Array
(
    [content] => 
    [params] => Array
        (
            [0] => /forum/index.php?threads/us-govt-says-banned-chipmaking-equipment-still-ends-up-in-china.19154/
        )

    [addOns] => Array
        (
            [DL6/MLTP] => 13
            [Hampel/TimeZoneDebug] => 1000070
            [SV/ChangePostDate] => 2010200
            [SemiWiki/Newsletter] => 1000010
            [SemiWiki/WPMenu] => 1000010
            [SemiWiki/XPressExtend] => 1000010
            [ThemeHouse/XLink] => 1000970
            [ThemeHouse/XPress] => 1010570
            [XF] => 2021370
            [XFI] => 1050270
        )

    [wordpress] => /var/www/html
)

US Govt Says Banned Chipmaking Equipment Still Ends Up in China

Daniel Nenni

Admin
Staff member
 SMIC.


Although the U.S. government has imposed extensive restrictions against the Chinese high-tech industry in a bid to prevent the latter from obtaining advanced chips and wafer fab equipment (WFE) made by American companies, Chinese entities can still get what they need using various loopholes, a report from the U.S.-China Economic and Security Review Commission has revealed. SMIC's high-volume production of 7nm chips illustrates that Chinese companies can definitely overcome U.S. sanctions.
The measures imposed by the U.S. administration in Oct. 2023 have notably impacted China's artificial intelligence, high-performance computing, and semiconductor sectors, in general — but loopholes and China's adaptive strategies have undermined their overall effectiveness, reports The Register.

The China export rules imposed last year require U.S. companies and individuals to secure licenses for selling equipment and technologies used in producing non-planar transistor logic chips on 14nm/16nm nodes and smaller, 3D NAND with 128 layers or more, and DRAM memory chips with a half-pitch of 18nm or less. These regulations also extend to foreign firms exporting U.S.-originated components.

However, tools that are used to make 28nm chips can be used in production of 5nm ICs and neither makers of WFE nor the U.S. government can control how these tools are used.

"With BIS using a 14nm restriction limit, importers are often able to purchase the equipment if they claim it is being used on an older production line, and with limited capacity for end-use inspections it is difficult to verify the equipment is not being used to produce more advanced chips," the report reads.

The report outlines a number of significant enforcement challenges and loopholes. For example, it shows how Chinese chipmakers are acquiring tools for producing chips that are only marginally behind the target technology nodes, bypassing the intent of the bans. They can still equip their fabs with fairly advanced tools and produce chips that are close to those being regulated.

Another way Chinese companies have managed to continue producing advanced chips is by procuring tools from countries other than the U.S., before sales were restricted. Chinese companies were able to purchase advanced tools from Dutch and Japanese companies for about a year after the U.S. first imposed restrictions.

Perhaps the best example of how Chinese companies have been evading U.S. sanctions is Semiconductor Manufacturing International Co.'s high-volume production of the advanced Huawei HiSilicon Kirin 9000S application processor using its 2nd Generation 7nm process technology. The company can apparently produce a boatload of such chips using tools that it obtained before it was blacklisted and WFE that it managed to acquire after it was blacklisted (using various loopholes).

 
The current US Gov sanctions are just all talk. All carrots and no stick. Every company under the sun is trying various ways imaginable or execues no matter how lame to skirt around the specificities in the regulations and having no fear if their actions seem obviously violating the intent of the regulations. Look at Nvidia, for the best example.
 
With BIS using a 14nm restriction limit, importers are often able to purchase the equipment if they claim it is being used on an older production line, and with limited capacity for end-use inspections it is difficult to verify the equipment is not being used to produce more advanced chips,
When the machine is being serviced by the manufacture wouldn't it be pretty obvious that it's being used on an advanced node?
 
When the machine is being serviced by the manufacture wouldn't it be pretty obvious that it's being used on an advanced node?
The machine setup is pretty much the same, whether it's running 40nm or 28 or 14 or even 7nm, for vast majority of equipment sets in a fab, perhaps the only exception maybe litho scanners because the masks will be different for different nodes.

If fab does not share what devices are being made, there is no way for a service engineer to know.
 
The machine setup is pretty much the same, whether it's running 40nm or 28 or 14 or even 7nm, for vast majority of equipment sets in a fab, perhaps the only exception maybe litho scanners because the masks will be different for different nodes.

If fab does not share what devices are being made, there is no way for a service engineer to know.
But when they see the giant SMIC sign on the building that's a bit of a give away.
 
I don't think so. There are certain features that would be overkill if you were using the equipment for 40nm. Some tools also just don't have the process capability for advanced nodes for example good luck trying to do 200L+ 3D-NAND on an etch tool from 2010 or even a more modern tool without all of the fancy add-ins that vendors shell out. Also vendor process engineers would be able to see clear as day what you are doing with their tools.
 
The machine setup is pretty much the same, whether it's running 40nm or 28 or 14 or even 7nm, for vast majority of equipment sets in a fab, perhaps the only exception maybe litho scanners because the masks will be different for different nodes.

If fab does not share what devices are being made, there is no way for a service engineer to know.
The machine setup for 40nm, 28nm, 14nm and 7nm is NOT pretty much the same. But if you classify them as litho, etch, film dep, implant, epi, clean, metrology, it is pretty much the same. The following is the tool re-use example from 20nm(planar) to 16nm(FinFET). Both shared the same backend D/R and scheme. the only difference is transistor architecture. FYI.
"It will now do tool reuse to convert more to 16nm, citing that it has 95% equipment reuse. The company can save a few hundred million on capex through the reuse and conversion of 20nm capacity to 16nm."
from: https://www.barrons.com/articles/ts...-month-at-credit-suisse-conference-1427328336
 
Back
Top