Chinese semiconductor industry

Status
Not open for further replies.
L

latenlazy

Brigadier
FairAndUnbiased said:
The vast amount of used and refurbished KrF and dry ArF equipment from the 90s and 2000s also doesn't help. For low cost low productivity equipment you have to compete with refurbished equipment too.

The other thing is that you have to match the lowest WPH rate for any equipment in the fab. So it doesn't make sense to have a fast photolithography system if your resist developed capability or other capabilities can't catch up. Wafers in process are liabilities so you either have to throttle your good equipment, buy all best equipment, or buy more equipment.

This makes it very expensive to keep at the leading edge because you have to upgrade or expand many things at once, not pick and choose. So it seems to naturally lead to monopolies.
Click to expand...
tokenanalyst said:
ASML has become like the Nvidia of the WFE world, fabs just prefer ASML, on spec Nikon scanners don't look that unproductive, its look like they can pattern the same wafer per hour than ASML or even more, but to archive that looks like they are using pretty hefty tricks that are coming at the expense of easy of use and seamless integration, add to that the culture of ASML of closely working with fabs and semiconductor industries in different countries, just make things more difficult for Nikon.

Another thing is that productivity is also relative and complicated, lets suppose an imaginary fab needs to produce 100 WPH to satisfy their costumers orders, an 275 WPH machine looks like an overkill and overprice for them, that machine will set idle half of the time, but what if they could buy a machine that can do 150 WPH at half of the price of the 275 WPH? That machine will just fit right to them, if the lower priced machine come with a very steep learning curve, is difficult to use and can't be integrated easily, then yes go with the expensive one. but apples to apples the 150 WPH fit better the 100 WPH fab than the 250 WPH.

The monopolization of the semiconductor industry has left no middle ground. i think the fabless model become popular no just because the difficulty of fabbing on smaller process nodes but also because the rat race towards never ever increasing productivity of big players left smaller players without tools that fit their needs at the right price. I massively oversimplifying this, but you get my point.​
Click to expand...
One of the major parts of the business that AMSL really cracked was the service model, because when you do inevitably have some instrument downtime, spending more on a service contract that helps get your operations back up much faster helps you reduce the cost of the downtime, which more than offsets the cost of the service support. And then once you have that service relationship you’ve folded yourself into the operations of a client and you have an extra bit of stickiness against competitors. But I think it’s also worth recognizing that the service model became a competitive edge in this space mainly because 1) the pace of key performance innovations slowed dramatically after the industry hit a wall in basic technology with ArFi, 2) as demand for chips boomed, and as cost of production increased at each new node, and especially as production developed greater reliance on multipatterning and thus increased utilization of lithographs, that magnified the opportunity costs of operations downtime for fabs, which then increased the value of good service models.
 
A

ansy1968

Brigadier
Registered Member
antiterror13 said:
woww, thats the most advanced DUV from ASML
Click to expand...
@antiterror13 Sir from @olalavn himself a 22nm DUVL in cooperation with Huawei.


olalavn​

New Member​

Registered Member
ansy1968 said:
@olalavn bro thank you and since we're on topic, do you have any info regarding the mythical 22nm DUVL? are SMEE or CETC doing some work on it or it's just a speculation on my part?
Click to expand...
SMEE is doing it with experts from Hsilicon
 
Status
Not open for further replies.
Top