Cable companies is an entirely different business, different distribution, close to fixed demand and with ultra minimalistic R&D.
ShintaiDK said:There are only 2 ways for you to get cheaper CPUs. One is cutting R&D. But that means the segment commits suicide. The other way is higher volume. In the long run I dont see any room for other companies. Simply due to expense costs. You can have 1 healthy company or you can get 2 or more dying companies that cant produce anything good.
Intel has a track record for doing illegal things all over the world at the expense of its competitors. It doesnt play fair and its been proven time and time again.
Cable companies is an entirely different business, different distribution, close to fixed demand and with ultra minimalistic R&D.
Which would you recommend:
(1)...I choose ONE and only ONE semiconductor manufacturer, to be solely responsible for ALL production/marketing/R&D/PRICING of DDR5/DDR6, or whatever future memory requirements are ?
(2)...I choose TWO or a small number of COMPETING semiconductor manufacturers, who, after agreeing the specifications of e.g. DDR6, will go out and compete with each other, on issues such as who is fastest to market, PRICE, reliability, available speeds, etc etc.
Now lets jump ahead 5 or 10 years into the future.
How will the specs, availabilty and especially PRICING, vary between options (1) and (2) ?[/COLOR]
Problem with 1:
TSMC... already being done by all the small companies going to the single large company.
Do you really want TSMC to pump out all the dies?
Theres a reason why Intel builds new fabs from scratch when a node change is done.
Do you know how long it takes TSMC to even accept a new node?
Problem with 2...
IS NO ONE CAN COMPETE.
I would also Chose 1 when if it came down to it.
One would have way more money in research.
One would have more staff to facilitate customer service.
One would have more security the company wont file for chapter 11.
Anyhow... no... Its too late to do anything to intel now.
The government needs to wait for another company to handle the offload from intel should intel leave.
You cant just Ban a monopoloy, when your country is built ontop of the monopoly.
And i can think of a million things wrong with india's internal laws b4 they can even consider sueing Intel.
I hate when the little guy tries to be a big guy, and think he has big guy powers.
It doesnt work that way... not with Americans.. not with an American Corporation.
This is why American Businessmen are always frowned upon...
And yet, the argument for competition being harmful in the two industries is the same one, that they are natural monopolies.
Is the exact argument used for cable companies. They argued that two (or more) companies wouldn't be able to afford the maintenance costs of running multiple networks, and so the government had to prevent competitors from coming into the market and lowering service quality. While the industries are very different, the argument for monopoly being the optimal market structure is the same.
If this isn't your argument, why do you think competition is harmful in the semiconductor industry?
Please note, the question is not "Can Intel act as a traditional monopolist without AMD/Qualcomm/Samsung/etc", it is "Why is [any] competition harmful in the semiconductor industry". Don't get me wrong, the former question is still very important, but conflating the two questions (as is often done on this forum) is not productive.
Let me try to get to the bottom of, if competition is or is not a good thing with Intel, and accusations against Intel, such as this threads title.
Please let me ask you a hypothetical question.
Imagine I was made in charge of the entire memory ddr3, ddr4 etc market.
(N.B. I am factiously making up the terms DDR5, DDR6 etc, which may or may not exist in the future).
DDR5++, needs R&D money, especially for the memory chips that will make it up, etc.
Which would you recommend:
(1)...I choose ONE and only ONE semiconductor manufacturer, to be solely responsible for ALL production/marketing/R&D/PRICING of DDR5/DDR6, or whatever future memory requirements are ?
(2)...I choose TWO or a small number of COMPETING semiconductor manufacturers, who, after agreeing the specifications of e.g. DDR6, will go out and compete with each other, on issues such as who is fastest to market, PRICE, reliability, available speeds, etc etc.
Now lets jump ahead 5 or 10 years into the future.
How will the specs, availabilty and especially PRICING, vary between options (1) and (2) ?
I can see your logic, and your post makes a lot of sense to me (I'm nodding in agreement).
BUT, really scared. Because if there was ONLY one memory manufacturer (or cpu manufacturer for that matter), the dangers of them charging sky high prices for the memory, and almost never bothering to develop new, faster and better versions, makes it a VERY SCARY prospect to me.
I'm still sore with the memory (excuse the pun) when memory prices (can't remember date, something like late 90's) became HUGE. It cost a small (large) fortune to buy memory for your computer, and the maximum/typical memory sizes, did not go up for years, at that time.
It was a horrible time for computer enthusiasts, memory wise. I don't want to ever get back to those days.
How many CPUs do you think Intel would sell if they raised prices 50%? It would not be 400 million CPUs. More like 100 million.
I'd love to see the statistical analysis and market research that led you to this conclusion.
Intel sells about 100 million chips annually now, so you're saying they'd be able to sell the same amount of product at 50% higher ASP. Tell me again why you think Intel wouldn't do this.
Ever heard of the DRAM chartel? Not to mention the endless row of bankrupt memory makers.
Competition hurts innovation in a fast moving high risk segment. Because nobody dares to take the investment needed in risk of losing. Why do you think we use slow memory today with parallel busses. Yet all other busses are serialized now.
People complain about IPC increase now. But the same people rejected IA64. IA64 costed billions and took multiple years. AMD64 was a hotfix hack made on a tissuepaper in a lunchbreak (dramatized). And guess where we stand now.
Although competition significantly reduces the available funds the smaller (than one big company) companies have available to pay for developments, there are also dangers, if there is only ONE company (= No competition).
What happens if the ONE single (competition free) company, decides to maximize profits, by delaying, squeezing the budgets of, and scope limiting R&D work.
So, I'm still not happy with the ONE company (no competition) model, but unfortunately, this seems to be the way we are all heading, with Intel at least.
- Delay R&D = more profits
- Spend less on R&D = more profits
- Reduce capability of what is being R&D = Cheaper = More profit
- Charge more for our (competition free) stuff = More profit
- Ignore risky R&D projects (only do safe ones) = More profits
- Delay for a long time, speed improving R&D = more profits
- Sack half the R&D department = More profits
Going back in history, they were worried that IBM had become a monopoly. I think their solution was to split IBM up and give them severe operating regulations, such as "must publish details specs" and "open source like" stuff.
I'm NOT a professional analysts, and don't know much about the IBM stuff (it was in the fifties, without giving my age away, I was not very active then), but I don't think the anti-monopoly ATTACK on IBM, has been good for IBM or probably the rest of the world.
(Although I'm typing this out on my 'IBM' compatible PC, which may still have cost $3999, if IBM had kept their monopoly, who knows, and I'd be using OS/2 or something).
' IA64'==>> Not all big R&D things work out. I'm not familiar with the exact details, but there were a number of fundamental issues with it, which contributed to its demise. Such as, it relied too much on compiler writers needing to re-order instructions, to maximum running speed (rather than doing instruction re-ordering on-chip, like it is, now).
You completely mix up things with different segments and different demands.
Let me try another way. Haswell most likely costed around 5-6B$ to develop. And the price only goes up. But the market is too small to support the development cost from multiple companies.
So sooner or later you will have to choose. Just like in any other industry and we can use the 3 way picture. Imagine quality, performance and competition. But you can only select 2. And forget pricing, the prices are already where Intel gains maximum profit. Unless you prefer the "competition" back in the K8 days. Then an i5 4670 would cost almost 3 times as much.
We can already see this with AMD. They cant afford it, not even close.
But claiming that Intel being the only one is the best outcome for consumers (and innovation) is completely unfounded.
yeah but the issue is intel has no real competition at the point.
BUT, really scared. Because if there was ONLY one memory manufacturer (or cpu manufacturer for that matter), the dangers of them charging sky high prices for the memory, and almost never bothering to develop new, faster and better versions, makes it a VERY SCARY prospect to me.
You keep thinking that people more or less will buy the same volume, nomatter price or innovation.
Im pretty sure that intel was proven guilty a long time ago.
Intel has a track record for doing illegal things all over the world at the expense of its competitors. It doesnt play fair and its been proven time and time again.
I also see this as other countries trying to get a piece of intel because there country is going into a financial crisis and want some pie.
Honestly at this point a boycott of intel would DESTORY a country more so then destory the company.
Now I think I understand WHAT you mean. But, I am NOT happy with it.
WITHOUT competition we may get:
6-core Intels at $999
Because if they charged $1999, they would get such a huge reduction in sales, that the profits would collapse.
If they sold it for $499, they would sell more, but profits per unit would be less.
Hence they go with $999 (or whatever).
WITH competition
6-core AMD and/or Arm and/or someBusiness $199, good enough performance
6-core Intel at $300, good performance
----------------------------------
I want to see $300 processor availability, not $1000 ones. And I also want the $1999 option, for less demanding applications.
----------------------------------
But I DO accept that I am NO expert on economics, and the current market situation is very complicated.
So how would you then suggest the spiralling prices of IC designs is to be payed? Half the current companies cant even afford 14nm IC designs.
To keep competition, you would have to give up something. because the money is not there to support it. Unless you are willing to pay x times the price of the product per competitor you want to have.
ShintaiDK said:And if anyone should be in doubt. The sole purpuse of capitalism is to create monopolies.
If this was true, then Intel would not need to play dirty.
If this was true, then Intel would not need to play dirty.