Hacker News Reader

Comment by etempleton | Hacker News Reader

I think if they can deliver 18a in 2025 and it is comparable to the competition they are going to be in a much better position. Intel dominated because their fabs were better. At some point they got complacent and stopped investing. Pat changed that and hopefully it continues. If Intel wants to succeed it has to have a comparable product at a comparable price to TSMC. There is more demand for leading edge chips than supply. Plenty of market share to go around if Intel can figure it out.

View on HN

>Intel dominated because their fabs were better. At some point they got complacent and stopped investing. Pat changed that and hopefully it continues.

I thought the board wanted to split the company in two, while Pat wanted to invest in the fabs. Now that he's gone, doesn't that mean the board will find someone who agrees with them about spinning off the fab business?

>doesn't that mean the board will find someone who agrees with them about spinning off the fab business?

I dont think that is even possible. Given the money they received from CHIP Act. They could spin off x86 business, but may be not the Fab. My guess is that The board actually want Intel to be itself back in the good old days, leading edge node and selling CPU at a premium. Hence they want to scale down their scale of investment on Fabs and possibly focusing capacity for itself only.

Obviously the board doesn't know much about the current Semi market.

>The board actually want Intel to be itself back in the good old days, leading edge node and selling CPU at a premium

Those days are gone and never coming back. How do they get people this clueless on the board anyway?

My guess is that is what those Fund Managers wants. They only invested in Intel for its dividends.

I think this is probably accurate. I think the board is struggling with the amount of financial outlay and the weaker than expected returns. I don’t think Pat or the board thought things would get this bad before they got better. I do think they could have managed public sentiment a bit better. While they have been struggling it would have been smart to have an enthusiast gaming cpu that outclassed AMD even if it they lost money on it. Most of the tech sites that review cpus are enthusiast oriented. I don’t believe things are as dire as they look for Intel.

For national security reasons, the fab side might be too important to fail, and the US government will continue to subsidize them in being 2 years behind TSMC. The x86 business is in a weirder place because switching to the latest TSMC node would be an immediate improvement and its problems are more tractable, but it's less strategically important (another domestic vendor with near drop-in replacements exists), and both datacenter and PC customers are seriously looking at other options.

Unless tame is a us based fab and a U.S.A. company even. Then … but probably before that Taiwan war started. Hence still have to have U.S.A. base in the next few years.

I think it is conjecture that is what the board wants. I myself have mused on here that is what they want, but I honestly think it would be a huge mistake, especially at this point. I think they are too far invested in this strategy to abandon it now.

>I honestly think it would be a huge mistake, especially at this point. I think they are too far invested in this strategy to abandon it now.

Don't assume any great intelligence on the part of these board members. If all these company-running people were so brilliant, Intel wouldn't be in this mess in the first place.

I think you're right, but

> if they can deliver 18a in 2025

What on earth makes you believe this is a possibility given their last 10 years?

For Intel's sake, I hope I eat crow. But fool me once, shame on you...

Intel made a strategic error in adopting EUV later than Samsung and TSMC, but they've always had good fabs. All the (non-fake) news about 18a seems to point to them being able to deliver in 2025.

This is the biggest thing that gives me some hope. The issue was they were trying to execute on old technology that was never going to scale smaller. This was a huge strategic blunder by leadership, but does not mean that the rank and file cannot execute when given the proper tools.

>What on earth makes you believe this is a possibility given their last 10 years?

You mean the 10 years prior to Pat? Since his arrival he plan of 5 node in 4 years has been executing well so far. The last node in that plan is 18A.

Executing so well that they’re fabbing basically all of their current products at their biggest competitor TSMC instead of their own fabs? That doesn’t really scream successful execution to me…

As you said, 18A is the last node of that plan, so by now if they were executing well on those nodes they should be fabbing all their products on the Intel 3 and 20A processes but they’re not. In fact they don’t even have a 20A process to use because they cancelled it, so clearly that wasn’t executed well.

And that is the big IF. Depending on who you believe 18a is either doomed or right on track. I think if they can launch shipping products that consumers can actually buy on 18a in 2025 it is a huge win. If not, well, Intel is for a lot more pain.

Reply on news.ycombinator.com