zlacker

You don't think $8.9B would do it?

>>andrew+(OP)
TSMC has already put $65B into the Phoenix fab and is adding at least that much more, so no. You're off by an order of magnitude.

>>andrew+(OP)
TSMC's estimated costs in 2020, were $12 billion for their first fab. In 2025, their updated estimates were $65 billion for the first three fabs and $165 billion for when they get to six such facilities. So, $8.9B is a lot of money, but isn't anywhere close to getting to the equivalent to what TSMC has in Taiwan.

>>rchian+y5
> getting to the equivalent to what TSMC has in Taiwan

That wasn't the question. The question, at least for me, is can you build non-zero chip production, enough to start building out a sustainable business. Obviously you're not going to compete with TSMC on day one, but there's a wide spectrum between that and "garage".

>>andrew+(OP)
This link contains a graph of fab costs over time. It looks like 9 billion might get you a cutting edge fab 15-20 years ago. but that's just the fab.

https://semiwiki.com/forum/threads/how-to-build-a-20-billion...

>>andrew+T5
China tried to do it, and they aren’t even close despite their massive state subsidy programs, so no.

>>andrew+T5
How would you build a sustainable business based on old processes though? The only reason fabs exist that use old processes is because they were once new processes, and once they've been built you may as well keep them running for a while. Building a new 50 nm fab would never be viable.

>>IshKeb+Cx
Those old fabs are still able to be useful at all because most applications don't need cutting edge chips. Chips have been Good Enough for decades. And again, if the goal actually is manufacturing independence, buying local chips that are a bit more expensive is totally worth it.