Bitget App
Trade smarter
Buy cryptoMarketsTradeFuturesStocksEarnInstitutionAI & More
Trump Opposes Global AI Regulation, While the US Tightens Overseas Computing Power Permits! Nvidia and Southeast Asian Data Centers Face Divergence

Trump Opposes Global AI Regulation, While the US Tightens Overseas Computing Power Permits! Nvidia and Southeast Asian Data Centers Face Divergence

智通财经智通财经2026/11/19 07:26
Show original
By:智通财经

President Donald Trump warned that any “strict” regulation of artificial intelligence would “stifle” its development. Meanwhile, the United States has imposed new restrictions on access to advanced artificial intelligence technology.

According to Odaily, former U.S. President Trump stated at the APEC Leaders' Informal Meeting held in Shenzhen on Thursday that the U.S. opposes global regulations that could hinder the development of artificial intelligence, while acknowledging that AI technologies do pose certain risks and require careful management.

At the meeting, Trump remarked that artificial intelligence is "almost entirely beneficial, but there are some downsides," emphasizing the need to address related risks appropriately but opposing any attempts to control or stifle its development.

It is noteworthy that Trump's comments came as the U.S. government had just further tightened export controls on advanced AI chips and offshore computing power services. The U.S. Department of Commerce recently introduced new regulations extending some export license requirements to remote access to overseas data centers' computing power, aiming to restrict companies in China and other targeted destinations from acquiring advanced AI computing capabilities through third-country data centers.

This means that while the Trump administration wants U.S. companies to accelerate the development and commercialization of AI technologies, it is also strengthening control over the end users of advanced AI computing power. Although these two policies target different areas, they both reflect the U.S. government's policymaking between promoting AI industry growth and maintaining its technological edge.

Meanwhile, media reports regarding former NVIDIA CEO Jensen Huang and his relationship with a Chinese businessperson have also brought further attention to the domestic debate on advanced chip export controls in the U.S. However, related espionage allegations have not yet been independently verified.

Trump Opposes Regulation that "Stifles AI": Why Is the U.S. Simultaneously Tightening Overseas Computing Power Licenses?

On November 19th, at the APEC summit in Shenzhen, Trump stated that the risks of artificial intelligence require prudent management, while warning that excessive regulation may impede technological development. This position was made almost simultaneously with the U.S. government's recent tightening of advanced AI chip exports and overseas computing power licensing.

The White House hopes U.S. companies can rapidly develop and deploy cutting-edge models, while the Commerce Department seeks to restrict specific overseas users from gaining U.S. computing power sufficient to train and operate these models. The former is concerned whether regulation may slow down technological innovation, while the latter is focused on who ultimately receives the computing power enabled by U.S.-designed chips.

Disagreements within the U.S. over AI regulation are mainly centered on who should assess the risks of frontier models and when such assessments should intervene.

Previously, the White House promoted the establishment of relevant assessment mechanisms, requiring that for frontier models reaching specific network capability thresholds, the federal government be allowed to conduct evaluations up to 30 days in advance before opening to other trusted partners.

Meanwhile, expert recommendations released in California on November 18 proposed that the state government set baseline AI safety standards and require companies to undergo independent testing during model training, development, and deployment phases. However, these recommendations are still in the policy discussion stage and cannot yet be regarded as regulatory requirements already in effect.

For model developers, the main difference between these two approaches is whether safety assessments are mainly carried out before model release, or if they must span multiple phases of model training, development, and deployment.

The recent overseas computing power licensing rules promoted by the U.S. government address another technological reality: Even if advanced AI chips remain in overseas data centers, customers can still remotely access entire groups of GPUs via the internet to train models, or continually access GPUs to process user requests after model deployment.

Therefore, U.S. regulation of advanced AI technology is further expanding from the physical export of chips to the end users of computing power services.

From Where the Chips Are Shipped, to Whom the Computing Power Is Rented

The new rules announced by the U.S. Department of Commerce on November 16, set to take effect on November 19, introduce two licensing requirements.

The first requirement concerns where the equipment goes: Exports, re-exports, or in-country transfers of designated advanced chips and related equipment to Singapore, Malaysia, Thailand, Indonesia, and Brazil generally require approval from the U.S. government.

The second requirement focuses on the end user of computing power, regardless of the five aforementioned countries. If the U.S. is aware that relevant equipment will be used in data centers outside the U.S. and will provide remote computing power exceeding regulatory thresholds to enterprises whose headquarters or ultimate parent companies are located in China or other restricted destinations, those equipment transactions also require licenses.

In addition, the U.S. government has provided a 90-day transition arrangement for some eligible transactions, and certain compliant commercial recommendation systems may receive exemptions.

This means that even if GPUs remain in overseas data centers in Singapore, Malaysia, and elsewhere, as long as those data centers provide computing services exceeding regulatory thresholds to restricted end users, the equipment transactions may still be subject to U.S. export licensing rules.

This rule is particularly noteworthy with regard to its impact on large Asian cloud computing operators and data center operators. For data centers that have already been built and energized, if the original customers' computing power usage is limited, operators may have to find new compliant tenants. Re-signing and deploying new business takes time, during which data center utilization and leasing revenue could be affected.

NVIDIA and Data Center Operators Face Different Impacts

This also explains why the same regulation could have significantly different impacts on the profitability of NVIDIA and overseas data center operators.

For chip suppliers, if some customers cannot continue to obtain advanced AI chips or computing power services, the equipment can still be resold to other buyers compliant with U.S. regulations. As long as global AI computing power demand remains robust, suppliers may not have to bear the entire loss in demand.

BofA believes that NVIDIA's global demand still exceeds supply, and some orders may be taken up by other compliant buyers. Goldman Sachs also points out that after NVIDIA considered the new leasing restrictions, it still raised its growth expectations, reflecting ongoing market confidence in its overall demand outlook.

NVIDIA's guidance for fourth-quarter revenue is $130 billion, excluding data center computing revenue from China. For investors, the key question is whether export restrictions will materially dampen NVIDIA's serviceable global demand, or simply change the ultimate flow and client structure for its chips.

In contrast, overseas data center operators may face more direct risks.

Data center construction involves upfront investments in land, power supply, facilities, servers, and financing. Even if original customers cease using computing power, operators must still bear depreciation, facility, and financing costs. If new compliant clients are not found promptly, a reduction in billable GPU usage could directly impact leasing income and project returns.

Morgan Stanley accordingly notes the other side of the risk: model developers reliant on overseas clusters may delay their training schedules, and companies renting GPUs themselves to offer overseas inference services may be more directly affected by the new rules.

On November 17, shares of Malaysian data center infrastructure firms MN Holdings and Critical Holdings both fell nearly 20%, while NVIDIA's local partner YTL Power dropped nearly 10%. Such market performances reflect how investors are starting to evaluate the potential impacts of tighter overseas computing power licensing on local data center project construction and future leasing demand.

Overall, this tightening of overseas computing power licensing by the U.S. does not signify a decline in global AI computing demand itself, but rather raises the compliance threshold for cross-border advanced AI computing flows. For chip suppliers like NVIDIA, whether demand can be absorbed by other compliant customers will determine the actual revenue impact of export restrictions. For Southeast Asian data center operators, the ability to promptly find replacement tenants and maintain utilization rates may become a key factor affecting project returns.

0
0

Disclaimer: The content of this article solely reflects the author's opinion and does not represent the platform in any capacity. This article is not intended to serve as a reference for making investment decisions.

Understand the market, then trade.
Bitget offers one-stop trading for cryptocurrencies, stocks, and gold.
Trade now!

You may also like

IEA raises energy 'demand alert': Cooling surpasses heating to become the primary growth engine, AI data centers push the power grid to its limits

The International Energy Agency (IEA) annual flagship report "World Energy Outlook" points out that global cooling demand exceeds heating demand. Additionally, the rapid growth of data centers will push power grid loads to their limits.

智通财经•2026/11/19 06:41
IEA raises energy 'demand alert': Cooling surpasses heating to become the primary growth engine, AI data centers push the power grid to its limits

The euro's decline shows no sign of stopping after falling below the 1.10 mark! France's 2027 election may become the biggest risk; the most pessimistic bet in more than six years emerges

The euro has fallen to its lowest point in 18 years, with little sign that the decline is easing. The market increasingly expects that, as France faces both intensified political division and fiscal pressure, next year's French elections will continue to exert pressure on the euro.

智通财经•2026/11/19 06:41
The euro's decline shows no sign of stopping after falling below the 1.10 mark! France's 2027 election may become the biggest risk; the most pessimistic bet in more than six years emerges