6 ms·
Intel need to split into two. 1. Intel fabless: open up x86 like ARM. work on Xscale (ARM) and RISC-V (SiFive). 2. Intel foundry: a pure play foundry with no
by MangoCoffee 5y ago
Intel need to split into two.
1. Intel fabless: open up x86 like ARM. work on Xscale (ARM) and RISC-V (SiFive).
2. Intel foundry: a pure play foundry with no tie to its fabless side to win customers trust.
the foundry ecosystem is going to hurt Intel. the current trend is everyone design its own chip and outsource to foundry for production.
edit: look like the deal to buy SiFive fall apart but the idea is Intel fabless can focus on chip design regardless the platform.
- jjoonathan 5y agoThis kills the foundry.
- ngomez 5y agoCould you elaborate? AMD spun their fabrication arm off into GlobalFoundries a while back, and they seem to be doing okay. I thought that would be a good precedent for Intel divesting its fabs too, but I'd love to hear any counterpoints to that.
- totalZero 5y agoAMD spun off GlobalFoundries to save itself, and suffered greatly for many years afterward. Let's not pretend it was some kind of masterstroke done by a company that was firing smoothly on all cylinders. Do a Google search for the stock price. AMD announced the spinoff in October 2008.
- jjoonathan 5y agoGlobalFoundries completely gave up on leading edge and went downmarket. It would be a shame to see Intel's fabs follow.
- throwawaysea 5y agoWhat you're suggesting is what AMD did when they spun off Global Foundries, correct? https://en.wikipedia.org/wiki/GlobalFoundries https://en.wikipedia.org/wiki/GlobalFoundries What would make Intel more successful at operating a foundry than anyone else? Or is it just a way to make the best of assets they already own?
- MangoCoffee 5y ago>What would make Intel more successful at operating a foundry than anyone else? Intel is very successful with their 14nm, with a engineer CEO like Pat. I think Intel can catch up on node process.