You are currently viewing SemiWiki as a guest which gives you limited access to the site. To view blog comments and experience other SemiWiki features you must be a registered member. Registration is fast, simple, and absolutely free so please, join our community today!
Do not say I did not tell you so. Maybe they will have more luck fabbing leading edge CPU and GPUs from AMD and NVIDIA. Turns out the desktop version of AMD Zen 5 is still using N4X. The N3E is only used in the Zen 5c server chips.
Samsung got infected with IBM disease. They did not use to be like this chasing the bleeding edge in the old days. Back when IBM, AMD, and Samsung were in the design alliance, typically IBM had the most bleeding edge process, AMD was more conservative, and Samsung was the most design...
It is kind of lame but it is progress I guess. They used to be stuck at 90 nm.
That 65 nm machine is probably not from SMEE. Since it is not on their website at all.
Old processes without EUV are still around 50% of TSMC revenues.
At the low end they will face competition from China. While at the higher end they will be facing competition from Intel.
At least that is the theory. We will see in 2 years time if the Chinese and Intel fab expansions go...
Simply not true. Huawei's HiSilicon develops its own ARM core designs just like Apple. The Taishan core family for example is used as the big cores in the Kirin 9000S. They also develop their own GPUs named Maleoon. Also used in the Kirin 9000S. The big Taishan CPU cores are also used in their...
Had the US waited to apply the sanctions until it had an actual conflict with China it would have been much more effective. In this way China will have its own counter-measures in place by the time a conflict does happen.
It is also one thing to ban supply of tools you had not supplied before...
This is a consequence of Rock's law.
https://en.wikipedia.org/wiki/Moore%27s_second_law
Fab construction costs double every two process nodes. You will see them double again when High-NA comes out.
The amount of businesses with enough capital to build leading edge fabs naturally tends to...
SMIC could have redeployed the immersion lithography machines they already had at their other facilities to make the simpler layers. And used the leading edge machines at their FinFET fab to make the harder layers. The import ban was only applied after the fab expansion shell was complete as...
Exactly. He worked for several years in the solar cell business after leaving SK Hynix in 2010, and then he moved back into DRAM in China in 2015. Even people who sign confidentiality clauses typically have a time limit. And in his case he resigned from SK Hynix apparently. Probably got zero...
https://www.reuters.com/technology/indictment-details-plan-steal-samsung-secrets-foxconn-china-project-2023-06-27/
https://www.ft.com/content/fc7f6ea0-08f6-40f3-897f-e723cff9fd8c
For what it's worth 20nm is not considered leading edge and Chinese companies like CXMT have better DRAM technology...
Rick Tsai goofed up by messing up the next process transition and focusing on whale clients while dismissing smaller clients. That is why Morris had to come back.
Intel sold a 49% stake on those factories. They are keeping 51% of the capital and control of the fabs. So it is not as bad as you are talking about. Intel will then use this money from the sale to complete their new fabs in the US. The plan makes perfect sense in my opinion.
This is IMHO one of the main failures by Intel. Altera still makes all their FPGAs at TSMC. They released like a single product with an Intel process then gave it up. I hope Intel management will fix this long term. They keep wanting clients for IFS and here is one in-house and they do not bother.
CUDA is used for more than just AI inference. The available software base is immense and hard to compete against.
I expect the AI neural networks market long term to evolve towards specific NPUs.
SMIC doubled the size of its FinFET production facilities. We are supposed to believe there are chip supply issues. I doubt it.
There are claims of delays with Huawei's new operating system (HarmonyOS NEXT) and application support. Their new OS is going to remove Android backwards...
The biggest weakness of Intel and AMD in GPUs is the software base. It is much weaker than NVIDIA's. And NVIDIA has been accumulating software for over a decade at this point.
StrongARM ended up landing on Intel's lap as part of the DEC lawsuit. They never particularly wanted it. But it had...
