The U.S. Commerce Department plans to release a proposed rule as early as Monday that would ban the use of connected car hardware and software made in China and Russia, according to people familiar with the matter.

The Commerce Department has been meeting with industry experts in recent months to address security concerns posed by a new generation of so-called "smart cars." The action would include bans on the use and testing of Chinese and Russian technology for autonomous driving systems and in-vehicle communications systems. While the bans focus primarily on software, the proposed rules would also cover some hardware.

Many gasoline and electric vehicles today are equipped with devices that connect to the internet or cloud services, making them potential targets for hackers. The upcoming restrictions stem from an investigation into cybersecurity risks of Chinese vehicle software launched by President Biden in March.

The Biden administration's main concern is to prevent China or Russia from hacking or tracking vehicles by intercepting communications from software systems created by their domestic companies. The rule would also have a certain protectionist tone because most new cars are connected to the Internet at least through their infotainment systems, so Chinese automakers could be banned from selling their vehicles in the United States if their vehicles use their connected technology.

In May, the U.S. government imposed a 100% tariff on Chinese electric vehicles, arguing that the Chinese government is subsidizing its auto industry and increasingly exporting its excess capacity while American companies increase production of electric vehicles.

The U.S. Commerce Department declined to comment. Lael Brainard, director of the White House National Economic Council, plans to speak in Detroit on Monday about the Biden administration's efforts to "strengthen the U.S. auto industry."

China has become a leader in electric vehicles and smart car components, thanks in part to extensive government subsidies and support. BYD Co. sold more all-electric vehicles than Tesla Inc. in the fourth quarter of last year, and global automakers are increasingly relying on Chinese suppliers for the technology needed for connected cars. China says it respects data privacy and the security of its foreign customers and adheres to the principle of fair competition.

The new restrictions, which will be enforced by the Commerce Department, are intended to prevent Chinese companies from collecting data on American drivers, especially individuals, and transmitting it back to China. The rules will also effectively prevent Chinese suppliers from establishing a larger foothold in the United States, giving the U.S. auto industry time to build its own connected car supply chain.

The Commerce Department's proposal includes a phased implementation schedule for different software and components, sources said. Officials plan to issue the final rules by January 2025, and there will be a 30-day public comment period after the proposal is published.

The rules will also govern the hardware of vehicle communication systems in addition to driver assistance and autonomous driving systems and software that uses maps and satellite positioning to track vehicles. This could include so-called V2X systems, where vehicles communicate with road infrastructure, other equipped vehicles and the cloud.

Will H20 also be banned?

According to domestic media reports, some Chinese manufacturers are unable to place orders for H20 chips. They pointed out that Nvidia has stopped accepting H20 orders since August, but has not officially notified them. Although they are looking forward to the latest news, the possibility of no longer scheduling production is quite high.

Public information shows that Nvidia's H20 chip was launched to comply with US export restrictions and is an AI accelerator designed specifically for the Chinese market. H20 is based on Nvidia's Hopper architecture, has 96GB of HBM3 memory, and provides 4.0TB/s of memory bandwidth. In terms of computing power, H20's FP8 performance is 296 TFLOPS and FP16 performance is 148 TFLOPS. Compared with H100, its GPU core count is reduced by 41% and its performance is reduced by 28%.

It is worth mentioning that in July this year, investment bank Jefferies pointed out in a report to clients that when the United States reviews its semiconductor export policy in October, it may impose a sales ban on Nvidia's H20 products. If implemented, Nvidia is estimated to lose $12 billion in revenue.

Nvidia's move is most likely to respond to the upcoming new regulations of the US Department of Commerce and prevent problems before they happen.

On October 17, 2023, the US Department of Commerce issued new control rules to supplement and update the export controls on advanced computing integrated circuits, semiconductor manufacturing equipment, and items supporting supercomputing applications and end uses issued on October 7, 2022.

The most important part is the change in control parameters. The interim final rule deleted "interconnection bandwidth" as a parameter for identifying restricted chips under ECCN 3A090, and instead changed it to if the chip exceeds one of the two parameters (3A090.a and 3A090.b) marked in ECCN 3A090, the export will be restricted.

According to the documents of the US Department of Commerce, the revised 3A090.a control parameters will control integrated circuits with one or more digital processing units, and the "total processing performance" of these units needs to be 4800 or above, or the "total processing performance" is 1600 or above, and the "performance density" is 5.92 or above. The new ECCN 3A090.b will regulate integrated circuits with one or more digital processing units that have one of the following: a "total processing performance" of 2400 or more but less than 4800, a "performance density" of 1.6 or more but less than 5.92, or a "total processing performance" of 1600 or more, and a "performance density" of 3.2 or more but less than 5.92.

In addition, the rule establishes a license exception that creates a new "License Exception Notice Advanced Computing" for consumer-grade ICs with AI capabilities below the restriction threshold, which applies to two products, one is a chip designed or sold for use in data centers, and the second is a chip not designed or sold for use in data centers, with a "total processing performance" of 4800 or more.

Compared with the rules of October 7, 2022, the new rules have once again expanded the scope of regulation. Under the dual rules of total processing performance and performance density, A800 and H800 are affected, and other NVIDIA products are also affected. The L40 and L40S for the inference market, and the RTX 4090 in the consumer field are also included in the ban.

View more at EASELINK