The thing about all of these articles analyzing Intel's problems is that nobody really knows the details of Intel's "problems" because it comes down to just one "problem" that we have no insight into: node size. What failures happened in Intel's engineering/engineering management of its fabs that led to it getting stuck at 14 nm? Only the people in charge of Intel's fabs know exactly what went wrong, and to my knowledge they're not talking. If Intel had kept chugging along and got down to 10 nm years ago when they first said they would, and then 7 nm by now, it wouldn't have any of these other problems. And we don't know exactly why that didn't happen.
Intel's problem was that they were slow getting their 10nm design online. That's no longer the case. Intel's new problem is much bigger than that at this point.
Until fairly recently, Intel had a clear competitive advantage: Their near monopoly on server and desktop CPUs. Recent events have illustrated that the industry is ready to move away from Intel entirely. Apple's M1 is certainly the most conspicuous example, but Microsoft is pushing that way (a bit slower), Amazon is already pushing their own server architecture and this is only going to accelerate.
Even if Intel can get their 7nm processes on line this year, Apple is gone, Amazon is gone, and more will follow. If Qualcomm is able to bring their new CPUs online from their recent acquisition, that's going to add another high performance desktop/ server ready CPU to the market.
Intel has done well so far because they can charge a pretty big premium as the premier x86 vendor. The days when x86 commands a price premium are quickly coming to and end. Even if Intel fixes their process, their ability to charge a premium for chips is fading fast.
We actually have a lot of insight in that Intel still doesn't have a good grasp on the problem. Their 10nm was supposed to enter volume production in mid 2018, and they still haven't truly entered volume production today. Additionally Intel announced in July 2020 that their 7nm is delayed by at least a year which means they figured out their node delay problem.
> We actually have a lot of insight in that Intel still doesn't have a good grasp on the problem. Their 10nm was supposed to enter volume production in mid 2018, and they still haven't truly entered volume production today. Additionally Intel announced in July 2020 that their 7nm is delayed by at least a year which means they figured out their node delay problem.
Knowing something happened is not the same as knowing "why" it happened. That's the point of my comment. We don't know why they were not able to achieve volume production on 10 nm earlier.
I'll also add that it's fascinating that both 10 nm and 7 nm are having issues.
My understanding (and please correct me if I'm wrong), is that the development of manufacturing capabilities for any given node is an independent process. It's like building two houses: the construction of the second house isn't dependent on the construction of the first. Likewise, the development of 7 nm isn't dependent on the perfection of 10 nm.
This perhaps suggests that there is a deep institutional problem at Intel, impacting multiple manufacturing processes. That is something more significant that a big manufacturing problem holding up the development of one node.
I think that's not quite right. While it's true that for each node they build different manufacturing lines, generating the required know-how is an iterative/evolutionary process in the same way that process node technology usually builds on the proven tech of the previous node.
SemiAccurate has written a lot about the reasons, for me the essence from that was: complacency, unrealistic goals, they didn't have a plan B in case schedule slips.
I think it's just a difficult problem. Intel is trying to do 10 nm without EUV. TSMC never solved that problem because they switched to EUV at that node size.
Wild speculations: "newness" budget for 10nm was already used up by other innovations. Or they earmarked all EUV resources for 7nm or 5nm. EUV steppers don't exactly grow on trees.
Wasn’t the issue that the whole industry did a joint venture, but Intel decided to go it alone?
I worked at a site (in a unrelated industry) where there was a lot of collaborative semiconductor stuff going on, and the only logo “missing” was Intel.
Samsung is the opposite of Intel: gaining market as mobile takes over in the collapse of Intel's former moat. They have more money to solve their problems.
I think it's pretty clear from the article what happened. They didn't have the capital (stemming from a lack of foresight and incentives) to invest in these fabs, relative to their competition.
If you look at this from an engineering standpoint, I think you'll miss the forest for the trees. From a business and strategy standpoint, this was classic case of disruption. Dominant player, Intel, was making tons of money on x86 and missed mobile opportunity. TSMC and Samsung seized on the opportunity to manufacture these chips when Intel wouldn't. As a result, they had more money to build/invest in research to build better fabs, which could be funded by the many customers buying mobile chips. Intel, being the only customer of their fabs, would only have money to improve their fabs if they sold more x86 chips (which were stagnating). By this time, it was too late.
