This is what I think Intel should do.

Intel should seriously start a legit foundry business. Lots of big companies make the mistake of fighting inside the castle wall syndrome. When you are winning, you can't take TSMC's customers. When you are losing, they can take your customers. Just by probability, you will lose in the end.

You might say. Why now? we are behind. First, if Intel remain behind forever, all bets are off and intel fabs are dead. However, if Intel does recover and catch up, you want to exploit your winning to the fullest. Foundry business does take more than 1 day to build. So, double down the bet now is the way to go.

Comment! It's anonymous!

8 replies (most recent on top)

Several factors prohibit Intel from being a foundry business

TMG is incentivised for yield and volume at a much higher scale. They could miss the node advancement , yet collect huge performance bonus for churning out boatload of 14nm+++ products. This causes compromises at one node to get yield , which completely magnifies the issues at next node.

Manufacturing and product design work hand in hand and does not go for volume production until 90pct+ yield is achieved. For foundry customers , getting to the market in time is more important with 40% yield than getting to market year late with 90% yield. Moreover seemless internal feedback loop processes wont be as seemless when documented handovers are needed with 3rd parties.

At Intel, TMG maintains a top dog attitude , and often takes an upper hand with product groups. With BK this problem worsened. In ICF business , the same attitude shows up with customers. Not good for business. You need to have humility to accept that your customers have valid point.

Intel internal processes are riddled with red-tape and committee based vetting. Customers want answer now.

Here's what I think you should do before posting, OP... Do some research and learn some facts first.

Intel already has an external facing foundry business
That business is a joke

If they do this, they had better be willing to put money into updated equipment and software. No more user interfaces that remind me of the days I was just starting to type stuff in elementary school!

Using TSMC node already dead for Intel.

ICF was a joke, and could not accomplish anything. how about ICF 2.0?

legit foundry business. Don't blame customers on too hard to use your node. If intel can use TSMC's node, TSMC customers should be able to use Intel's node. If Intel can't do that, then Intel is behind in more than one way.

Uhh what did you call ICF? Intel foundry sounds like a good idea, but it is too hard to onboard customers onto their node vs. TSMC and also they have to be careful about manufacturing competitor chips.

Hah, Intel's foundry is such a cluster it might as well consider using TSMC. After 14++++++++++ will it take 10++++++++++ to be viable?

Thread regarding Intel Corp. layoffs

This is what I think Intel should do.

8 replies (most recent on top)

Post a reply