Intel is on life support. Can anything save it?(economist.com) |
Intel is on life support. Can anything save it?(economist.com) |
If it gives up on GPUs then its finished.
I worry more about their inept handling of recent CPU bugs than I do about their stock price or reduced dividends.
That was my impression, until the news came out that they had a massive round of layoffs. That looks like the opposite of investing, poisons the climate, causes people who can to seek greener pastures (potentially further disrupting those projects they allegedly invest into), ...
I'm sure there were other bad news that caused their stock drop from $30 to $20, and it's hard to separate the impact of the layoff from those, but I wouldn't be surprised if the layoff news (that generally tend to push stock values up for companies in a different situation) contributed to it (due to lowering confidence that the investments will happen and work out).
Yes, layoffs suck, but Intel seems to be in a position where something had to give, and giving up on building a leading edge fab would almost certainly lead to a death spiral. Additionally, Intel was and still is a huge company. They have more employees than TSMC & AMD combined (an imperfect comparison, but probably as close as one can get).
It slipped by me how much AMD eating a little more every quarter really added up over the last ~5 years. (~9% to ~50% of datacenter revenue)
Also, the foundry stuff just isn't working as a financial exercise: they can't split it out as a division and massively subsidize it and seem responsible. And there's ~0 light on the horizon. No one wants to use it, even Intel is falling back to using TSMC's leading nodes now, which is letting it tread water against ARM in laptops.
If their confusing plan to do 5 nodes in 2 years or whatever works out, that'd enable them to start reversing the tide: they'd still have to build the same muscles TSMC has from always being a foundry, like having generic designs that are already in the market available, and convince people to switch suppliers, which is always risky, and usually done over years.
> The tests conducted by Broadcom involved sending silicon wafers - the foot-wide discs on which chips are printed - through Intel's most advanced manufacturing process known as 18A, the sources said. Broadcom received the wafers back from Intel last month. After its engineers and executives studied the results, the company concluded the manufacturing process is not yet viable to move to high-volume production.
and
> A Broadcom spokesperson said the company is "evaluating the product and service offerings of Intel Foundry and have not concluded that evaluation."
The make or break is not this summer. It's next. Gelsinger himself said he bet the company on 18A. If it doesn't work out then yeah, Intel is in very deep trouble but until then, Intel still has 29B cash on hand which is not chump change.
It's not that I stan for Intel, the heck do I care, I dislike the reporting that goes around this topic.
Can their long term investments bear fruit? One might hope... but continuing to flail helplessly in the face of the first actual competition in decades may not allow them to harvest that crop effectively.
The near two-decade vacation Intel took from actually being good is insane in hindsight. They thought their competition was dead (due to Intel's own illegal behavior), so they just kept copy/pasting the same garbage over and over, shoveling it to the population with ever increasing price tags (anyone remember the days of $5,000-$10,000+ "gaming" CPU's?).
If any company deserves to fail, it is Intel. Their story is simultaneously hilarious and sad.
Holy smokes. I just looked at the INTC 5 year chart and it is, in fact, a bath of blood. I don't think it happens very often that the market cap of a company is less than the worth of its physical assets.
This was when basically every article about the company used the word beleaguered, they couldn’t actually ship the computers people wanted to buy, and they were flailing around with their next gen os.
The real problem is the manufacturing process. This used to be the power house for Intel. They kept it 5 years ahead of their competitors before the Krzanich era. But in all the geopolitical and infrastructural contexts, is it even feasible to restore that in America? It is not very clear. Intel needs to build up all the ecosystems for semiconductor manufacturing but America is far, far behind of Taiwan (and eastern Asia in general).
So far 18A seems to be a success. Already sub-0.40 D0 defect density and launch at some time 2025Q3 and fully ramped up in 2026.
Fabs will bleed money 2024 and 2025, there is no way around it.
Just because they are second place, doesn't mean they have lost the race.
You really have 3 cutting edge players in fabrication. TSMC who is the best by a significant margin, and Intel/Samsung. TSMC is certainly winning right now, but they have not "won."
If Intel doesn't choke on its debt payments or corporate culture in the next 3-5 years, they will have one shot at pulling the nose up out of their dive. Of course their altitude is low enough that if they fail to execute in even a small way they will hit the ground.
Intel certainly had those things. It has lots of interesting IP around processor design and tooling locked away in source control and so forth.
I think it is critically important whether Intel has retained the engineers who knew how to build world class products. Nehalem was excellent and shipped in 2008 so I'm sure they had the skills back then. They used to have a reputation for paying well.
It seems plausible that Intel is a bureaucratic horror show that no longer pays competitively, in which case it would be difficult to see why the engineers would still be there. Especially with redundancy offers waved around roughly annually.
I reckon they're dead.
McDonnell Douglas takeover Boeing is an example.
Prediction: Intel is done. It will become a zombie company like IBM or Oracle in a few years, without the license lock-in moat keeping those rotting corpses alive, but may be able to offer cheaper fab services for "it just was the state of the art" chip manufacturing market. They'll continue to focus on becoming a financial optimization organization and never return to being a technical leader. They've become a company of pure inertia.
Reasons for the prediction:
1. Intel is not longer the process node leader and hasn't been for a while (when even was that? 14nm? 2014?).
2. Intel seems to be struggling to keep lead engineering talent, board members, and leaders of all kinds. Jim Kelly, Lip-Bu Tan, and so on. They still have Jeff Wilcox, but who knows for how long? At least the CEO is an engineer finally...I guess.
3. They continue to launch then retract in otherwise profitable markets, not taking any particular lead, but not continuing to invest until they get it. They still have really only one core business.
4. Their branding strategy seems to be built around confusing consumers in the hopes that they accidentally buy the wrong things. Successful companies can be clear in their product naming, zombie companies end up trying to stuff every single market niche with as much microtuned, barely differentiated, SKUs as possible. In actuality it's probably financial engineering to figure out what to do with various classes of yield issues. Disable the bad part of the chip, slap a K or KS, or KF, or T, or whatever on it and get it on the shelf. I counted over a dozen different Core i5 14 gen SKUs, and over 20! 13 gen Core i5 SKUs. There's probably 200 different brand-new-from-Intel CPU SKUs on the market today. Insane.
5. They're getting feature for feature beaten by smaller, lower stakes competitors both in and out of their ISA. Those same competitors are almost always cheaper, and made at other fabs. In otherwords, Intel doesn't lead anywhere and has no lock-in.
6. Their main moat, their ISA, can be comfortably virtualized or emulated elsewhere, bugs and all, at entirely useful speeds meaning there is a ready exit ramp as customers wish to take it.
There's probably more, and it's probably continues to paint a very poor picture, but there is very little reason today, or in the near future, to stay with Intel as a manufacturer except for inertia.
This was like the stupid ass DEC hostile giveaway to Compaq while DEC had almost 2 gigabucks+ a year in enterprise services revenue that was going to continue forever. A single year of services revenue was almost as large as the entire merger deal.
And then, as PCs crashed and burned, Compaq couldn't figure out what to do with all that cash, either!
Executive morons all around.
Intel lost money the past two quarters. Revenue is still high, but having high "cash flow" doesnt work indefinitely if more of it flows out the door than in.
That's scary.
Fabs are so insanely expensive now. US$1 billion to US$4 billion each. Without the money to keep up, Intel becomes a niche player.
If I had to guess, I think the above is true. Well too bad Wall Street :)
If Intel is really in bad shape, then maybe it is time for some form of nationalization.
The era of WinTel is over. Megalithic data centers tend not to use MS OS and thus are free to choose AMD and increasingly ARM CPUs. For Intel to rush into AI would be, too little, too late.
* They missed out the rise of mobile
* They missed out on the rise of GPU computing
* AMD has been making gains in the server cpu market
What do they have left other than consumer computer market where their main advantage is name recognition?
Keen to see what the team that split off to do a RISC-V DSP chip end up producing. Even if they added dual RX/TX SDR front ends on a SoC, than it could open up a single-chip smartphone solution etc.
Intel is entrenched with process-people obsessed with features no market asked for like hardware RATs, and betting on imaginary AI economies.
There are paths to move forwards technologically, but Intel as a company has structured itself to be its own worst problem. =3
China needs TSMC to have economic and political leverage. US wants TSMC to become an American owned enterprise so it can prop up Intel.
South Korea or Samsung rather see TSMC burn to the ground giving it monopoly marketshare. It benefits if US and China destroys each other in the process.
There's like only 3 guys in town with TSMC dominating the market and it just happens to be in a very risky region.
Just like the pipelines that were blown up by Ukranians I can't help but feel we will be seeing similar sabotage to TSMC should a war occur.
This is great for the free market, tsmc engineers make less than half what US software folks make.
This is a win/win for the US if TSMC keeps puttering along as normal.
Aren't those extremely cheap? Do they have good margins on them?
I feel like if they sold a million of them $55MM in revenue isn’t moving the needle much at all, even with a 100% profit margin.
Yes, Intel is currently in a slump. Just like AMD was for a decade before Zen 2/3.
Poor business? Gimme a break. They could exterminate a smaller ~~African~~ scratch that, Central or South American country with zero consequences.
Joking aside, I can see the US engaging in heavy-handed protectionism to help Intel.
Yes, you can[0].
Also, you can see it when we punish other businesses who had better strategy
https://www.wsj.com/tech/ai/nvidia-justice-department-probe-...
Imho it's pretty likely that this will happen. Intel's IP will be sold off, while their manufacturing business remain, both due to the geopolitical situation and because although they have seen setbacks in recent years, they are still damn good at making chips.
This effect is more obvious in small countries, but large ones are not immune to it either, especially if their fortunes turn.
Is it the role of a for-profit public company to "support" another for-profit public company?
Free market, supply and demand, live by the sword die by the sword and all that.
If Intel made decisions that lead to it's downfall, surely that is on them.
It is (or it can be) if you want to avoid antitrust lawsuits. Same reason Microsoft bailed out Apple in the 90s.
Apple used to dual-source their chips from Samsung/TSMC before they went all in with TSMC, but it's clear they're now vulnerable to geopolitical tensions.
They support lowest cost manufacturing in China even when those suppliers have been proven to have stolen the tech (free R&D!) with state support.
Not really, it also has to do with timing, fab space, extra cost of maintaining that node, etc. I don't want to get into details as I work on 20A/18A development but I'll say that they share the same major process, 20A is just an earlier version. So it's not like Intel is dropping the process, it's just deciding to put it into production later.
There's some more information here: https://www.intel.com/content/www/us/en/newsroom/opinion/con...
i think Samsung is in a far better environment.
if Warren Buffet dumps his TSMC shares then we can't ignore these real risks to it
Kind of. If Intel keep making shit, it doesn't help anyone if they succeed doing it. They are nothing making products that are a viable alternative to TSMC, so it makes no difference if they vanish.
Not to mention that Apple had atleast a salvageable IP if done well. Nobody else could compete with them on shipping MacOS machines. Intel has no such moat.
Doesn't that need R&D though (i.e. people, and preferably motivated ones)?
Also, unless you know some non-public information, its not clear they are laying off anyone doing R&D. The only specific groups mentioned in the layoff memo are sales & marketing.
How many of them are inclined to stay after the value of any stock-based comp was utterly destroyed and layoffs have presumably smashed morale?
It's the microcontrollers and FPGAs controlling weapons systems that really should be made within the country boundaries and that should be completely fine on GF's machinery.
Also, the military semiconductor supply chain is already protected and has serious export restrictions, you just don't hear about it often.
Don't forget that today's leading edge nodes are tomorrow's trailing edge, so you really can't escape the issue.
There's been a continual drive towards better semiconductor processes. That's really important for things like GPU compute. I think we're past diminishing returns for a lot of other things. There's probably a ~14nm arm chip in the keyboard I'm typing on that's only there because it was the cheapest option. Does the ECU in a car benefit from being sub 7nm? Maybe, probably not by much.
Intel received a small slap on the wrist, shrugged it off, and then proceeded to do nothing with their illegally-won cleared playing field (making what is perhaps the single biggest business blunder in human history - they effectively were left with zero competition, and squandered the opportunity).
Meanwhile, AMD got scrappy and found a way to bounce back, and today is slowly swallowing Intel whole, one inch at a time.
Does Intel have two decades of runway to recover? I guess we're going to all find out together...
That's the special thing about leading semiconductors: they enable other technologies. From the personal computer to IoT and AI, all of these were at some point enabled by advances in semiconductor scaling.
You're only focused on what leading edge nodes provide today, but they enable much more than just speculative AI hyperscaling in the long term.
For years, AMD survived as a cheaper solution to Intel. They got competitive with Opteron (and invented x86-64), while Intel was trying to wring out the last drops out of Netburst.
Then Intel Core came out and Opteron got back to being a cheaper but slower solution.
The point you’ve given no argument for was that AMD didn’t suffer from Intels illegal practices. Not who cooks eggs better.