Intel's Problems Are Even Worse Than You've Heard
wsj.com"You may think you know how much Intel is struggling, but the reality is worse."
From what I can tell, none of these problems really matter if their 18A bet pays off, and this article doesn’t talk about their 18A progress, from what I could quickly read.
Even if the 18A process is awesome, that doesn't help without the right product(s) to run through the fab.
Have a look at the graveyard of firms that tried being Intel Foundry customers over the past 15 years. Achronix was the perfect early foundry customer for Intel: FPGAs have regular physical structures replicated over the entire die. Yet they abandoned Intel as their early fab partner. Then Intel bought Altera and fucked it up as well.
Intel is on par with Google for killing non-core products. Look at StrongARM, various network processors, ultrawideband, ia64, their cellular modems... Anything that isn't going to directly help sell an x86 server, desktop or laptop CPU can be chopped, or at least not given the support needed to flourish. This was fine for decades, and Intel made a tonne of money doing that because that's what happens with monopolies: they print money. However, that strategy fell apart once the cost of fabs and fab R&D became high enough that more products across different market segments (read as: non-core market segments that Intel middle management doesn't care about) are required to generate a fair return on investment. Just look at how many top tier customers TSMC is able to amortize its fab R&D over today compared to Intel.
Any significant culture change at Intel is going to be nearly impossible given the inertia of 124,000 employees doing what they've always done.
I appreciate the in-depth reply. But if the foundry is able to produce the next gen of their core ultra processors, wouldn’t that lead to a sudden flock of renewed interest, especially if the foundry becomes the new core business model? Not doubting that inertia is there, or that management is incompetent, but they’ve bet the entire company on foundry, unlike before, which shows a lot of commitment this time around. Especially if people aren’t able to get access to the cutting edge node in Taiwan due to giants buying out the entire capacity, it seems like intel would be a natural second choice for anything bleeding edge, no?
1 "good" process node does not a trusted partner make. Leading edge semiconductors have design lead times measured in years. This means that the design decision for what fab to use for a given product that is being manufactured today was made anywhere from 2 to 5 years ago. Sure, designs can be ported from one fab to another if the processes are similar enough and the design is not overly aggressive, but that is much more difficult for complex designs like high end CPUs where power management and frequency scaling are intricately woven together to optimize performance within a given process technology. Just look at the difficulties Intel had with stability in shipping 13th and 14th generation Core CPUs where the wrong voltage was being used causing damage to the CPU. When hardware teams are pushed too hard too fast, quality suffers.
Risk in hardware is many orders of magnitude worse than in software, so the people making the decisions about what fab to use for a product tend to make much more conservative choices to help minimize risk. Intel as a fab partner is risky. Who in their right mind would select Intel when even Intel's own internal design teams made the decision to outsource manufacturing of some CPUs to TSMC in recent years? Intel has a lot of trust to build amongst hardware designers before that will happen at any significant scale.
From what I understand, Intel went with TSMC because their own nodes weren't ready yet, which makes sense to me. I imagine once their node is production ready, they will switch back. It's also interesting that for a single generation or two, they were able to switch to TSMC, with the plan to switch back, which, as an outsider, suggests to me that changing between nodes is not as tall of a task as one might think.
Additionally, I feel a lot of people are forgetting their biggest surefire customer--the United States military. If they get anything operational, even at 2x or 3x the cost of TSMC, the US will buy them and give them the resources and experience to serve larger markets, as the US has done with computing in the past.
They have a mountain to climb, for sure, but the path towards recovery still seems possible as long as Uncle Sam wants to buy from them.
The US military will never be the "biggest customer" of a fab. The only true volume products the military needs are bullets.
For reference, Intel CPU sales are on the order of 50 million units per year. As a comparison, Apple is selling on the order of 232 million iPhones per year. This is why missing the entire mobile market has hurt Intel, and there's no way the US military is buying more than 232 million of anything other than bullets (and maybe paper products).
The subsidies being paid to Intel and other manufacturers aren't a substitute for volume manufacturing of semiconductors. The more volume you can push through a fab, the more information you can collect and use to understand how to tune the process. Without the kind of volume that TSMC is pushing through its fabs, Intel doesn't have a way of getting the data it needs to tune its fabs. Less data on the process means that it's harder to develop and tune the next generation process. This means that the longer the negative aspects keep feeding into the next iteration, the harder it becomes to pull out of the negative feedback cycle.
Intel needs volume, and if it doesn't find something to run through their fabs in the hundreds of millions of units, then things can only get worse for Intel.
Good point, We'll just have to see, I suppose. I'm really curious how far Intel will get before they fail, if they do indeed fail. Given the current demand for cutting edge chips and TSMC's risky monopoly, I know a lot of people are hoping they somehow pull through
I think that Intel being split up into separate design and manufacturing firms might well be the optimal path forward. It is quite clear that AMD did the right thing when it spun out Global Foundaries 15 years ago. Were Intel to do so, the urgency of winning customers to survive might well give the manufacturing business the clarity of purpose and drive needed to do what needs to be done. Plus the new board of directors would not be able to allow the ship to continue drifting.
None of the problems are worse than I have heard. In fact it sounds like a rehash of what I wrote on HN. Especially the way it was presented and certain points. Not that I care about Intel now Pat is gone.