Nvidia and AMD: Weapons in the AI ​​Chip War – TipRanks.com

On the 20th^ththe geopolitics of the last century centered on oil; on the 21st^st the geopolitical dominance of centuries is determined by technological supremacy. As in the previous century, the United States is defending its territory against another superpower, this time not the Soviet Union but China; the new oil is the chips needed for the development of AI technology.

As the technological rivalry between the two empires continues to intensify, the main focus is shifting more and more to Artificial Intelligence (AI). It is much more than ChatGPT’s ability to write great texts and Midjourney’s ability to produce superb images, although these skills can do a lot of political, economic and social damage if they fall into the wrong hands (think disinformation, deep fakes, espionage, etc. ).

Scientists and tech entrepreneurs agree that Artificial Intelligence will bring about a major technological change comparable to the invention of the Internet; the technology has the potential to increase economic productivity and competitiveness, as well as impact or even predetermine military dominance. Earlier this year, the Pentagon, backed by tech leaders, sounded the alarm about China’s development of AI-based warfare practices aimed at stockpiling advanced capabilities to counter the United States in a potential conflict.

The AI ​​arms race officially began in 2017, when China announced its plan to become the global leader by 2030. It was clear then, as it is today, that modern global leadership ambitions are not being realized. mass-producing cheap shoes for Western buyers, but winning on the front of technological progress; AI technology is at the forefront of these efforts. Since no AI advancement is possible without the latest chip technology, US efforts to counter China’s AI advances are focused on chips. China is investing heavily in the development of the domestic semiconductor industry, but still lags behind the United States in the AI ​​chip sphere by a few years. Furthermore, the progress of Chinese chip-making ventures depends on their access to advanced Western chip-making equipment.

The Biden administration has continued and escalated previous US government policies on China. Last year, the administration imposed export controls aimed at preventing the sale of the most advanced chips needed to develop and train AI models and the chip-making equipment needed to produce them to entities operating in China. Back then, it was thought that sweeping restrictions would deal a major blow to companies in the industry, especially suppliers of chips and chip making equipment with high exposure to China, such as Applied Materials (AMAT), Lam Search (LRCX extension), KLA Corporation (KLAC extension), and other. However, while earnings have suffered, the industry has adapted; This year’s rally in all things AI has passed to the left of all concerns of recent years.

Source: Google Finance

Now is the time for the next phase of the US-China AI divorce saga. According to reports, the US government is planning more restrictions on AI chips and equipment in China to further ensure that the Asian giant stays behind the US in the AI ​​arms race.

After the restrictions were put in place in recent years, the United States quickly discovered that Chinese companies had found ways to circumvent them. While smuggling or buying through third parties and shell companies is tough to crack down on, the US government is taking steps to close major loopholes through which China gained access to AI chips, post-ban.

Chipmakers are partly to blame for the next wave of restrictions. In their attempt to defend their share of the Chinese market, they have made every effort to circumvent government demands.

The main culprit for this is Nvidia (NVDA extension). Following the announcement of the earlier restrictions, the company engineered a slightly lower-performance version of its A100 AI GPU (graphics processing unit) that fell within government-set limits for chips shipped to China. The Chinese chip, dubbed the A800, runs at 70% the speed of the A100 GPUs that power AI computations in data centers in the US, but is still faster than anything made in China; moreover, it can be used in automatic targeting systems for drones or surveillance systems. Under the new export restrictions, A800 chips will need a government license for exports to China, i.e., effectively banned.

In addition to using Nvidia chips previously restricted by law, Chinese companies have exploited other loopholes to gain access to technology that enables artificial intelligence. While restrictions have focused on physical exports of the chips, Chinese tech firms have leased access to controlled chips via the cloud. Top Amazon Cloud Service Providers (AMZ extension), Microsoft (MSFT extension) and Alphabet (GOOGL) have long offered remote access to the computing power of AI chips through leasing; China has made accelerated use of these offers since the restrictions were introduced last year. In March of this year, Nvidia said it also plans to lease supercomputing power to Chinese companies willing to leverage their demand for its high-end technology.

Apparently, the Biden administration doesn’t view this approach lightly. According to reports, the government is planning to limit the rental of cloud services to Chinese companies. If that weren’t enough, the government is reportedly outlawing US investment in advanced semiconductor, artificial intelligence and quantum computing companies in China, intending to prevent US money and expertise from supporting the development of advanced technologies made in China.

The new restrictions will surely hurt the revenues of US chipmakers like Nvidia and Advanced Micro Devices (AMD), as well as companies looking to challenge the dominance of Nvidias and AMD in AI chips, such as Intel. They will also apply to companies that sell products containing advanced chips, such as AI data center servers. Dell Technologies (DELL), Hewlett-Packard (HP extension) and Super Micro Computers (SMCI extension) fall into the category. Cloud providers will also be impacted by the ban on their AI rental service, if only marginally.

Nvidia controls about 80% of the market for accelerator chips in AI data centers operated by Amazon, Alphabet and Microsoft, and derives about 22% of total revenues from China, with revenues from data center chips sold to the Asian superpower ranging between 7% and 10% of the entire revenue pie.

Source: Google Finance

AMD, the second largest GPU maker, aspires to a larger market share; intel (INTC), which was a latecomer to the AI ​​game, is just starting to make headway in the AI ​​chip sphere. While Intel, a newcomer to AI, still has nothing to lose from the ban on AI chips in China, AMD will suffer. While it has significantly reduced its reliance on Chinese sales over the past decade, they still make up about a quarter of its total revenue (however, much of the income comes from non-AI related products).

Obviously, the new restrictions would be a serious headwind for both Nvidia and AMD in terms of revenue generation in the near term, until they figure out a way to adjust. Citigroup (c) said demand for AI will outstrip supply in the near term; the two chipmakers remain at the forefront of a surge in AI development. The Bank of America (bac extension) predicts that the Chinese market will ultimately account for less than 10 percent of the total addressable market for AI-related accelerator chips, which is expected to reach more than $100 billion; Evercore ISI called the additional restrictions “a small slowdown on the AI ​​highway.”

The United States is not alone in this battle, as all of its allies have now returned to the anti-China banner, reminiscent of the Cold War era. The domiciles of global manufacturers of advanced chip making equipment, Japan and the Netherlands, have also limited tool exports to China. Dutch tech mega-cap ASML Holding NV (ASML) is facing an export ban on its chip making machines to China; the restrictions should provide a model for other European governments to follow suit.

ASML is the world’s only manufacturer of high-end semiconductor lithography equipment, so the ban on its equipment exports is one of the most effective bottlenecks for the progress of Chinese AI chips. The company is a vast ecosystem of knowledge and effectiveness; there is no chance that China will be able to replicate its products at least within a decade. ASML is so overwhelmed by global demand for its equipment that it does not see export restrictions in China as a danger to its bottom line.

Japan also joined the ranks of the United States earlier this year in restricting its domestic companies’ exports to China of 23 types of cutting-edge chip-making technology and equipment, which now need to be upgraded. government licenses. The restrictions apply to companies such as Tokyo Electron (TOELF), Screen Holding Co. (DINRF), Nikon Corporation (NINOF) and Lasertec Corp. (LSRCF extension), which are vital parts of the global AI chip supply chain. These companies derive 20% to 25% of their revenues from sales in China, which means they will be seriously affected by the ban, at least in the short term.

However, on a larger scale, this pain is worth the long-term gain. With Japan and the Netherlands joining the restrictions club, Chinese efforts to make rapid advances in AI will be severely curtailed, ensuring this round of the game is won by the US and its Western allies.