tariffs on electronic products, the surprise exemption is a win for Apple and the consumer electronics industry, which still largely depends on China for production.

Before the latest exemption, the iPhone maker had a plan: adjusting its supply chain to produce more US-bound iPhones in India, where they would be subject to much lower taxes.

Apple executives believed this would be a short-term solution to avoid Chinese tariffs and prevent high price increases.

Given that iPhone facilities in India are producing at a rate of more than 30 million iPhones per year, production in this country alone could meet a significant portion of American demand. Apple sells approximately 220 million to 230 million iPhones annually these days, with about one-third going to the US.

Implementing such a change smoothly would have been difficult, especially as the company is already approaching production of the iPhone 17, which is primarily to be manufactured in China. Fears had grown in Apple’s operations, finance, and marketing departments about the impact on the new phones’ launch in the fall.

According to Bloomberg, the company would have to accomplish the daunting task of moving more iPhone 17 production to India or elsewhere in just a few months.

In this case, it would probably have to raise prices (which is still possible) and fight with suppliers for better profit margins.

Another concern for Apple was: How would China retaliate if the company increased its production outside of China even faster?

Apple derives approximately 17% of its revenue from this country and operates dozens of stores, making it an outlier among US-based companies.

According to Morgan Stanley estimates, the iPhone is Apple’s biggest money-maker, and approximately 87% of these products are manufactured in China. About four out of every five iPads and 60% of Macs are also produced in this country.

Together, these products account for approximately 75% of Apple’s annual revenue. Yet, the company now produces almost all of its Apple Watches and AirPods in Vietnam. Some iPads and Macs are also produced in this country, and Mac production is expanding in Malaysia and Thailand.

According to Morgan Stanley estimates, the company makes approximately 38% of its iPad sales and about half of its Mac, Apple Watch, and AirPods revenue in the US.

It is unlikely that Apple will completely decouple from China, which has been its manufacturing hub for decades. Even if Trump forced Apple to manufacture iPhones in the US, the lack of domestic engineering and manufacturing capabilities could make this nearly impossible in the short term.

On the other hand, according to US Commerce Secretary Howard Lutnick, smartphones and other electronic devices that won exemptions will be part of the new tax applied to semiconductors.

Speaking on ABC’s “This Week” on Sunday, Lutnick signaled that the tariff delay was temporary and reiterated Trump’s long-standing plan to apply a different, specific tax to the sector.

Since the announcement of the tariff wave on April 2, Apple and other tech companies’ lobbyists have been pressing the White House for exemptions.

But discussions have become more urgent in recent days after a series of retaliations between Washington and Beijing led to a 145% tax on imports from China.

The potential impact has become even more acute after Trump halted higher tariffs on other countries. This meant an advantage for Apple’s competitor, Samsung Electronics, which produces its phones outside of China.

Apple and other companies have emphasized to the Trump administration that while they are willing to increase their investments in the US, there is little benefit to moving final assembly to this country.

Instead, they argue that the US should focus on bringing back higher-value jobs and encouraging investment in areas such as semiconductor manufacturing.