Hedge Fund Third Point Urges Intel To Explore Deal Options (reuters.com) 115
Activist hedge fund Third Point LLC is pushing Intel Corp to explore strategic alternatives, including whether it should keep chip design and production under one roof, according to a letter it sent to the company's chairman on Tuesday that was reviewed by Reuters. From the report: Were it to gain traction, Third Point's push for changes could lead to a major shakeup at Intel, which has been slow to respond to investor calls to outsource more of its manufacturing capacity. It could also lead to the unwinding of some of its acquisitions, such as the $16.7 billion purchase of programmable chip maker Altera in 2015. Third Point Chief Executive Daniel Loeb wrote to Intel Chairman Omar Ishrak calling for immediate action to boost the company's position as a major provider of processor chips for PCs and data centers. The New York-based fund has amassed a nearly $1 billion stake in Intel, according to people familiar with the matter.
Intel is insufficiently managed, it seems to me. (Score:2)
Micheal Dell (Score:2)
says, sell the assets and give the money back to the shareholders.
Sell everything and give the money to investors (Score:2)
Seriously, why anyone who runs a business and isn't planing to gut it like a fish would ever listen to a hedge fund manager is beyond me. Those guys exist to devour useful companies and shit out the remains.
Re:Sell everything and give the money to investors (Score:4, Informative)
That's not what a hedge fund is.
You're thinking of "vulture funds."
A hedge fund is a minimally-regulated fund only accessible to certain experienced investors, that trades in high risk investments and can engage in leverage. When a hedge fund is invested in distressed securities it would normally be in the form of derivatives. The hedge fund invests in derivatives generally in the hopes that they'll turn around. A hedge is a side investment that might go way up when the rest of the market goes down. I just use a VIX ETF for that, myself, but I'm only a small active trader not an accredited investor.
Failed derivatives get sold to the vulture fund after they fail. Liquidating distressed assets is a specialist investment with lots of moving parts and lots of lawyers.
Re: (Score:2)
Re: (Score:2)
Compare Intel and AMD's performance this year, and you'll easily see why a hedge fund would be investing in Intel right now; because it is probably a bad investment, but that's already baked into the price, and if they pull a rabbit out of their hat it is big money. That's what hedging is all about.
That said, this isn't something a person who is investing in a company says; this is something somebody who wants to invest in the company says. But they want a lower price than the current price, so they want th
Re: (Score:2)
sell the assets
Who would they sell them to? Their fabs have fallen behind TSMC, so how much are they worth? The only companies interested would be those required to manufacture domestically for defense systems.
But if they are going to sell, the time to do so is now. There is no point in sitting on a depreciating asset.
Re: (Score:2)
Fabs are worth a lot if they're either modern, or equipped for things still being made. Fabs that don't have value to others would be ones that don't have value to Intel either, because they're using obsolete processes.
I'll bet TI would love to have it, they have lots of US fabs, and lots of government customers who appreciate it.
Re: (Score:3)
Even obsolete fabs have quite a bit of value, and as the IoT rolls out they're possibly more valuable to that market segment than the new ones since they tend to be more flexible. Older fabs with larger die sizes are also good for ICs that function in difficult environments but don't need the specialty tweaks like milspec radiation hardening (IoT stuff, again). As an example Mercury's LNL-3300, one of the most installed access control boards in the world, uses a 59 megahertz CPU because the thing has to b
Re: (Score:2)
Intel's fabs are not that far behind. The finance media is fixated on nanometer numbers. Process technology has lots of idiosyncrasies beyond that one number. Intel has been struggling to move their process to newer generations, but they have been making improvements within the current generation. There are also factors that make their 10nm as good as or better than TSMC 7nm. But, TSMC is ramping up 5nm.
Intel needs to get their issues resolved and get their next generation ready soon. They have lost t
Re: (Score:2)
Intel's problem isn't fab process..., but competition from ARM
Intel's problem is that, as customers switch to ARM or AMD, Intel has less and less money to invest in their fabs. So they fall behind further, causing more customers to abandon them.
They are in a vicious cycle with no obvious way out other than going fabless like everyone else.
We're more likely to see Intel pivot from general purpose CPUs to more specialized processors for things like AI.
There is already a ton of competition in that space from Nvidia, Google's TPU, Tesla, and Apple's Neural Engine.
Unlike Intel, none of these companies are weighed down by the cost of running their own fabs.
Re: (Score:2)
Intel's problem is that, as customers switch to ARM or AMD, Intel has less and less money to invest in their fabs. So they fall behind further, causing more customers to abandon them.
That isn't Intel's problem, lol. They are still pulling in money hand over fist. They aren't falling behind because of money, if that were the case it would be as simple as dumping more money in, which they currently have plenty of.
Re: (Score:2)
Actually their revenue started a decline in Q3 2020, and it's going to get worse. They can't sell 14nm Xeons forrever.
Re: (Score:2)
Re: Micheal Dell (Score:2)
The problem is, if you want to really push the limits of technology, you NEED your own fabs, or you'll never be able to leapfrog over the PRESENT state of the art.
Here's a historical example: Andrew Carnegie started Carnegie Steel, which eventually became a vast global market leader worth billions. But... he didn't do it because his goal in life was to become a fabulously rich oligarch. He was an engineer who wanted to design bridges capable of spanning the Mississippi River (and others). He came up with cl
Re: (Score:2)
You've got a rosy view of Intel's fab problems. The fabs themselves are worth quite a bit, and were Intel to spin off or sell their fab business, the infrastructure could be quite valuable to companies like TSMC and Samsung (as would some of Intel's IP). Intel has a lot of cash-on-hand, so retooling could be done to order within a few years' time to generate a decent lasting income source for Intel. TSMC is capacity-constrained and also in need of fabs that are geopolitically out of reach of a bellicose
Re: (Score:2)
Aside from, you know, making and selling a heap of chips.
Not everything needs to be on the latest nm, Intel is still manufacturing chips for embedded systems, still manufacturing ASICs, they still manufacture older CPUs from 4 or 5 years ago:
https://ark.intel.com/content/... [intel.com]
https://ark.intel.com/content/... [intel.com]
The last one there is still on 22nm process.
Yep the assets are depreciating, the moment they buy them, they're depreciating, that's what they do. As long as they're profitable while doing so is what matte
Re: (Score:2)
TSMC and Samsung could definitely use extra fab space, assuming Intel is willing to retool.
Re: (Score:2)
Just because they're not cutting edge, doesn't mean they're not worth something. There is a huge market for tech that is not the latest and greatest. Previous gen tech is tried and tested so has a use where reliability is key.
Previous gen tech is generally cheaper, so can easily sell into markets which are price sensitive and/or not performance critical.
Re: (Score:2)
Intel can't continue to command huge margins on yesterday's nodes.
Re: (Score:2)
Indeed Intel could become the next Motorola. A massive selloff of assets and know how to China etc and a huge short term bonus to shareholders. American capitalism is really good for the 0.1%, just not very good for American businesses. Enjoy the ride suckers.
Re: (Score:2)
Find a new meme, bro... your "[ ] is insufficiently managed" meme is all played out.
We need leaders who are logical about ALL issues. (Score:3)
One story about extreme examples: Trump always ignores our most important problems [post-gazette.com] (Feb. 6, 2019)
My grandparent comment gives links to major issues about Intel management: Intel is insufficiently managed, it seems to me. Three of the MANY examples. [slashdot.org]
Re: (Score:3)
When people are talking about something businessy, that isn't even fun or funny, they're probably not meme-ing. It is probably just a recurring theme.
I gotta disagree (Score:2)
Intel's got problems, but only because AMD got their shit together and ARM is now fast enough to replace them. Intel is still way, way faster in a lot of work loads though.
None of this is to say
Re: (Score:3)
Re: (Score:2)
Intel's problem is their culture is rotten to the core. There's been a massive talent exodus from the company. You can take away AMD and ARM and they'd still be a mess.
yeah (Score:4, Insightful)
Just focus in ONLY the product that makes money.
Like that amazing company that used to make absolutely everything and dozens of their products where in every lab in the world.
And now they sell "ink subscriptions".
Insufficient management of HP (Score:2)
One story about the on-going insufficient management of HP: Carly Fiorina's disastrous record as HP's CEO [fortune.com]. (Sept. 21, 2015)
Re: (Score:2)
Weird, somebody should tell them about HPE.
Re: (Score:2)
I'm wondering how much more mileage there is in the 'Carly wrecked HP' trope. HP was a zombie corp before Carly crapped on it even harder, and it's been over 15 years since she was put to pasture.
At some point — even if it was her fault in the first place — whatever is wrong with HP isn't Carly's fault anymore. Not that I have any love for Carly Fiorina, but blaming her for stuff she hasn't had a hand in for the life span of someone that can drive a vehicle starts to look like a cop out. Do
Re: Insufficient management of HP (Score:2)
Also remember she is just one person.
It takes a certain corporate culture to choose somebody like her in the first place, and to let her do what she did.
It's not like that suddenly appeared or disappeared with her.
Exactly: Overall corporate culture is the issue. (Score:2)
Thanks for the excellent logic. Overall corporate culture is the issue.
Re: (Score:2)
Right, right, that's why GE, Samsung, Soft Bank, Panasonic, Honeywell, etc all do so poorly; too many revenue centers.
Re: (Score:2)
That company is still around. HP split off the medical and electrical test equipment division as Agilent, and then Agilent split off the electrical stuff as Keysight.
Keysight gear is still decent. Some of their oscilloscopes could do with a bit of an upgrade but basically they are good bits of equipment, if over-priced in the face of competition from people like Rigol.
Re: (Score:2)
Just focus in ONLY the product that makes money.
Like that amazing company that used to make absolutely everything and dozens of their products where in every lab in the world.
And now they sell "ink subscriptions".
Intel reminds me of Sears. Sad to see the slow death happening. Intel, to survive, needs progressive new blood with industry knowledge.
Yea outsource Intel to Asia. (Score:5, Insightful)
Re: Yea outsource Intel to Asia. (Score:2)
Maybe not that, but they better do SOMETHING or all those fine American jobs will disappear when intel goes bankrupt while AMD, nvidia, Apple and TSMC absolutely eat Intel's lunch.
Re: (Score:2)
Because that's what happens when you listen to greedy shareholders.
If you have shareholders who aren't greedy, but invested money in your company anyway, they'd be an even worse source of information.
The subject is investing, investors, or management of a public company, depending on your perspective. In all these cases, that somebody was greedy is irrelevant other than as an prerequisite to be a stakeholder.
The reason not to listen to him is that a "hedge" is a side bet you make in case the main bet fails. Eg, hedge fund managers are contrarian experts who are, of course,
Re: (Score:2)
So a hedge fund manager's job is to spout gibberish until they're lucky 1/10 times. Pretty much what I thought.
Re: (Score:2)
No. A hedge fund manager's job is to make bad investments until they're lucky.
Spouting gibberish is expected of everybody in the industry, that's part of all the financial jobs.
Re: Yea outsource Intel to Asia. (Score:2)
I otherwise ageee, but:
What exactly is the point of nationalism?
Honest question.
Being nice to others and being nice to oneself are not mutually exclusive, you know? The whole point of social behavior is that it is advantageous for both, to work together.
Or are you one of those lizard brain psychopaths that aren't happy until they can eat their own babies? (Lizards have been seen doing that.)
Maybe you just assume it's gonna be bad because people are gonna use you and not give back because that's how you trea
Re: (Score:3)
Are you the type that spits on your neighbors? Or do you even live in the US?
The US has been so nice to China they have all our jobs, manufacturing, and technology. It is time to improve ourselves for once and stop this short sighted crap.
Re: (Score:2)
Intel's problem is that they are stuck on 14nm and it's hurting their CPUs. They run hot and consume a lot of power (i.e. crap battery life).
The fabs in Taiwan and China are years ahead of them now and Intel either has to catch up or start outsourcing. Catching up is hard, especially when your government is making it difficult to work with Chinese companies or bring in skilled individuals to help you develop your own fabs.
Obviously investors just want to make a fast buck so as far as they are concerned outs
Re: (Score:3)
What makes you think Intel thinks any more of our country than its greedy hedge fund investors? If they cared about the United States of America, they wouldn't have failed in moving to their 10nm process, and their 7nm process would be right around the corner. USA! USA! USA! Instead they screwed up everything. Now another American company - AMD - is out-designing them, while a company in a country we fiercely defend (Taiwain, RoC) is out-fabricating them. And yet another company in a country that our sol
Financial companies interested in immediate profit (Score:5, Insightful)
Carving off pieces of a company, selling them for immediate profit, and renting back some or all of that functionality is typical of financial companies looking to loot an existing company.
If a company actually makes the mistake of going public (and making a one-time profit by a sale of the stock) by letting voting shares get into the hands of the public (Google's voting shares are strictly controlled - its publicly traded shares are non-voting [investopedia.com]) then I guess it becomes susceptible to financial companies attempting to dismember it.
This concept of selling off and renting back real estate [cnn.com] or other functionality [forbes.com] seems especially dubious for long term company health and jobs, but it is good for consolidating wealth right now for the devouring company.
Re:Financial companies interested in immediate pro (Score:4, Interesting)
It seems like AMD had to do it because they were financially on the ropes a decade ago. Is Intel's situation anything comparable? Because to the untrained eye, keeping manufacturing in-house looks like one of Intel's biggest advantages over AMD.
Intel's fine (Score:2, Insightful)
Re: (Score:2)
No, Intel isn't fine. Their 10nm process is a complete failure up til this point. If delays continue, their fabs will take the entire company down with them.
Re: (Score:2)
keeping manufacturing in-house looks like one of Intel's biggest advantages over AMD.
It was when Intel's fabs were the best in the world or at least very close to it.
The problem as I understand it is Intel has been struggling with new fab processes*, while TSMC has been pushing ahead. Having your own fabs creates strong internal pressure to use your own fabs, even if that is not what will deliver the best product for your customers.
* Intel released their first 14nm desktop parts in early 2017 (mobile parts came a bit earlier), now it's nearly four years later and their desktop parts are sti
Re: (Score:2)
* Intel released their first 14nm desktop parts in early 2017 (mobile parts came a bit earlier), now it's nearly four years later and their desktop parts are still on 14nm.
You missed the step when they stopped using 14nm 3D trigates and started using essentially the same 16nm finfets as the rest of the industry was up to, while still calling it 14nm, and then the step where they did sort of shrink their 16nm finfets while still calling it 14nm
Re: (Score:2)
You're way off. The 14nm i7-5775C and i7-6700k both launched in 2015.
The best foundry was advantageous. Fourth best now (Score:2)
Having the best semiconductor manufacturing facility in the world in-house was a major advantage for Intel for a number of years.
Then, Intel was the FOURTH company to produce 14nm parts.
In the leading-edge chip business, being forced to use an outdated manufacturing process that can't compete with TSMC and GlobalFoundries isn't an advantage. It's a liability.
Re:The best foundry was advantageous. Fourth best (Score:4, Interesting)
I was under the impression that the nanometre classification was irrelevant, and Intel's 10nm process and TSMC's 7nm process produce the similar transistor densities (approximately 100 MTr/mm2 for Intel and 91 MTr/mm2 for TSMC); is there some other advantage to TSMC's approach (other than yield, which appears to be improving)?.
AMD's problem is they won't be able to supply enough chips to OEMs and distributors due to competing with everyone else for resources (hence all their current CPUs, GPUs, and both videogame consoles being out of stock); Intel has an edge there, given the in-house manufacturing.
Re: (Score:2)
nm, turns out that, like everything else in the world, it's not that simple. The density is only one factor, but the "10 nm" and "7 nm" are just marketing names. This video uses an electron microscope to inspect both processes [youtube.com], which was cool.
Re: (Score:2)
Intel still manufactures the vast majority of their CPUs on their aging 14nm process. 10nm is only responsible for some Atom-based network appliance CPUs, some FPGAs, and their Ice Lake/Tiger Lake mobile CPUs. Intel's bread-and-butter Xeon lineup is still stuck on 14nm with no end in sight. Ice Lake-SP is completely MIA outside of a few leaked benchmark results. Intel won't have a 10nm desktop part available until MAYBE Q3 2021, and that's if you're being really optimistic. And nobody knows what availa
Re: (Score:2)
AMD sold their fabs because their proprietary tech had basically failed to be different or better than what others were doing, and they don't have a diversified product line that lets them make efficient use of excess capacity.
Intel has done better on their fab tech, so it has more value, and they also have a lot of different hardware products they can make if they have short term excess capacity.
Like with machining; there are lots of factories that could have their own machine shop, because they need a fai
Re: Financial companies interested in immediate pr (Score:2)
Well, there's a reason hedge fund companies are compared to both vultures and pimps.
They are like pimps, if pimps would be sucking blood instead of money, and devour the corpse later too.
Re: (Score:2)
Intel has had significant issues with process tech for more than 5 years. When Intel deployed their 14nm process tech they were 2 full nodes ahead of ALL the competition. They are now 2 full nodes behind (if you consider that 10nm process is still broken).
Management should have made drastic changes to the process side 3 years ago after 10nm was delayed twice. Particularly after Zen 2 landed and neutralized their entire advantage in x86. But Intel management was too worried about disruption, they should have
You'd think we'd ban the practice (Score:5, Insightful)
Re: You'd think we'd ban the practice (Score:2, Troll)
Well, the solution is simple:
America has no problems *creating* that so-hated regulation. It's just that is usually comes from a corporation that temporarily wore a government skin via one of their lobbyist politicians. And it is usually done for anti-competitive reasons. ... you know ... the actual institution that serves the will of the people ... look bad so it can b
Which conveniently lets them bitch about being the poor poor victims, if a corporate enemy happens to do it to them. And make the government
Re: (Score:2)
If only we gave in to revolutionary socialist/communist bureaucrats with no industry experience. That would make for a much-better regulatory system!
Or you could just hire people with experience (Score:2)
Re: (Score:2)
Carving off pieces of a company, selling them for immediate profit, and renting back
Except that there would be no rent-back.
Intel's fabs are falling behind TSMC. They have been unable to fix that in the past and have no viable plan for the future.
So they would sell their fabs to someone who doesn't need top-end performance, and then Intel's products would be fabbed by TSMC.
So Intel would become fabless, just like AMD, Apple, ARM, Xilinx, Nvidia, Broadcom, Qualcomm, etc. It is a successful model that is working for many other companies.
Re: (Score:2)
Intel uses proprietary tech nobody else is ready to fab with. If it is an old generation, nobody would touch it. Unless it can be used for analog power components it will be liquidated when it becomes obsolete.
With Intel, who has valuable buildings cited near each other to maximize network effects, when they liquidate the equipment and replace it, in effect building an entirely new fab building inside the old walls, nobody knows but investment nerds who read their corporate filings. Other companies would ha
Re: (Score:2)
So Intel would become fabless, just like AMD, Apple, ARM, Xilinx, Nvidia, Broadcom, Qualcomm, etc. It is a successful model that is working for many other companies.
At least until China has its next serious leadership hiccup and the leadership decides to retake Taiwan to keep its populace happy.
"We had to destroy the province in order to save it."
Not going to do much fabbing with a smoking crater where TSMC (and much of the rest of Taiwan) used to be...
Re: (Score:2)
Jesus Christ. A) you act like "unable to fix that in the past" extends more than a few years. Like Intel has been desperately trying to catch up for a decade. What nonsense.
B) You call out your own problem in your final sentence, I'll let you figure it out. Hint: That's a lot of companies, huh? Wonder how many chips TSMC can crank out and what will happen when 50 companies are trying to get in their schedule. Might get weird, huh?
Re: (Score:2)
Except that there would be no rent-back. Intel's fabs are falling behind TSMC. They have been unable to fix that in the past and have no viable plan for the future. So they would sell their fabs to someone who doesn't need top-end performance, and then Intel's products would be fabbed by TSMC.
They have the same equipment as everyone else in the industry: cleanrooms, high-end scanners from ASML and Nikon (those are a large part of the capital investment) and all the machines for etching, coating, and inspection. What they don't have is TSMC's IP (both patents and trade secrets) to use that equipment for particular device designs.
Intel could sell a fab for good money to TSMC, although TSMC might not want to risk their IP leaking through former-Intel employees that still see their buddies from Int
Re: (Score:2)
TSMC has already signaled that they are not interested in becoming Intel's fab partner. They have other higher-priority customers in line ahead of Intel (which is why Intel is only getting "a few" 6nm wafers from TSMC, presumably for DG2 and Ponte Vecchio). The only way Intel can truly become fabless is to lease/sell their fabs to a company like TSMC or Samsung, pay to retool them with better process tech than what Intel has now, and then take wafers from those fabs.
Which, in my opinion, is the only way I
Re: (Score:2)
Once upon a time, vertical integration was considered a good thing: We still demand it from software giants (including Facebook and Google) but they're not acquiring land and factories, they're getting people and IP.
In the 1990s, "core function" was the new catch-word: (One of many; "down-sizing", "team" and "synergy" were others.) Governments were instructed to sell-off land and factories so they could create comparative advantage via private enterprise.
The only advantage I noticed, was government of
Re: (Score:2)
If a company actually makes the mistake of going public (and making a one-time profit by a sale of the stock)
It isn't a one-time profit. Equity in a public company is worth about 10x what equity in a private company is worth. And when the economy grows, you issue new shares to sell, and that doesn't dilute the company, it generally increases the value. Also if you need to raise funds you can issue new stock, and if other people agree that raising money was a good business decision, then the stock goes up, not down, and so isn't diluted.
Nobody is going to sell off something they would need to rent back unless it is
Re: (Score:2)
Intel is WAY too big and held by way to many institutional investments (mutual funds in particular) for this activist to be able to acquire enough shares to materially affect Intel's strategy through normal corporate channels.
Because of this they are trying the only strategy available, public pressure.
I put the odds of success here at 0.0001%. They will likely succeed in getting management to consider options, but Intel has always had management with significant investments in the company so management is l
Re: (Score:2)
Intel can't manage their fabs worth a damn. Everything they've done since 14nm has been a categorical failure, and even the 14nm process was pretty badly-delayed. If they sell off their fabs to Samsung and TSMC in exchange for access to better node tech (or specialized hybrid nodes, which is a thing: see IBM's POWER after GF acquired IBM's old fabs), then Intel could continue to design competitive CPUs and regain control of the markets they once dominated. If they DON'T go that route, their integrated de
Oh yeah, I always take my business advice (Score:5, Funny)
from hedge funds and venture capitals. They are known to ensure a healthy and long future for any company they invest in.
Amazing (Score:2)
Can you imagine reading this story and on AMD here on slashdot 10 years ago? Intel lead the way of chip fabrication and was arguably the inventor of the CPU on a single chip.
It comes to show mobile is where the true innovation is where as Intel is like IBM of legacy un innovation of old things. I guess the mobile and web revolutions killed Microsoft too as 15 years ago it would be unthinkable someone would not use Internet Explorer and that MS is playing catch up with Visual Studio tools and WSL.
I don't kno
Re: Amazing (Score:2)
Well, what is left of the original Microsoft?
They are only one husk away from beoxming a Linux & "cloud" company with some gaming side business.
Can you imagine Ballmer accept Linux inside Windows? And on their servers? He'd die of a heart attack, just to turn over in his grave at 2600RPM.
Re: (Score:2)
2600? More like 7200.
How might Intel be dismembered? (Score:2)
What would the most profitable way to butcher that cash cow look like and who might buy the steaks?
Let the foundries go (Score:2, Interesting)
Obviously, Intel should have started the process of divesting the foundries at least two years ago when it had already become obvious to anyone with as many as two brain cells to rub together that no single company has the scale to opertate its own foundries as processes edge ever closer to the physical limits of lithography. The longer Intel brass takes to admit reality, the longer INTC will continue to be the sick dog of technology stocks.
Re:Let the foundries go (Score:4, Interesting)
That's complete nonsense, TI has a dozen fabs, including ones they've bought from others. Some of the times when you saw a company selling their fabs, and came up with the false analysis that "no single company has the scale to opertate its own foundries," you didn't even notice that they'd sold it to some other company that is using it for their own products.
Microchip has 3 US foundries, including one it bought from Fujitsu and another it got when it acquired Atmel.
Rohm Semiconductor has 13, including 10 in Japan, 2 in the US, and 1 in Thailand. Do you think that Rohm would exist without their own fabs? When I'm shopping for ICs at Mouser I sometimes wonder, why is this company's version so much cheaper? Oh, Rohm has lots of their own fabs, that's why.
Murata has 6, honestly I thought they just made capacitors and filters, but it turns out they make lots of MEMS sensors and fancy doodads, in addition to their discount caps.
Bosch only has 2, with a third under construction, but that's because they built them freakin' huuuge. Everything they make is durable, they're great.
Samsung has as many as TSMC.
STM has 9.
Infineon has 22 in an impressive list of countries; Austria, Germany, Malaysia, Hungary, South Korea, USA, Indonesia, Singapore, Mexico, and China.
Look at the list of closed plants; only a few. And when you look at all the plants owned by somebody other than who built them, the vast majority were bought by companies that make their own products. There are only a few fabs that do the commercial production for fabless designers; most companies who try to do fabless design fail. And most companies that own their own fabs, are the ones who are successful in the marketplace.
https://en.wikipedia.org/wiki/... [wikipedia.org]
In short, you're an idiot and your analysis falls down instantly.
Re: (Score:2)
Lots of companies can make good money manufacturing specialized parts on obsolete fab nodes. But a PC CPU maker does not have that option. See?
Re: (Score:2)
Well fucking duh.
Oh, wait... did you think you were contradicting me? That's as deep as you got into the issues, and you still through it at the wall? LOL
Re: (Score:2)
You must be a lot of fun at parties.
Re: (Score:2)
Hurr hurr, no I would not have attended your party.
Re: Let the foundries go (Score:2)
Uum, no? That would mean Intel's foundries would die, and they would become TSMC's bitch too. Which can easily become worse than death.
Obviously, they should have made the same management decisions as TSMC, to get to the same process nodes too.
Alas, we do not know why they failed. Maybe someone on the inside will release a book in a few years.
Re: (Score:2)
Umm, yes. If Intel sells its fabs (unfortunately for pennies on the dollar, that's just the price of delay) it will be to an operator who intends to take on fab work for multiple customers. Intel tried this itself but they never fully comitted to it, and face it, Intel doesn't have that many core competencies. They are really, really good at milking a monopoly, but when that monopoly weakens they just flop around like a fish out of water. Investors need to put a stop to that, starting with bringing in a hat
That would kill them (Score:3)
At least in the CPU space, Intel never had really good designs, what they always had was superior manufacturing. If they had good CPU designs, a dwarf as AMD would not have been able to embarrass them time and again (also remember that the current architecture is the "AMD64" one, as Intel failed and failed and then failed again to come up with a 64 bit architecture that could follow i386). Intel can only try to reacquire that advantage and that means their own fabs. Anything else basically means long-term they can close shop. Or become a patent-troll.
Re: (Score:3)
Intel didnt really know what they were making so had unrealistic expectations. But what they did have was a bunch of big early contract, which might look like others agree ab
Re: (Score:2)
Intel didnt really know what they were making so had unrealistic expectations.
Which is a recurring topic for Intel in the CPU space. And if you look at company history, Intel comes from _memory_ (hence the superior Fabs) even if they try hard to bury that. AMD comes from signal processors and they were always better with regards to CPU engineering than Intel as a result.
Re: (Score:2)
The Intel Itanium disaster was only so because it was meant to replace x86. The engineering and design principles were great as a compute-focused cpu that offloads as much operation scheduling as possible to a compiler
The problem with your analysis is that intel never delivered a compiler that in turn delivered on their performance promises. Several systems which had nothing to do with x86 (in that they were not replacing x86 systems) were built with Itanium, and they were all overpriced too.
Intel's x86 compiler was long considered the best in the business*, and they thought they could somehow work the same magic to produce a compiler that would make Itanic performant. They were wrong.
* It might still produce the best-op
Re: That would kill them (Score:2)
Exaclty. Itanium wasn't bad. It just was bad at being a x86 drop-in, and developers were lazy.
I would have preferred a fresh start, dumping as much of x86 as possible.
Because now, ARM and RISC-V are taking over that part, and I hope AMD can see the trouble on the horizon that will mean for them, when the cruft of x86 limits their competitivity.
Re: (Score:2)
RISC-V isn't taking over from x86. It's cutting into embedded ARM territory.
Re: (Score:2)
It's already killing them. If Intel continues to cling to their fabs, they will shrink into irrelevance. Their entire revenue stream will be devastated. You will not be buying Intel PCs, workstations, laptops, or servers. And if they sell off Altera (bad idea) they'll be doubly dead.
Sure, that's a long term outlook... (Score:5, Insightful)
I sure trust a hedge fund to come up with solid idea for long term growth and stability of companies. They wouldn't just sow chaos on a market segment to profit of the instability. Bonus, even less jobs in the US. Excellent.
Intel outsourcing fabrication for CPUs and other critical parts would be a disaster. They do very, very well in the server and cloud markets because they can control every aspect of production to met orders and forecast availability. Making their supply chain more complex is not good.
TSMC is playing fast and loose with process definitions. We don't know what their yields are. Also, they can change on a dime. Rumors surfaced that they are reducing wafer availability to customers to increase profit two weeks ago. And they can do that at any time.
When customers want 40K CPUs, that's too much of risk. Intel would be better of trying to nab key talent from TSMC and others, get a bit humble and refocus on bending sand better than anybody.
Re: (Score:2)
Apologies for the lack of editing here. A bit embarrassing, to be honest.
Intel's reply: (Score:3)
"Who the hell are you? ... And who gave you this e-mail adress?"
Maybe I should write suggestions to Intel too... ... ;)
Or do you have to be the corporate equivalent of a pimp too?
Cause I don't think I can snort enough cocaine to inflate my sense of self-importance that much
Please dont pay attention... (Score:2)
So you can suffer as much as you made AMD suffer with your illegal tactics.
Like being forced to sell your headquarters.
Better yet, please die already.
Hysteria. (Score:4, Insightful)
So much idiocy in some of these threads. Let me ask a simple question: What fab, either cutting edge or cutting edge-2 generations, is sitting idle right now because there aren't enough fucking chips to be sold?
Fucking none. It'll be a genius business move for all involved when the whole world is competing for a small set of fabs to produce every CPU/SOC made for anything more powerful than a low-end phone.
How many chips do you people think TSMC can make?? Those "ancient, dilapidated" fabs you all seem to imaging Intel running are money factories, half of you think they are going to be mothballed and obsolete in 6 months.
Unless they let the inmates run the asylum, Intel isn't selling their fabs double-lol.
Re:Hysteria. (Score:4, Interesting)
From my first day at Intel orientation...
A fab isn't a factory and it's not even really a building.
It's a machine that just looks like a building.
They take the purest chemicals in one end, and put wafers out the other.
They are built, run a planned load for a planned time, and have a planned decommissioning.
Demand builds another fab. It doesn't keep one running longer.
It doesn't pay to chase diminishing yield from a fab that's wearing out.
Intel's edge was not so much the fabs they had, but that they knew how to build them.
hedge fund and long term profit (Score:2)