I'm sorry but comparing Intel to IBM is quite simplistic and naive. Yes, Intel has stumbled badly and there is a lot of work to dig out of the hole. But, unlike IBM, Intel has always been number 1 or number 2 in semiconductor fabrication and their CPU design teams are still quite capable but were hobbled badly by the 10nm missteps. IBM in comparison was never top 1 or 2 in in PC sales for that long. Their decline was more pronounced.
Saying that AMD runs circles around Intel CPU designs is hyperbole. Yes, on the server-side Intel has been knee capped quite a bit due to their process deficit, lack of disaggregation and validation issues. But their core CPU microarchitecture designs are fundamentally not something that AMD runs circles around. AMD might be somewhat better overall for now and wins in some, Intel in others and we shall see what happens once they are on TSMC 3nm and next gen server CPUs come out next year. Intel's core problem right now is fixing the process deficit.
A bit disappointed in this shallow take. I suppose it has become fashionable to show extreme negativity towards Intel and count them out. That's good for those of us who see the risk/reward as attractive and can see the possibility of rays of sunshine that could lead to a turnaround.
Full Disclosure: I own $INTC, $AMD and $TSM (for different reasons and bought at different times).
I’ve been about as supportive as I could be on Intel 18A.
And man, I don’t know what to say but it’s like you didn’t read it. IBM had a fab back in the day. Nor do I think AMD is exactly in a great spot either, who needs a CPU? Seriously? client is making not that much.
I literally think that Intel is in a worse spot than IBM, and if you disagree with me sure, but there’s almost no hope for commodity design (Intel CPU) and its foundry or bust, which is an objective worse business than the Intel IDM juggernaut of 1990s-2010s.
Sorry if this didn’t fit your priors, but right now 18A is fledgling. Even if they pull it off, we are talking about a business that won’t make that much money, at least compared to the Intel of old. Disagree.
I did read the article. From a business point of view, yes Intel did miss some platform shifts. But unlike IBM which was never number 1 or 2 in PC sales or semi fabrication for long, Intel has not been that far behind technology wise other than the 10nm stumble. Their problem has been execution, management and strategy.
A high-performance CPU design is not a commodity. There are only a handful of teams that can do that. And CPUs are needed ultimately to do many things in computing, so that business is not going away.
Assuming they pull it off, even with slightly lower margins as a fab (compared to the IDM model) doesn't mean they won't make money. If the world is going to need a lot of fab capacity to build all these AI accelerators in the future, then getting a big enough chunk of that volume will suffice.
I'm not saying Intel will easily accomplish this, but they aren't quite dead yet. We shall see what happens.
Dude it is a commodity. Why is it that everyone can have their own that is +/- 10% of the best one possible w/ a few fees to ARM.
I think that the problem is that 18A rules, foundry matters, but it's really hard to see the way out for design unless they start cutting hard yesterday. RIF should be 40%.... then we are cooking.
Yes Lip-Bu Tan leaving is big negative I agree. I have changed my views. See my twitter post about that.
But you are being delusional if you think everyone else is within 10% by just paying fees to ARM. $QCOM only got there because they bought the ex $AAPL crack design team via the Nuvia acquisition and they still can't quite match Apple. Yes, the gap has narrowed but it's not that easy. Most of the ARM cloud CPUs so far are just scale out and not high performance. BTW even $NVDA tried mightily in this space and failed. Look up the history of Stexar and Project Denver.
I'm sorry but comparing Intel to IBM is quite simplistic and naive. Yes, Intel has stumbled badly and there is a lot of work to dig out of the hole. But, unlike IBM, Intel has always been number 1 or number 2 in semiconductor fabrication and their CPU design teams are still quite capable but were hobbled badly by the 10nm missteps. IBM in comparison was never top 1 or 2 in in PC sales for that long. Their decline was more pronounced.
Saying that AMD runs circles around Intel CPU designs is hyperbole. Yes, on the server-side Intel has been knee capped quite a bit due to their process deficit, lack of disaggregation and validation issues. But their core CPU microarchitecture designs are fundamentally not something that AMD runs circles around. AMD might be somewhat better overall for now and wins in some, Intel in others and we shall see what happens once they are on TSMC 3nm and next gen server CPUs come out next year. Intel's core problem right now is fixing the process deficit.
A bit disappointed in this shallow take. I suppose it has become fashionable to show extreme negativity towards Intel and count them out. That's good for those of us who see the risk/reward as attractive and can see the possibility of rays of sunshine that could lead to a turnaround.
Full Disclosure: I own $INTC, $AMD and $TSM (for different reasons and bought at different times).
I’ve been about as supportive as I could be on Intel 18A.
And man, I don’t know what to say but it’s like you didn’t read it. IBM had a fab back in the day. Nor do I think AMD is exactly in a great spot either, who needs a CPU? Seriously? client is making not that much.
I literally think that Intel is in a worse spot than IBM, and if you disagree with me sure, but there’s almost no hope for commodity design (Intel CPU) and its foundry or bust, which is an objective worse business than the Intel IDM juggernaut of 1990s-2010s.
Sorry if this didn’t fit your priors, but right now 18A is fledgling. Even if they pull it off, we are talking about a business that won’t make that much money, at least compared to the Intel of old. Disagree.
I did read the article. From a business point of view, yes Intel did miss some platform shifts. But unlike IBM which was never number 1 or 2 in PC sales or semi fabrication for long, Intel has not been that far behind technology wise other than the 10nm stumble. Their problem has been execution, management and strategy.
A high-performance CPU design is not a commodity. There are only a handful of teams that can do that. And CPUs are needed ultimately to do many things in computing, so that business is not going away.
Assuming they pull it off, even with slightly lower margins as a fab (compared to the IDM model) doesn't mean they won't make money. If the world is going to need a lot of fab capacity to build all these AI accelerators in the future, then getting a big enough chunk of that volume will suffice.
I'm not saying Intel will easily accomplish this, but they aren't quite dead yet. We shall see what happens.
Dude it is a commodity. Why is it that everyone can have their own that is +/- 10% of the best one possible w/ a few fees to ARM.
I think that the problem is that 18A rules, foundry matters, but it's really hard to see the way out for design unless they start cutting hard yesterday. RIF should be 40%.... then we are cooking.
Lip Bu Tan left for a reason man
Yes Lip-Bu Tan leaving is big negative I agree. I have changed my views. See my twitter post about that.
But you are being delusional if you think everyone else is within 10% by just paying fees to ARM. $QCOM only got there because they bought the ex $AAPL crack design team via the Nuvia acquisition and they still can't quite match Apple. Yes, the gap has narrowed but it's not that easy. Most of the ARM cloud CPUs so far are just scale out and not high performance. BTW even $NVDA tried mightily in this space and failed. Look up the history of Stexar and Project Denver.
Actually I wonder , in your opinion, what gotten Brookfield and Apollo interested in investing billions is those scips?