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!
New Fab customers are 2 years out best case for meaningful revenue. How can a customer commit to Intel with all this uncertainty?
When you look at re-use vs greenfield, you often find reuse isnt as cheap as hoped (There are lots of details why). reusing old fabs is not option (maybe 42?)
Best...
so the options for customers are:
1) go with the best company who delivers and has a proven history but is located in a foreign country IF people want.
2) use US company, who has no history of foundry work because they are headquartered in America.
Do we want to allow people to make a choice...
Speaking of Japan. The Rapidus pilot line is up and running 2nm now, correct? Have they shown pictures of the first wafer out of their single wafer processing, low cycle time, IBM 2nm process in Japan
Ireland has been up a running for a while (a year) but the volumes have been low. Maybe this is an announcement of ramping Granite Rapids and Sierra Forest. As I have mentioned before, the ramp of datacenter CPUs is much slower than most people realise. Ireland needs to ramp a lot to be successful.
If only someone had predicted that all of the wild Fab expansion dreams were not going to become reality
In other related news, NOAA said it has measured a increase in March temperature over that in February and January of this year .... but only in the northern hemisphere.
They attribute the...
We looked at it when I was in memory fabs. The theory was that it is a good reuse for memory equipment (mainly lithography). But like everything, you have to be good at it and efficient at it....
TSMC Fab delays were blamed on various issues. That doesnt mean it was the cause.... it might have been a good excuse. Micron and TI have been successful in fast builds.
Scotten's numbers are spot on. If it turns out that the equipment utilization is much worse than in TW, then TSMC costs will...
The faster the growth, the bigger the impact when it levels off. In general, there will be a 1 quarter warning where everyone says "this is just a short term adjustment" and "We see recovery next quarter". and "fundamentals are still good"
One quarter later it will be a full correction and...
this is great. There is a myth (that is very easy to disprove if you do the math) that TSMC makes all its money on old mature nodes.
Also of note: In 2024, I do not believe Nvidia or AMD had any N3 products. The are never on leading node. (correct me if I am wrong).
If we think Intel and TSMC are in a race, lets track actual foundry revenue from 18A vs N2 in then next 4 quarters. Predictions and models today are not worth arguing about. Actuals matter. I think Vegas is taking bets on revenue and margins for the two processes :ROFLMAO::LOL:
Get your popcorn. Its about to get real.
Be sure to ask every 10 minutes "Is this an April Fools Joke???"
Or maybe the whole meeting is a April fools joke?
