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!
Just curious - what node do you think was Intel's last "good node" for only internal volume? (Assuming Intel also migrated GPUs such as Battlemage to Intel fabs). Were they financially broken already at "10nm/Intel 7"? earlier? later?
The board switching tactics twice in 18 months (during Bob Swan's tenure) definitely hurt the success rate of going all in internal fabrication. The Tower acquisition failure certainly also put a nail in the coffin too.
14 process and product was very good technically. I made fun of the ++++ but they were still way ahead of TSMC and did amazing stuff. The process was way too expensive but it didnt matter since it was so far ahead..... then came 10nm and for the first time in 30 years, intel didnt deliver..... Intel did the finances and said "we cannot afford to do internal development and manufacturing due to our scale and not being OGA" They correctly decided to outsource in the future. Then Pat came along and reversed the plan saying "we will do foundry and be selling billions in wafers in 2025" ... 18A works from what I hear, but it cost too much money and has too little return so they cannot afford to ramp it.
Intel will decide to outsource everything again, it will save the product company and IFS will merge into a foundry like IBM did with similar terms (Intel hired the IBM foundry people).
14 process and product was very good technically. I made fun of the ++++ but they were still way ahead of TSMC and did amazing stuff. The process was way too expensive but it didnt matter since it was so far ahead..... then came 10nm and for the first time in 30 years, intel didnt deliver..... Intel did the finances and said "we cannot afford to do internal development and manufacturing due to our scale and not being OGA" They correctly decided to outsource in the future. Then Pat came along and reversed the plan saying "we will do foundry and be selling billions in wafers in 2025" ... 18A works from what I hear, but it cost too much money and has too little return so they cannot afford to ramp it.
Intel will decide to outsource everything again, it will save the product company and IFS will merge into a foundry like IBM did with similar terms (Intel hired the IBM foundry people).
They are talking with everyone. Seems like TSMC said no. they would be a huge help to GF, UMC, Tower and it could be a JV partnership (AMFAB). Those companies do foundry and IFS can bring advanced tech
I think hiring BK/Swan as CEO was nail in the coffin if some good person has been CEO and they gave the fab enough funding 10nm may not have been botched and it wouldn't be expensive as hell also people say products are good their design PPA Sucks tbh
They are talking with everyone. Seems like TSMC said no. they would be a huge help to GF, UMC, Tower and it could be a JV partnership (AMFAB). Those companies do foundry and IFS can bring advanced tech
they need more customers and those companies need advanced processes. Intel spends more than any other company and the employees still complain that they do not spend enough... so yes, thats a problem that the new owners will need to deal with.
Imo, if 18A was really good and would have appropriate PDKs and IP, Intel would not have problems finding funding and customers. Currently, 18A still seems to be in a state where people don't really know if it works or not (or does it work only on Intel designs). If Intel manages to secure an order from someone like Broadcom or Nvidia to produce 10-20k wpm, I'm sure there will be investors / JV opportunity. Intel desparately needs cash flow and there are many fabless companies that would love to produce chips without TSMC's margin. If Intel expects TSMC margins on 18A then they are stupid and deserve to fail.
I think for margin, it should be viewed relatively. Given TSM US investments, their margin numbers are also depressed. For Intel, assuming they remain IDM, they can stack their margins.
TSMC investment in the US is very much backend loaded. They won't be able to produce chips in volume in US until a few years later. And even for that, we are assuming TSMC actually follows through...
Trump is a showman. A press conference with him, announcing a big round-number investment figure, goes a long way.
they need more customers and those companies need advanced processes. Intel spends more than any other company and the employees still complain that they do not spend enough... so yes, thats a problem that the new owners will need to deal with.
I dont know.... but if I could choose anyone to look into it and lead them to be more cost effective, that person would probably be Naga. Also stealing hiring someone from TSMC would be great. Wei Jen Lo and Kevin Zhang are both former Intel VPs but neither is working in R&D. Group leaders and department managers might be better than VPs. I know some TSMC R&D people who can describe differences
Often the challenge is just a completely different way of doing things. It cannot be adjusted. I worked with Intel and Micron on technology development. everything was different and micron was just twice as fast .... Intel response was typically "that wont work". Nvidia vs Intel product development is same. People who worked at both companies just say "its completely different" ... and of course nvidia is 2x faster.
Personally, I like to see their incremental improvements. They should strictly follow a plan to break even by 2027 for IFS. Doing so will remove the drag on Intel's valuation regardless of any M&A deals.
Personally, I like to see their incremental improvements. They should strictly follow a plan to break even by 2027 for IFS. Doing so will remove the drag on Intel's valuation regardless of any M&A deals.
I dont know.... but if I could choose anyone to look into it and lead them to be more cost effective, that person would probably be Naga. Also stealing hiring someone from TSMC would be great. Wei Jen Lo and Kevin Zhang are both former Intel VPs but neither is working in R&D. Group leaders and department managers might be better than VPs. I know some TSMC R&D people who can describe differences
Often the challenge is just a completely different way of doing things. It cannot be adjusted. I worked with Intel and Micron on technology development. everything was different and micron was just twice as fast .... Intel response was typically "that wont work". Nvidia vs Intel product development is same. People who worked at both companies just say "its completely different" ... and of course nvidia is 2x faster.
Breaking Intel into smaller pieces is a must. Intel is just too big, has too many internal contradictions, and is too complicated to find someone, a superhero, to lead. People often blame Intel's CEOs, Board of Directors, or culture for all the troubles we witness. I believe without changing the structure of Intel organization and business model, any new CEOs or new board of directors will walk on the same path to failures.
You can say I believe it because they are taking active steps to drive costs down like delaying the operation of the first Ohio fab.
It is a responsibility for them to do. They are supposed to deliverer financial stability as the most minimum requirements.
I still do not understand they could let PG drive the company to such a state. Completely unnecessary IMO. I think partly US government played a role here by enticing the CHIPS Act but without concrete financial supports and relevant policies. But I think PG should bear most of the responsibility here. But there are good things coming out from his efforts, like the abilities to provide foundry services, which take time to get to this state, i.e. Intel 12 with UMC, bringing external IPs to IFS, etc.
You can say I believe it because they are taking active steps to drive costs down like delaying the operation of the first Ohio fab.
It is a responsibility for them to do. They are supposed to deliverer financial stability as the most minimum requirements.
I still do not understand they could let PG drive the company to such a state. Completely unnecessary IMO. I think partly US government played a role here by enticing the CHIPS Act but without concrete financial supports and relevant policies. But I think PG should bear most of the responsibility here. But there are good things coming out from his efforts, like the abilities to provide foundry services, which take time to get to this state, i.e. Intel 12 with UMC, bringing external IPs to IFS, etc.
Personally, I like to see their incremental improvements. They should strictly follow a plan to break even by 2027 for IFS. Doing so will remove the drag on Intel's valuation regardless of any M&A deals.
I think Pat mentioned betting everything on 18A? Or going all-in? I believe that to be true now. Intel 18A must succeed in the foundry business. I don't think Intel will get another chance. This is a must win situation for the manufacturing side.
Let's wait until the Intel Foundry event next month. I'm hoping for some really good news!
