Liarfoundries and China know a 14nm fab would be too obvious of a lie - even though their fantasy could be exactly that. 22nm FDSOI is the best alternative to Finfet's for the Chinese, without having the means to get the right tools to build finfet's, that's all.
In the end, China will grab anything it can, to get them one step closer.
I agree that liarfoundries is a complete joke and a mubudala sewer. But a 130 nm fab is going to be much cheaper than a 22 nm fab. I think the person discussing this 10 billion dollars number hit it right on the head. Intel is spending 7 billion on a 7 nm fab. So this 10 billion has to be some totally bogus number for a multi fab complex over multiple years.
A new 180/130 nm fab has to be around 2 billion max. The fab tools cost is completely different from 22 nm FDSOI. You can get the 180/130 nm tools used for cheap.
There should be strict export resitrictions in place to prevernt the 22 nm FDSOI from going to China. The litho tools alone should be restricted into china from both EU and USA. If you don't have the litho tools, you don't have a process.
This is all a fantasy anyway. ASIC starts are decreasintg not increasing. Why do you think there is so much disruption in the semi industry? Mentor Graphics got bought by Siemens. That's because you don't need 3 tool vendors anymore. Core EDA tool biz is shrinking every year due to reduction in ASIC starts.
The industry is moving to fewer players, not more players. Go ahead and build some unused fab in chengdo china. TSMC is the undisputed number one leader in foundry and always will be with over 50% market share. UMC and SMIC pick up all the other accounts that TSMC says no to.
Please show me some real numbers anyway about how much is spent on FDSOI. I'd like to see some real evidence about how much fantastic market traction this unused process really has. What a laugh.
China will go after all the technology it can grab out of Globalfoundries, plain and simple. Neither LiarFoundries nor China will admit what's going on. It depends on what you can build with 130nm, some leading edge technologies are better built on older platforms, at lower cost.
To make FDSOI economical, China needs lots of volume, or a big fab. It is not about how much profit FDSOI delivers today, but what it will deliver in 2 years because it's geared towards low powered applications in the future. But the cost is no issue for China, really.
Even if the whole thing failed, it's a great platform for China to jump into. The IP alone justifies any cost. China will have means for leading edge silicon technology, something they desperately want, and will get; at cost of future of US jobs thanks to LiarFoundries.
I read the developments on the winners and losers of Glofo's deal in China somewhat differently
The fact is that Glofo's initial transfer is 180nm/130nm technology which is >15 yrs old and delivers very low profit margin. That doesn't change much dynamics.
There was mention about FDSOI winning big but how much revenue and profit margin does FD SOI actually delivers today to any foundry? Even if it does deliver , why wouldn't the winner be Samsung who is already ahead on 28nm FDSOI or given TSMC's legendary time to market, why wouldn't TSMC just race ahead on FDSOI or a derivative if the market opportunity presents itself ?
Finally, isn't USD 10 billion a bit excessive for a Fab which is starting with 180nm/130nm technology and later introducing 22nm FD SOI (As a benchmark, INTEL is building their latest advanced 7nm Fab (in expensive USA) for USD 7 billion / UMC Xiamen's 55/40nm Fab is USD 6 billion / TSMC's Nanking 16nm Fab cost USD 3 billion ).
what i found out is: in any corner of the world (google, apple, amazon etc.) it's always people group A fighting group B inside. you have to play by rule to survive. only the losers will whine about job transfer.
Globalfoundries violates many things. It can be done under the covers now that the company is operating as a single entity over multiple countries with most of the information being stored overseas. The flow of information is mostly in one direction, moving outside of this country. It can happen without anyone outside knowing.
It is allowed under the pretext that the company will keep jobs in NY State, in return they were given big government grants and other State support. They do anything that they want, as long as they advertise that they are "keeping jobs in NY State"; their cover up slogan. The real situation is that within a short period of time, the technology that should have never ended up in the hands of a company like Globalfoundries, will soon be handed over to the Chinese and UAE. With most of the talent already gone, unable to create anything meaningful themselves because they can't keep talent, the best that LiarFoundries management can come up with is to ship IP to China for a very low price so the can protect their own paychecks - ala CEO Sanjay Kumar "The Golden Parachute Millionaire" Jha.
It is all very shady to me, if you are truly concerned I suggest you contact the NY State governor Andrew Cuomo.
Nuts. Suicidal. Also I thought there were bans on the export of advanced lithography tools to China and Russia. Last I heard Taiwan doesn't manufacture lithography tools. I mean they don't even let them buy Intel Xeons why should they let them buy the machine tools?
How are the US and EU governments even allowing this deal to happen? SOI AFAIK has applications in rad-hard defense applications, which is already bad enough, but then they compound the problem by allowing 22nm production in China in 2019. A switch from 130 nm to 22nm. WTF? Nuts. I think the best plants China has right now are like 65nm (yeah I know SMIC is supposed to have 45nm but I still haven't heard of any products shipped from there). Even 32 nm would be pushing it but 22 nm is bonkers.