Anyone with billions of dollars can do that...
Is Intel's dominance just dumb luck? *spawn
- Thread starter Esrever
- Start date
Right, and how they [Intel] got their billions was by out engineering the alleged competition.
Intel really gets where they are because they were basically a monopoly from the get go in a growing consumer space. I don't really see what that has to do with intel being a supplier of custom SoCs for consoles tho. When you have 90% of the money in a massive industry for 30 years, the engineering can easily be bought. Intel can throw 5 billion dollars a year into buying market share in mobiles and still make an extremely profitable year simply because they control all the money in the sectors they are majorly invested in.
Because they out engineered all the alleged competition.
You said anyone with billions can compete because they have billions. You made it sound like Intel inherited the money without actually earning it or having engineering talent, I'm simply pointing out they got those billions through hard work and earning it.
Intel had a monopoly before home pc were even a thing because they had the first integrated circuit designs. Everything since then is just them throwing more money into the pot than the competition because intel had more money to throw at it. Intel owns x86 with an iron grip. They make money without the need of really pushing anything new. You make it sound like its hard to make money when you have a monopoly.
They started mostly making DRAM chips and then were out competed and moved into processors. Like many massive companies, it was started by a few smart people who executed well, then faced challenges, and managed to overcome them, with luck playing a role in a couple of places (IBM picking the 8088 and then being cloned basically won Intel the home computer CPU market despite lots of competition in that space from better designs). Their advancement in semiconductor fabrication ahead of all the competition came from smart investment - they weren't 'born' that way, neither with superior manufacturing nor with more money than everyone else. The started with what they could do, made a profit, invested it wisely and not so wisely, competed, and grew. Apart from the 'Wintel monopoly' which AMD still managed to compete with through better designs, I don't know that Intel had any competitive advantage over anyone else. Just having more money doesn't lead to success.
Intel weren't first in IC production, but they had the first microprocessor in a single chip, the 4004.
They weren't a monopoly though, they had a lot of competition from Z80, Motorola 6800 and CMOS 6502.
IBM made them when they picked the 8086 (8088 actually) over the M68000 for the PC.
Cheers
The second IBM picked intel, there was no way they could fail at making money. They had the entire desktop market which was fast growing and has been extremely lucrative since then. They also held the rights to deny anyone who even wants to compete by having a completely locked down and proprietor ISA which they do not license to anyone. That and Windows software ecosystem keeps x86 in desktops.
Of course not all companies continue to be successful even with large funding but I think you will find them to be a lot more common than companies that don't have any money and becoming successful. And AMD never really offered any competition. In general, companys with more money make more money than companies with less money and that has almost always been the case. By the the timescale of things, Intel made more money than AMD ever will and will continue to do so even if AMD wasn't already pretty much dead from lack of funding. Even when AMD had a competitive product, Intel outsold them 4 to 1 simply because they are bigger and had more marketing and more connections. Then Intel crushed them with 10x the R&D. Its pretty hard to not win a fight when you have 100x the resources.
Holding 80% of the marketshare and 95% of the revenue is an easy way for intel to keep their engineers paid, get their designs made and get their products shipped. All they have to be is 1 step ahead of a company 1% its size and intel gets to do what ever it wants. Blowing billions into buying market share is pretty easy from there. Nobody can beat intel at pricing because intel's chips are free whenever they want them to be.
Of course not all companies continue to be successful even with large funding but I think you will find them to be a lot more common than companies that don't have any money and becoming successful. And AMD never really offered any competition. In general, companys with more money make more money than companies with less money and that has almost always been the case. By the the timescale of things, Intel made more money than AMD ever will and will continue to do so even if AMD wasn't already pretty much dead from lack of funding. Even when AMD had a competitive product, Intel outsold them 4 to 1 simply because they are bigger and had more marketing and more connections. Then Intel crushed them with 10x the R&D. Its pretty hard to not win a fight when you have 100x the resources.
Holding 80% of the marketshare and 95% of the revenue is an easy way for intel to keep their engineers paid, get their designs made and get their products shipped. All they have to be is 1 step ahead of a company 1% its size and intel gets to do what ever it wants. Blowing billions into buying market share is pretty easy from there. Nobody can beat intel at pricing because intel's chips are free whenever they want them to be.
D
Deleted member 11852
Guest
You can say this with the benefit for 40+ years of hindsight but at the time the prevailing view was why does you average consumer or office worker need an IBM PC? Intel's competitors, who were producing microprocessors for entertainment devices like home computers and games consoles, looked a much safer bet because entertainment is an easy sell. Particularly as video game arcades were sucking us vast amounts of money and you could save money by playing at home.
I think the hindsight is precisely Esrever's point. At the time, Intel were just another tech company, but because of how the PC market developed, which no-one predicted or even was pursuing (it wasn't IBM's intention to produce a machine that'd be pirated - they didn't particularly believe in PCs as a concept and just threw something out there to have a presence IIRC), Intel got a huge lucky break. If IBM had picked Motorola, we'd be talking about Motorola now instead of Intel in all likelihood. I don't know if Intel took an active role in that decision or not.
But Intel were still a normal company with the same competitions faced and same issues to tackle. If it wasn't for the mobile chip department, they might have crumbled already under the weight of their legacy hardware.
But Intel were still a normal company with the same competitions faced and same issues to tackle. If it wasn't for the mobile chip department, they might have crumbled already under the weight of their legacy hardware.
tongue_of_colicab
Veteran
Still the IBM PC came out in what? 1981. That was atleast a decade before pcs became somewhat mainstream. If Intels subsequent cpu's would have been far worse from a cost/performance pov than competing designs there is no way people would have sticked with them for so long.
A strange statement, considering the PC was not exactly a runaway success initially. Also, the PC was locked into mainly home and office roles for a very long time. It wasn't until the late 90s when the big iron CPU manufacturers had finally been marginalized into mostly irrelevancy or entirely squeezed out of the market that Intel could start dominating the lucrative server workspace as well.
Today, that same space that was once owned by HP PA-RISC, DEC Alpha, Sun Sparc, and MIPS Rxxxx might largely get taken over by less costly, more lightweight ARM chips. (IBM Power still soldiers on to some extent.)
http://en.wikipedia.org/wiki/Intel#Anti-competitive_allegations
Notably 38% of Dell's operating income in 2006 came out of Intel's rebates, and this was when it mattered.
Intel was sitting smugly patting themselves on the back for taking 4 years to upgrade the 386 to the 486 (which had inferior performance to other contemporary architectures) until they lost their iron grip of x86 due to an AMD lawsuit in the early 90's and got some serious competition _in the x86 space_. This last point matters as businesses relied on software that ran on this instruction set and had little other actual choice once that Wintel juggernaut started rolling. (EDIT: It matters much less today since we have the local horsepower and global infrastructure to be able to sit behind virtual machines and browsers.) It was indeed a harmful monopoly which prevented many consumers from enjoying the fruits of free competition through superior engineering.
This was amply demonstrated when they woke up engineering-wise in the mid 90's with the dynasty of the Pentium and PPro, forged and refined in the fire of competition. Prices plummeted and performance soared. Then when they faltered engineering-wise as manufacturing processes were not able to produce the frequencies the P4 was designed for, they resorted to "rebates." It took another salvo of legal troubles in the mid-2000's for them to actually return to the extended run of engineering excellence that we see today.
I'll give Intel credit where it's due for their engineering innovations but their business dept. is dirty dirty dirty. You might say any business focused on profits would do the same, but I'll maintain point of capitalism is to align profits with social value instead of turning yourself into a mindless maw of profits.
Last edited:
Inherently, capitalism strives to maximize market share and profits, above and beyond all else. Creating a monopoly, and then bleed your customers of as much money as you possibly can is the ultimate goal. Sometimes a company succeeds with that - or almost anyway. Thankfully there's some regulation still of the supposedly "free" market to rein in the worst aspects of it.
Intel has some truly horrific skeletons in its closet, and not too far back in the past either. I wouldn't buy their CPUs at all if AMDs weren't so pathetic in comparison. It's truly sad to see how AMD is breaking apart before our eyes. My first PC had a K6-166 in it, which happily overclocked to 208 with an Alpha PAL CPU cooler fitted with a 60mm YS Tech fan with newfangled 3-pin RPM monitoring "smart" plug...
Oh, those were the days! Don't want them to come back though! Nostalgia only tastes sweet for so long; the hardware from that period is SO damn primitive that it's ridiculous...Mass-production silicon VLSI is a vastly complicated and capital-intensive industry, and the high risk and cost leads to mandatory high volumes or high prices, which presents high barriers for entry and strong pressures for consolidation.
Actually designing and manufacturing a device on the leading edge nodes is itself vastly complicated and presents high barriers for entry.
Even in the absence of a desire to corner the market, we are seeing that the cost of duplicate capacity and the exponentially rising cost of break-even forcing either horizontal consolidation for manufacturers to avoid idling equipment and risk mitigation or vertical integration to allow designs that work or outside revenue to subsidize the effort.
If we want to see a real change from this, we need something that fundamentally changes the physical realities of this production, sort of like when Intel on switching from memory and risked going into integrated circuits.
Without a sea-change like that, the way the technologies work is going to demand somebody big will win out.
Actually designing and manufacturing a device on the leading edge nodes is itself vastly complicated and presents high barriers for entry.
Even in the absence of a desire to corner the market, we are seeing that the cost of duplicate capacity and the exponentially rising cost of break-even forcing either horizontal consolidation for manufacturers to avoid idling equipment and risk mitigation or vertical integration to allow designs that work or outside revenue to subsidize the effort.
If we want to see a real change from this, we need something that fundamentally changes the physical realities of this production, sort of like when Intel on switching from memory and risked going into integrated circuits.
Without a sea-change like that, the way the technologies work is going to demand somebody big will win out.
Silent_Buddha
Legend
It wasn't just that either. Intel's success can't be solely or possibly even mainly laid at the feet of Intel, unfortunately.
As someone mentioned above, a significant reason that Intel came to dominance was the ease with which competitors to IBM could make IBM PC/XT (these were brands at the time similar to Commodore 64, Apple ][, Apple Mactintosh, etc.) clones. Without that key piece, Intel likely never would have attained dominance, and if they had it would have been significantly harder.
As a comparison, other computer manufacturers at the time fiercely protected their computer designs. Apple originally started to allow other companies to make clones of their Apple computers. Well allow, isn't the right word. They originally didn't do anything to stop it. However, once those clones started to gain a market presence, Apple were quick to shut them down.
Had Apple, not done that, the Motorola 68k series of CPUs might have been able to challenge Intel much more effectively. As with a plethora of Apple clones, there would have been an alternative large homogenous computing pool to feed into them. As it was, Motorola had to do with a plethora of small disparate computing architectures using their designs compared to the homogeneous and growing IBM PC/XT clone computing architecture.
Note that prior to the IBM PC/XT architecture gaining dominance due to the massive number of cheap (keyword) clones, Motorola was able to compete quite effectively with Intel.
Entering the 90's the IBM PC/XT computing architecture was firmly entrenched (due to those clones) and from there it would be increasingly difficult to compete with Intel.
Now, imagine if IBM had protected their computer designs as well and as fervently as other computer manufacturer's had done at the time? Intel would have been at a serious disadvantage as their CPU would have been used in far less devices then Motorola's competing 68xxx series.
If not for the IBM PC/XT clones, we might be complaining about Motorola's predatory monopolistic practices.
Note - this isn't to pardon Intel for some of their business practices after they gained market dominance. This is to point out that to gain that dominance they not only had to execute well in the face of fierce competition, but they also had to rely on one of their key partners at the time, NOT protecting their intellectual property. The one key point, more than any other lead Intel to the position they are in now.
The fact that they then leveraged that to maintain manufacturing dominance which lead to performance dominance indicates that while Intel can be fat and lazy at times (386 to 486 transition, AMD's x86-64 initiative, etc.) they have been remarkably good on a technological front when pushed.
In other words, even without some of those monopolistic practices (as with the Dell example), they were never in danger of losing their market dominance due to their engineers as well as their management and the fact that IBM never protected their computing architecture as their competitors did with their proprietary architectures.
It wasn't the fact that IBM chose Intel that made Intel successful. It was the fact that IBM didn't protect their IP that lead to Intel being successful. You'll note that IBM personal computers basically failed in the market BEFORE Intel gained dominance over their CPU rivals. The IBM PC/XT clones however raked it in.
Regards,
SB
Last edited:
It did seem like the capital requirements are going up and that you need to be a certain size to sustain any kind of progress, however a competitive model works at any scale. With the rise of mobile devices, we're seeing renewed competition in the fab space, see the rise of Samsung and TSMC as fabs which are doing 14nm more or less concurrently with Intel. Hopefully the transition from silicon to alternative materials will be eased by the presence of multiple players.
Last edited:
I wouldn't characterize that as a rise at the leading edge: http://www.realworldtech.com/iedm-2005/2/
A little less than ten years ago, the list at the leading edge included the likes of AMD, IBM, Infineon, Freescale, Phillips, STM, Sony, Toshiba, and Chartered.
Texas Instruments used to be out there, and these are off the top of my head when 2005 was witnessing consolidation or parties dropping back on the treadmill.
Fujitsu, UMC, NEC, Renesas, etc.
DEC had a fab.
Back in the day, "Real Men" had fabs.
So Intel, TSMC, Samsung, and GF which is a sputtering satellite of Samsung made up of a Frankenstein's monster of pieces of a number of companies listed earlier.
One unified manufacturer/fab, a giant player in Taiwan, one globe-spanning chaebol tugging along an Abu Dhabi vanity project, and a lot of fading memories.
The list of former in-house leading core designs is long as well.
Addendum: given the marketing-driven numerology for nodes these days, whether the foundries are that much closer to the leader than in the past is open for debate.
A little less than ten years ago, the list at the leading edge included the likes of AMD, IBM, Infineon, Freescale, Phillips, STM, Sony, Toshiba, and Chartered.
Texas Instruments used to be out there, and these are off the top of my head when 2005 was witnessing consolidation or parties dropping back on the treadmill.
Fujitsu, UMC, NEC, Renesas, etc.
DEC had a fab.
Back in the day, "Real Men" had fabs.
So Intel, TSMC, Samsung, and GF which is a sputtering satellite of Samsung made up of a Frankenstein's monster of pieces of a number of companies listed earlier.
One unified manufacturer/fab, a giant player in Taiwan, one globe-spanning chaebol tugging along an Abu Dhabi vanity project, and a lot of fading memories.
The list of former in-house leading core designs is long as well.
Addendum: given the marketing-driven numerology for nodes these days, whether the foundries are that much closer to the leader than in the past is open for debate.
Last edited:
Without intense competition, the biggest player would just sit at the last available process for as long as possible. There were a total of 3 pitch refinements in processes in the 80's (not counting other refinements)
http://en.wikipedia.org/wiki/1.5_µm_process
versus 3 in this half decade alone. The effort put into the most recent three also utterly dwarfs those from the 80's. The capital needs are now astronomical, but this is the result of competitive pressure rather than some imperative of Intel that just because they have the most money to spend in advancing fab processes that they should. (This almost always the case, look at the progress your cable box has made in the last decade versus your cell phone.)
As for process name give and take, Intel itself isn't exempt from this either:
http://www.chip-architect.com/news/2010_09_04_AMDs_Bobcat_versus_Intels_Atom.html
With the commoditization of CPUs (shift from x86 to ARM, C/APUs now more tailored to customer needs rather than what Intel comes down the mountain with) and more companies opting to rent fab space rather than own, Intel's had to cancel a new fab and opt to sell excess capacity in existing fabs. It's not clear what design rules customers will need to follow and whether Intel retains advantage in an open environment where they fab for multiple customers.
I would agree that Intel's the leader, but despite the synergies, company specific fabs may be a thorn at Intel's side in the long run. Even they faced delays at 14nm, and it's not a give that they'll be able to meet capital requirements to be competitive going forward. I think there's good reason IBM paid Global Foundries to take that business off their hands and Apple doesn't own any fabs despite being able to buy all of Intel with cash on hand. What surprises me is that Intel's fabs haven't been spun off yet and that there's still as many cutting edge players in the world as there are. (Samsung, TSMC, GlobalFoundries, Intel)
http://en.wikipedia.org/wiki/1.5_µm_process
versus 3 in this half decade alone. The effort put into the most recent three also utterly dwarfs those from the 80's. The capital needs are now astronomical, but this is the result of competitive pressure rather than some imperative of Intel that just because they have the most money to spend in advancing fab processes that they should. (This almost always the case, look at the progress your cable box has made in the last decade versus your cell phone.)
As for process name give and take, Intel itself isn't exempt from this either:
http://www.chip-architect.com/news/2010_09_04_AMDs_Bobcat_versus_Intels_Atom.html
With the commoditization of CPUs (shift from x86 to ARM, C/APUs now more tailored to customer needs rather than what Intel comes down the mountain with) and more companies opting to rent fab space rather than own, Intel's had to cancel a new fab and opt to sell excess capacity in existing fabs. It's not clear what design rules customers will need to follow and whether Intel retains advantage in an open environment where they fab for multiple customers.
I would agree that Intel's the leader, but despite the synergies, company specific fabs may be a thorn at Intel's side in the long run. Even they faced delays at 14nm, and it's not a give that they'll be able to meet capital requirements to be competitive going forward. I think there's good reason IBM paid Global Foundries to take that business off their hands and Apple doesn't own any fabs despite being able to buy all of Intel with cash on hand. What surprises me is that Intel's fabs haven't been spun off yet and that there's still as many cutting edge players in the world as there are. (Samsung, TSMC, GlobalFoundries, Intel)
Similar threads
- Locked