Chinese semiconductor thread II
- Thread starter vincent
- Start date
In 2015 the Chinese government goal for 2020-2030 was to have:
2020:
250 nm I-Line machine. Wavelength = 350nm 0.8NA
110 nm KrF machine. Wavelength = 248nm 0.8NA
90nm ArF machine. Wavelength = 193nm 0.8NA
2025:
45nm Immersion machine. Wavelength = 193nm 1.35NA
And for 2030:
32nm EUV machine. Wavelength = 13.5nm 0.30NA
Not 7nm no 5nm no 3nm, that goal was to make chips in a planar process closer to the resolution of the machine because at the time,2015, overlay was still an issue. So for 250nm the overlay is 80nm, 110nm is 35nm, for 90nm is close to 30, for 45nm is 15nm and for 32nm is 10nm. Single exposures and big overlays.
My thinking is that the goal was not to replace foreign machines but to have machines that could make chips that are critical for the government, the space program and the military. That was the self-reliance that government wanted in with MIC2025, is to have 70% of the chips that the government use being produced locally, that include chips for the military, reducing the dependency of critical components that could be blocked. But US stooges in their infinite stupidity miss-interpreted this as China wanting to replace every single chip in China with domestic one and from there with the narrative of bellicose overpaid US think tankers everything when downhill with the situation escalating significantly under Biden in 2022.
I do think was bit foolish for the Chinese government not keeping their MIC2025 plans secret knowing that US clowns would misinterpreted it BUT at the same time I think that with MIC2025, the Special Project 02 and other projects China may have avoided a US made bullet.
2020:
250 nm I-Line machine. Wavelength = 350nm 0.8NA
110 nm KrF machine. Wavelength = 248nm 0.8NA
90nm ArF machine. Wavelength = 193nm 0.8NA
2025:
45nm Immersion machine. Wavelength = 193nm 1.35NA
And for 2030:
32nm EUV machine. Wavelength = 13.5nm 0.30NA
Not 7nm no 5nm no 3nm, that goal was to make chips in a planar process closer to the resolution of the machine because at the time,2015, overlay was still an issue. So for 250nm the overlay is 80nm, 110nm is 35nm, for 90nm is close to 30, for 45nm is 15nm and for 32nm is 10nm. Single exposures and big overlays.
My thinking is that the goal was not to replace foreign machines but to have machines that could make chips that are critical for the government, the space program and the military. That was the self-reliance that government wanted in with MIC2025, is to have 70% of the chips that the government use being produced locally, that include chips for the military, reducing the dependency of critical components that could be blocked. But US stooges in their infinite stupidity miss-interpreted this as China wanting to replace every single chip in China with domestic one and from there with the narrative of bellicose overpaid US think tankers everything when downhill with the situation escalating significantly under Biden in 2022.
I do think was bit foolish for the Chinese government not keeping their MIC2025 plans secret knowing that US clowns would misinterpreted it BUT at the same time I think that with MIC2025, the Special Project 02 and other projects China may have avoided a US made bullet.
I mean, it's probably not something that was predicted, at least not with certainty, but US re-actions to MIC2025 was very much a net positive to its completion.
Chinese government can do a lot of things, but one of the hardest thing to do even for China is influencing public behaviour and overcome network effects. Without America's open hostility it would have been very hard to get consumers to buy domestic, and without US giving up on entrenched marketshare it would have taken much longer for Chinese companies to gain a foothold.
e.g. China invested in EVs because EVs are sufficiently differentiated and superior to ICEs to overcome entrenched western ICE market dominance, but if there were no EVs and China has to compete on ICE, it would have been much harder without somehow forcing errors from western ICE carmakerse.
The original 2015 semiconductor plan was equivilent to an auto industry that tries to compete with the west on ICE, the oppertunity for signfiicant acceleration only really became aviliable because of US reaction to MIC2015.
this is almost two years old research .. interesting
This might also have been originally protectionist in nature, with the military/intelligence justification added later. The sanctions on Huawei originally came at a time when the most performant ARM designs outside of Apple from of HiSilicon (Qualcomm was struggling, especially on the modem front. It wasn't until 2 years after the Kirin 990 that they introduced an integrated 5G modem, and another year until it was actually good), on the enterprise front they were eating into the sale of more established players like Cisco and HPE, and they were winning contracts for 5G roll-out across the world.
Obviously no one here has the actual justification to why the US levelled these sanctions, but I believe even without the published MIC2025 plans the US would not stand to see their core industries out-competed by foreign rivals (just look at what happened Bombardier CSeries).
TCL Huaxing has their own 8.6 Gen OLED project signed in Guangzhou and will produce 22.5k wpm when complete.
AMAT got in big trouble in 2023 for selling Hydrogen Ion Implanters to SMIC through its SK subsidiary . Looks like that won't happen again.
Huahai Qingke: Ion implantation equipment has been delivered in batches to several leading domestic integrated circuit manufacturing companies
On October 21, Huahai Qingke announced that its wholly-owned subsidiary, Xinyi Semiconductor (Shanghai) Co., Ltd., has delivered a 12-inch, high-beam ion implanter, independently developed, to several leading domestic integrated circuit manufacturers. This marks the official acceleration of Huahai Qingke's ion implantation equipment's large-scale application.
The company said that this batch delivery not only reflects customers' high recognition of Huahai Qingke's product performance and R&D capabilities, but also demonstrates the company's technical strength and strategic determination in the process of localizing high-end ion implantation equipment.
Since its initial investment in ion implantation equipment, Huahai Qingke has made rapid progress and continuously improved its product portfolio. From its high-beam ion implanter series to the rapidly expanding high-energy ion implanter series, the company has established a comprehensive product portfolio covering the process needs of multiple fields, including logic chips, memory chips, power semiconductors, CIS (image sensors), and silicon wafer manufacturing.
In terms of product performance, Huahai Qingke's latest generation of high-beam ion implanters has reached the same level as leading international competitors in terms of replacement rate, stability, and process compatibility, meeting the stringent requirements of advanced processes for ion implant uniformity and accuracy. Currently, this series of equipment has successfully passed verification at multiple clients and has received repeat orders due to its excellent performance.
The company continues to promote the "equipment + service" platform strategy. Its product line has covered key equipment areas such as CMP, ion implantation, thinning, dicing, edge polishing, and wet processing, and provides supporting wafer regeneration and consumables maintenance services, building a complete process solution system.
Huahai Qingke stated that it will continue to promote technological innovation and R&D investment in the future, provide higher-end and more advanced equipment and process integration solutions for the semiconductor industry, and work with industry chain partners to promote the construction of an independent and controllable semiconductor industry ecosystem, and help the industry develop with high quality.
Last edited:
Muxi Co., Ltd. releases the domestically produced general-purpose GPU Xiyun C600
Recently, Pukou District announced on its official Weibo account that Muxi Co., Ltd. has officially released the first domestically produced general-purpose GPU, the Xiyun C600, marking a significant breakthrough in computing power. The press conference was reported by the Pukou District's officially authorized "Pukou Release" account. At the conference, Muxi Co., Ltd. Senior Vice President Sun Guoliang noted that as computing power becomes the "hard currency" of the digital economy, competition in general-purpose GPU chips will become a key issue for countries.
According to reports, Xiyun C60 is a fully localized chip that integrates large-capacity storage and multi-precision mixed computing power, and supports MetaXLink super-node expansion technology. It fully meets the training and inference requirements of the next generation of generative AI in terms of hardware performance and software compatibility.
Leveraging its expertise in GPU chips and understanding of application scenarios, Muxi has independently designed dozens of core IPs. Manufacturing, packaging, and testing are all handled by domestic companies. Muxi has established a "1+6+X" strategic strategy for applications. Leveraging its proprietary GPUs, Muxi aims to deeply empower six key industries: finance, healthcare, energy, education and research, transportation, and entertainment, while also promoting the implementation of X emerging scenarios.