There is plenty of talk about breaking Intel into a design side and fab side, so they can be split. The US government has an interest in on-shore fabs. Intel is the only real choice to do that. It seems to me, if China collapses, Intel future is far less bright. As long as China doesn't collapse, the US government is going to pour money on Intel to lithograph the smallest transistors it possibly can. I'd like to see Intel succeed but I'd also like to see China fail. I don't want to see China completely decimated, just collapse enough to reduce their capacity for war long enough for someone to take out Xi and put a civilized human being in power.
The market is responding extremely well to rumors of Elon Musk as an Intel suitor. I feel the same. Even though I have lost a whole lot of respect for Elon, he might be the best possible owner of Intel.
Hell no! Too much power for one man. A very unstable man at that. Assuming the foundry part can get back to being profitable, I'd rather Nvidia or Apple have it and consolidate Intel toward making designs only, since the x86 license cannot be transferred (which is total market manipulation crap).
I am tempted to buy a small quantity of INTC in hopes of Elon Musk buying the company. Whoever buys it will probably pay the somewhat standard 15~20% premium. This might be a good short term hold. Or, it could be a bucket of crap that strips people of their money. Place your bets. If Intel's 18A process is 10% as good as Pat Gelsinger directly stated, Intel has a tremendous future.
The new Intel GPU has some very positive user reviews and with a price under $300 it's a bargain, so that might be a good starting point for Musk.
At this point, I believe Elon Musk is the only outside buy out possibility for Intel. Intel sells a ton of chips into China. They require this revenue. To continue doing this, Intel will require buyout approval from the CCP. The CCP have not approved a couple of ideas, like Qualcomm. China is trying to kill the US. They will surely target the US chip industry, just as the US has targeted the Chinese semi industry. All China has to do is boycott every potential Intel buy out, they will condemn Intel to a 20 year, slow, death. Intel will cease to be relevant in two years. China is in a strong position here, since any CEO Intel will choose will be incompetent and made even more so by lack of board support. Possible outcomes. 1) 20 year slow death 2) Someone buys Intel without Chinese approval and call their bluff. It would hurt china to lose Intel parts but that leverage will go away in time so the sooner and more aggressive, the more likely this path will succeed. 3) Elon may have singular sway with the CCP. I'm inclined to think he has the best chance of gaining CCP approval. 4) Forced merger between AMD and Intel. Both companies already have access to the Chinese market but the CCP could still block this in China.
Further. If Intel and AMD were to merge, and I understand the US government has floated this idea, it would be a pretty unusual situation. AMD doesn't need Intel. The design side of Intel would just bog down AMD who would surely cannibalize the patents and send nearly all staff down the road. It would be like if Ford bought GM. They don't need each other. The only thing that would make sense is if Intel were subdivided and the foundry side sold to AMD, perhaps along with a couple of extremely specific parts of the design side. This would turn Intel into AMD circa 2004 but without Dresden. I think the question is if AMD would be open to this, more than if Intel would be open to it. Intel will have almost no choice. I'm not sure AMD is open to it. So, that brings us back to Elon.
Back to Elon. Intel is said to sell 50M CPUs per year. I'm confident this is not true, at the moment. I'd be surprised if they are selling at an annualized rate of 40M and that number will shrink. Tesla sold about 1.8M cars in 2024. Each vehicle has an AMD CPU in it. If Tesla were to switch from AMD to Intel, it would be a significant hit to AMD and a 4.5% volume gain for Intel. Elon buying Intel makes sense on a couple of levels. You may notice I have not considered the possibility of Elon not going for this idea. First, he said he was interested. Second, his hubris will be in control of this, not his business acumen.
ARC is a brilliant move by Intel. That's why I think 99% of Intel CEOs will kill the project. I really do estimate CEOs ability to understand industry dynamics near zero. Intel seems to be working toward getting ARC onto M.2. If they can make a viable product, and it seems clear they can, they could create a whole new market segment of gaming laptops with modular GPUs. M.2 is 4 lanes of PCIe. The top performing video card, right now, is the 4090. The 4090 uses 16 lanes of PCIe 4.0. That is substantially less bandwidth than 8 lanes of PCIe 5.0. (Note: PCIe 5.0 has double the serialization rate plus some other efficiency up lifts) PCIe 6.0 is just around the corner and the PCIe 7.0 spec will be frozen shortly. 4 lanes of PCIe 6.0 will be the same or better than 16 lanes of PCIe 4.0 which is what the top GPU of today uses. It's plausible to move top tier gaming to laptops. The constraints will be power and heat. The winner of the power and heat war will be ARC, not nVidia or AMD. If we can get 4070Ti levels of performance out of a 90W laptop, it would be possible to reform the entire gaming segment and kill the desktop market as we know it. Let's be clear. I didn't just invent this strategy. This has been Intel's strategy all along. I just don't think the Intel Board has the patience to succeed on this, or any, strategy.
The way other investors and analysts think hurts my head. It's difficult to conjure much respect but I feel I must in order to conduct myself how I wish. Consider this Yahoo News article that compares AMD to Intel and comes to the conclusion that Intel's far, far worse performance makes them a better investment. They seem oblivious to the idea that Intel is on a trajectory to receivership. Their board has demonstrated poor judgement and zero attention span. They are clearly going to select a CEO who will promise the world and deliver nothing but heartbreak. The only hope Intel has is for a board shakeup. This company is poisoned from the top. That might come in the form of a buy-out or merger. If Intel foundaries are separated and sold, the new company might do well with them. Meanwhile, AMD is among the best run companies in the world. Lisa Su has watched over AMD as it went from zero to hero over the last 10 years. All of that tech advance came from her team but she was smart enough to embrace the great things coming from her team and she was part of an epic trajectory of success. What then, is going on in this "analysis" article? First, not much analysis. Second, I think they are basing the up side on the idea that all stocks return to peak and Intel has far more gain potential, based on this theory. It's like they don't realize there is a business behind the ticker symbol or perhaps they do and don't consider it relevant. https://finance.yahoo.com/news/better-tech-stock-intel-vs-100000730.html
"Meanwhile, AMD's free cash flow of $1 billion compared to Intel's negative $12 billion suggests AMD is in better financial standing. It's true that AMD's shares are trading at a premium compared to Intel, with a significantly higher forward P/E. However, it feels like the more reliable investment option, with a more stable position in the technology sector." Seems to me that they are pushing for AMD, as would any sane person right now.
You are right. Either I posted the wrong article or they changed that one. The article I read had a few paragraphs from Motley Fool at the bottom of the article talking about how much money could be made by investing in Intel. I apologize for any confusion I have caused.
No worries at all. Your comments made me want to read the insanity. Could have been changed after the fact, like you said.
The rumor is that Intel hopes to bring 18A online in the second half of this year. If they can pull that off, even with a paper launch, that should get them close to the leading edge again. It won't be enough to lead but it will be a miraculous step forward.
I would love for this to be true, but like with anything Intel these days, I will believe it when it when the wafers are going out the door.
We share the same view on this, RTN. It would fit their operating profile to paper launch it. Even if yields are terrible, they could easily make some product and get a tiny amount to retail, they work on the process to make it commercially viable. I would guess, they have running 18A parts right now.
If someone wants to geek out on process nodes, this is a solid presentation from ARM in 2019. He discusses the path to 1nm. He also goes over feature size versus node number. He also talks about, and predicts, node expansion. I believe we are starting to see the potential of migration to larger nodes, right now. If 2nm is 2% better but 3x the cost, it will hurt anyone who migrates to 2nm. At that point, scale can happen with stacking modules, instead of process shrink. We see this with AMD EPYC. They are using a slightly inferior, but still near leading edge, to pack more cores on a chiplet and build a CPU complex that currently scales to 192 core. He also has an interesting segment on activation budget. Basically, it isn't possible to activate all transistors at once as there simply isn't sufficient power. It comes down to ratio limits of what ratio of transistors can be active. Adding more transistors is detrimental to this ratio, in the absence of power routing gains. Lots of engineering and a bit of science in this video.
Any thoughts on the leaked i9-15900 benchmarks? They seem 90% likely to be a 14900 scam but who knows what's going on at Intel. I expect they have early 15th gen cores and some modest 18A fab capacity in the engineering lab.
