3 ms·
Question: why is it that Intel can’t just poach a bunch of the best TSMC engineers and catch up that way? Is it a lack of knowledge how to do this manufacturin
by jonplackett 4y ago
Question: why is it that Intel can’t just poach a bunch of the best TSMC engineers and catch up that way?
Is it a lack of knowledge how to do this manufacturing? Or is it more like a skills shortage of engineers to do the actual manufacturing?
Where exactly is it in the process that they can’t compete, and why is it so hard to catch up, given their virtually infinite R&D budget.
- ksec 4y agoThese questions tends to come from software side, especially web developers where deployment and distribution lead time is practically zero. Even if Intel knew, the exact method to catch up, the time required to build Fabs, even if it was perfectly planned both internally ( as in Staff, Operation, recruitment ) and externally, everything from Land, Water, Electricity supply, equipment all readily available would still take 8 months to build and 1 year to operation. That is the fastest record to date to bring up a Fab by Apple and TSMC with lots of help from Government ( in terms of Regulation and infrastructure approval ). And if you include the planning prior to it that is more like 2 years. And it is already consider an astonishing achievement because most other players ( including TSMC themselves ) need 3-5 years. And right now, ASML has a backlog of orders to fill, assuming they cant increase their rate of production both due to the complexity and chip shortage, they are talking about 2024 if not 2025 orders. That is ignoring other part of the supply chain like Tokyo Electron which are also running extremely tight. ( Or Applied Materials if you want to look it up ) Again, all of that is assuming you knew exactly what to do to catch up to leading node.
- jonplackett 4y agoHey thanks for the reply (and yeah, software dev here!) Follow up question: does anyone know what it is they don’t know? Because while that does sound like a long time, they’ve been faffing around for years now trying to make 10nm work. Even on that time scale they should have been able to build a new fab - intel should really know how to do that anyway. If there is something they don’t know, why can’t they just buy it in. That’s the thing I don’t get. They have so much money still.