What’s behind China’s Nvidia monopoly investigation? – Asia Times

The State Administration of Market Regulation in China is looking into American tech leader Nvidia for possible anti-monopoly violations and a deal Nvidia reached with Mellanox Technologies in 2020.

The initial claim only makes sense as retribution for the most recent round of US sanctions against China, but the next one may hold up. Nvidia is disputing the claims.

Nvidia has undoubtedly refuted rumors that it intends to cutback sales to China, which generated 15 % of its revenue during the three months to October ( Nvidia’s fiscal third quarter ). Nvidia is expanding its presence in China while focusing on industries like automatic driving, which are not subject to trade settings.

Because US Commerce Secretary Gina Raimondo and her department’s Bureau of Industry and Security ( BIS ) have already prohibited China from exporting its most advanced GPU processors, it seems illogical for China to accuse Nvidia of monopoly practices.

Perhaps China should congratulate the US government for granting Huawei and another Chinese Circuit design firms the chance and a strong opportunity to launch their own AI chipset business while their most formidable rival’s hands are tied.

In 2022, the BIS banned imports of Nvidia’s top-end A100 and H100 chips to China. In 2023, it banned exports of the A800, a dumbed-down variation of the A100 designed specifically to meet BIS needs.

The A800 was a best seller in China so the BIS lowered the bar, forcing Nvidia to style another device, the H20, with even lower efficiency. But Chinese AI chips, led by Huawei’s 910B, proved to be economical with the H20, which has not sold also in China. And Huawei claims that its leader, the 910C, matches the effectiveness of Nvidia’s H100.

In any case, buyers from Baidu, Tencent, and another Chinese companies started switching to Huawei and various domestic chips once it became apparent that US sanctions were not based on strict technical standards as US federal officials claimed but could be changed at any time to just condemn China.

Export of Nvidia’s fresh and considerably more powerful Blackwell B200 AI chips to China are banned under existing punishment, but a dumbed-down type that may be called B20 is apparently being prepared. But why would it succeed as much as the H20 did?

Regarding Mellanox Technologies, the Chinese state approved the purchase under the condition that Nvidia may continue to offer its wire technology to Chinese customers without discrimination, no bundle them with its own GPUs, and guarantee interoperability of its own GPUs with another wire products.

Nikkei Asia points out that” It is unclear which of these words Nvidia is alleged to have been broken.”

InfiniBand and Ethernet connection adapters, switches, and other applications are created and provided by Mellanox, an Israeli firm focused on high-performance technology, data centers, sky storage, and financial services. Nvidia made arrangements to buy the entire company in 2019

Where might China have come up with the idea of conducting an antitrust investigation? In July, France confirmed that it was looking into Nvidia for alleged anti-competitive practices, and in September, the US Department of Justice subpoenaed the company for information regarding its potentially restrictive marketing practices. Nvidia’s share of the market for AI processors is thought to be about 90 %, aside from China.

US sanctions prevent China from purchasing EUV lithography systems made by ASML of the Netherlands, which not only restrict the export of advanced AI processors from Nvidia and its smaller rival AMD to China.

Due to this, it is impossible for Chinese designers to create integrated circuits ( ICs ) with design standards less than 5 nm, and 7 nm is the closest they can go with sufficient efficiency. Blackwell processors, on the other hand, are made by TSMC using a 4nm process. In the quest to create more sophisticated AI processors, Nvidia is currently steps ahead of the Chinese.

However, chasing Nvidia is not the only way to make AI progress. In recent months, Chinese research institutions have reportedly developed AI processors based on RISC-V ( pronounced “risk-five” ) open-architecture design standards and large language models for military use based on open-source models, including Meta’s Llama.

The Chinese aren’t the only ones using RISC-V. In February, US-based up-and-coming Nvidia competitor Tenstorrent made an announcement to license its RISC-V CPU technology to Japan’s Leading-edge Semiconductor Technology Center for the development of an AI processor made up of chiplets produced and packaged in Hokkaido by Japan’s new IC foundry Rapidus. &nbsp,

China is using chiplets to circumvent its lack of access to EUV lithography equipment. According to the MIT Technology Review, as stated:

In contrast to traditional chips, which integrate all components on a single piece of silicon, chiplets take a modular approach. Each chiplet has a dedicated function, like data processing or storage, they are then connected to become one system.

Since each chiplet is smaller and more specialized, it’s cheaper to manufacture and less likely to malfunction. In order to improve performance, individual chiplets in a system can be replaced at the same time, while other functional components remain the same.

Based on Reduced Instruction Set Computer Design Principles, RISC-V is an open standard instruction set architecture. It is a free, non-proprietary platform for the development of IC processors.

RISC-V is a great option for China as well as for the EU and to smaller companies and IC designers looking to establish a low-cost independent presence in the semiconductor and computing markets. It is an alternative to Arm, Intel, AMD, and Nvidia.

China has been forced by US sanctions to accelerate the creation of a national ecosystem for advanced ICs based on RISC-V.

The RISC concept was conceived at the University of California, Berkeley, in 2010. The Institute of Computing Technologies of the Chinese Academy of Sciences was one of the founding members of the RISC-V Foundation, which was established in 2015 to support and manage the technology. Other Chinese members of the foundation include Huawei, ZTE, Tencent and Alibaba.

The foundation left the United States in 2020 to avoid potential interference from US President Donald Trump by forming the RISC-V International Association. Beyond the reach of US sanctions, China is now estimated to account for about half RISC-V core shipments worldwide.

Meanwhile, DigiTimes reports that Nvidia has hired “hundreds” of new employees in China to work on autonomous driving. For autonomous electric vehicles, BYD and about a dozen other Chinese automakers have adopted its DRIVE Orin system-on-a-chip ( SoC), which generates annual revenues of more than$ 1 billion for Nvidia.

Looking ahead to 2025, at least five Chinese EV makers – BYD, XPeng, GAC-Aion, Li Auto and Zeekr – plan to use DRIVE Thor, the successor to DRIVE Orin. But DRIVE Thor incorporates the generative AI capabilities of Nvidia’s leading-edge Blackwell architecture. Will the US government also make an effort to stop this?

The Chinese are hedging. BYD, Li Auto, ChangAn and many other Chinese automakers are also working with Horizon Robotics, China’s leading computing solutions for advanced driver assistance systems ( ADAS ) and automated driving ( AD ) for consumer vehicles. So is Volkswagen. Horizon Robotics went public in Hong Kong’s largest IPO this year in October.

Follow this writer on&nbsp, X: @ScottFo83517667