How is entering the GPU industry too expensive?

[BFG10K]

Their image quality has actually improved alot over the earlier IGPs. Graphic anomalies are few now.

Sub-par AF at the hardware level is not an anomaly, it’s a purposeful design decision.

The driver support is alot better as well. The last few revisions improved HD3000/ 2000 performance significantly. Intel doesn't gain performance by intentionally cheating IQ. They can squeeze so much performance out of a 12 EU igp is really their silicon technology that allows clocks of 1.35ghz with ample overclock headroom to spare. IVB IGP on 22nm 3d transistors will be even more impressive.

It doesn’t matter what they do with the core given it’ll always be castrated by system bandwidth. It’ll compete with low-end discrete GPUs at best. A 5570 is too slow to play anything 3D, yet it’s still a magnitude faster than Intel’s solutions.

 

[BFG10K]

I wouldn't pay engineers peanuts, but I would pay myself less if I thought it would help things out and I would employ the optimum number of people and make their work conditions as good as possible.

LMAO.

Be sure to ask those optimum engineers of yours to implement forcing specific Z buffer depths in old games.

 

[s44]

IVB IGP on 22nm 3d transistors will be even more impressive.

Fixed AF, twice the horsepower... Unfortunately that only puts them about at par with Llano at a point when Trinity will probably be out.

 

[Anarchist420]

You're being philosophical and unrealistic (are you one of those Austrians?). Your no-IP stance only makes sense in an era where individual workers churned out products using their skilled hands using simple tools. So stealing 'IP' then could not reproduce the same products unless the rogue co. also had workers of the same skill.

Nvidia/ATI designs are embodied in their graphics cards so if a rogue co. steals the designs it can make products as good as Nvidia/ATIs without expending an extra drop of sweat. Its the same with books, music cds, designer fashion etc. Do you think pirate music cd sellers should use the defense by saying they didn't steal anything because Beyonce or whomever will always remember what they sang?

Yes, I'm an Austrian. I don't think Pirate music cd sellers should have to have a legal defense. But if they have to be brought to court, then I guess that's a good defense.

 

[bononos]

Yes, I'm an Austrian. I don't think Pirate music cd sellers should have to have a legal defense. But if they have to be brought to court, then I guess that's a good defense.

I knew it (Austrian).
So what you are saying is that a poor and hungry budding musician somewhere could be putting in alot of time and effort into making an album/track and when he is finished, I could just take his CD and run a million copies and sell it just for myself and pocket the profit because hey, I didn't steal the song thats in the musicians head did I?
Do you get the problem there, which makes capitalism impossible without some sort of IP?

 

[HeXen]

I knew it (Austrian).
So what you are saying is that a poor and hungry budding musician somewhere could be putting in alot of time and effort into making an album/track

see thats the thing, they don't care if the musician starves. Caring is the key word here, cause you have to really give a damn first.

 

[Anarchist420]

I knew it (Austrian).
So what you are saying is that a poor and hungry budding musician somewhere could be putting in alot of time and effort into making an album/track and when he is finished, I could just take his CD and run a million copies and sell it just for myself and pocket the profit because hey, I didn't steal the song thats in the musicians head did I?
Do you get the problem there, which makes capitalism impossible without some sort of IP?

I see no problem.

The musician can perform live. The musician can do a lot of things that could make them money, and it's the publishers who gain the most from IP, not the artist. Getting rid of IP would definitely destroy the big publishers more than it would the artist.

 

[bononos]

I see no problem.

The musician can perform live. The musician can do a lot of things that could make them money, and it's the publishers who gain the most from IP, not the artist. Getting rid of IP would definitely destroy the big publishers more than it would the artist.

I'm well aware about the problem of oligopoly in the music industry with publishers acting as gatekeepers and adding little to no value to whatever they are doing - which is why I used the analogy of a hungry musician independently publishing his work (perhaps not clear enough).

So what would you say if the musician lived in a perfect world where there he could reap the benefits of his music without any gatekeepers?

How do you expect graphics corps to survive if their work could simply be copied by upstart competitors? Why do through the trouble of finding ways and means to pay scores of engineers to work for a few years to develop a product only to have it stolen right before their eyes?

And do you have a response to my #48 post about modern industry and how the Austrian no-IP theory does not fit in.

 

[Yuriman]

IP is there for this reason - nobody would ever bother to create anything new if they thought their time and effort would have no reward. Sure, you could have an R&D company that sells designs and produces nothing, but IP makes it so you don't have to spy on every employee and keep your designs under high security. You have a legal defense, and nobody would dare steal it.

 

[Anarchist420]

I'm well aware about the problem of oligopoly in the music industry with publishers acting as gatekeepers and adding little to no value to whatever they are doing - which is why I used the analogy of a hungry musician independently publishing his work (perhaps not clear enough).

So what would you say if the musician lived in a perfect world where there he could reap the benefits of his music without any gatekeepers?

How do you expect graphics corps to survive if their work could simply be copied by upstart competitors? Why do through the trouble of finding ways and means to pay scores of engineers to work for a few years to develop a product only to have it stolen right before their eyes?

And do you have a response to my #48 post about modern industry and how the Austrian no-IP theory does not fit in.

Profits would be at market levels, there would be plenty of useful inventions. What they'd do, is they'd come first to the shelves with something, then that would sell for a few days maybe a few years, then someone would copy it, then the person who was copied would be motivated to invent something new and be the first to bring it to the shelves. That too would eventually be copied, but there would still be a motivation to make things, especially not for profit but maybe because it was something they wanted for themselves or something someone asked them to make.

Also, people could make money off of services, land, and/or agriculture, not just complex industrial goods. We've gone too much into a tangible and non-agrarian economy IMO.

 

[Anarchist420]

LMAO.

Be sure to ask those optimum engineers of yours to implement forcing specific Z buffer depths in old games.

Andrew Lauritzen from B3D said it wasn't that hard for good programmers IIRC. Don't quote me on that, but I think that was what he meant.

 

[s44]

We've gone too much into a tangible and non-agrarian economy IMO.

Translation: not enough people are starving and miserably scratching away at subsistence "living".

No thanks.

 

[bononos]

Profits would be at market levels, there would be plenty of useful inventions. What they'd do, is they'd come first to the shelves with somethihttp://forums.anandtech.com/newreply.php?do=newreply&p=32352181ng, then that would sell for a few days maybe a few years, then someone would copy it, then the person who was copied would be motivated to invent something new and be the first to bring it to the shelves. That too would eventually be copied, but there would still be a motivation to make things, especially not for profit but maybe because it was something they wanted for themselves or something someone asked them to make.

Also, people could make money off of services, land, and/or agriculture, not just complex industrial goods. We've gone too much into a tangible and non-agrarian economy IMO.

Just like I thought. You have no real response except to say that the 'market' will (magically) take care of everything. Austrians are stuck in the past and have no understanding of modern industry.

You say the person whose product was copied would be motivated to invent something new but the reality is that the person (or graphics corp) would be demotivated instead. You also assume such a person would have the resources after losing their business to copycats to have the free time etc to indulge themselves and freebies for others.

You want more intangibles? Odd since global trade in intangibles like services/ finance instruments have grown much greater than the manufactures of 'real stuff' in recent decades yet inequality has grown greater, poverty has increased, economies have been less stable. The current view is to go back to tangibles. I agree that agriculture would be a good source of employment if it weren't monopolized by big corps.

 

[Capt Caveman]

Profits would be at market levels, there would be plenty of useful inventions. What they'd do, is they'd come first to the shelves with something, then that would sell for a few days maybe a few years, then someone would copy it, then the person who was copied would be motivated to invent something new and be the first to bring it to the shelves. That too would eventually be copied, but there would still be a motivation to make things, especially not for profit but maybe because it was something they wanted for themselves or something someone asked them to make.

Also, people could make money off of services, land, and/or agriculture, not just complex industrial goods. We've gone too much into a tangible and non-agrarian economy IMO.

Please stop posting this FUD. It costs 100s of millions to billions of dollars to research, develop, etc... What profits are you going to generate after a couple of days of selling a new item when it gets copied? The company would go out of business b/c it wouldn't be able to generate any money to cover their expenses and you would have zero innovation. Only a clueless person like you would think 1000s of people would want to work for free to develop a product so that someone else could copy and sell for less money.

Get the hell out of the basement and learn how the real world works b/c you obviously don't have a friggin clue.

Between going back to the gold standard and this crap, the world would be set back 100 years by kids living in their basement. Are you parents seriously feeding you this crap? Again, how old are you?

 

[bryanW1995]

OP, if I gave you $6B today, told you to spend it all over the course of the next 4 yrs by creating a company from the ground-up, hiring the thousands of engineers needed to design a competitive GPU (competitive with what the competition will be fielding in 4 yrs, not good enough to be competitive with what they are fielding today), buying licenses to the IP necessary to support DX9/10/11/12 and so on, building a support team, software driver team, AIB relations and support, etc...and let's say by some miracle you actually did it, 4yrs from now you have a chip that is competitive...what is going to be the ROR (rate of return) on my $6B investment into your business venture?

Have you seen the profit margins for AMD and Nvidia? After 4yrs my $6B would be gone, paid out in salaries, depreciated equipment purchases, and licensing fees, and it would take you 15 yrs to simply recoup my initial investment (if profits are a consistent $100m per quarter) let alone to actually generate a positive return on my money.

And what would be my ROR had I simply invested that $6B into purchasing Nvidia stock, or Intel stock (no one in their right mind would purchase AMD stock, so let's not go there )...or investing the money in many other "safer bets" that can be made?

$6B * 19 yrs @ 3% = a veritable buttload more money than I'd get by starting a new GPU company.

The world isn't lacking investors with $6B to invest in such an endeavor, the problem is that the world has a plethora of more lucrative and lower-risk opportunities for investors with $6B to pursue.

The proof of the pudding is in the eating...do you see Warren Buffet investing in GPU businesses? or do you see him investing billions into banks? :hmm:

Don't sell yourself short. You could buy AMD lock/stock/barrel, even at a 20% premium that would only be 5 billion. Then sell off the cpu side to the Qatari govt, and you probably only have a billion or two left invested in amd. Then all you have to do is find 1 good guy to run the entire gpu operation, probably Al Gore since he did create the internet.

THAT shows how bad it is right now. The ATI division has been moderately successful over the past few years, but at their current rate they will never "pay back" the $5 billion acquisition cost when adjusting for inflation. Nvidia is certainly doing better, but only 20% or so of their stock price is due to consumer gpu sales. So extremely high startup costs + low estimated ROI = no starter. And on top of everything else, intel is still lurking in the background and will continue to nip at amd/nvidia's heels for the forseeable future.

 

[bryanW1995]

I see no problem.

The musician can perform live. The musician can do a lot of things that could make them money, and it's the publishers who gain the most from IP, not the artist. Getting rid of IP would definitely destroy the big publishers more than it would the artist.

Wow. So where's the incentive for intel to spend eleventy billion dollars designing new cpus if Bill Gates could just hire a couple of their top engineers and steal the entire design with zero consequences? What you propose would do more than just run a few publishing houses out of business, it would create, um, well, anarchy. You have a very appropriate name.

 

[Anarchist420]

Please stop posting this FUD. It costs 100s of millions to billions of dollars to research, develop, etc... What profits are you going to generate after a couple of days of selling a new item when it gets copied? The company would go out of business b/c it wouldn't be able to generate any money to cover their expenses and you would have zero innovation. Only a clueless person like you would think 1000s of people would want to work for free to develop a product so that someone else could copy and sell for less money.

Get the hell out of the basement and learn how the real world works b/c you obviously don't have a friggin clue.

Between going back to the gold standard and this crap, the world would be set back 100 years by kids living in their basement. Are you parents seriously feeding you this crap? Again, how old are you?

They wouldn't be working for free if they were creative enough. And it doesn't always take 100s of millions of dollars to develop something. In fact, it rarely does. A lof of research is tax payer funded, so they waste it since it's not the researcher's money.

Just like I thought. You have no real response except to say that the 'market' will (magically) take care of everything. Austrians are stuck in the past and have no understanding of modern industry.

You say the person whose product was copied would be motivated to invent something new but the reality is that the person (or graphics corp) would be demotivated instead. You also assume such a person would have the resources after losing their business to copycats to have the free time etc to indulge themselves and freebies for others.

You want more intangibles? Odd since global trade in intangibles like services/ finance instruments have grown much greater than the manufactures of 'real stuff' in recent decades yet inequality has grown greater, poverty has increased, economies have been less stable. The current view is to go back to tangibles. I agree that agriculture would be a good source of employment if it weren't monopolized by big corps.

Bur why is agriculture monopolized? Because of the government of course. The wealthiest 10% of agriculture firms get >70% of subsidies, and the regulations including, but not limited to regulations on raw milk is why agriculture is monopolized. They have all of these regulations, which are lobbied for by big corporations, and if you take away the FDA from them, then they can't lobby for them under the guise of public safety. Since when does the FDA care about our safety?

As for finance instruments, that's because of the state created banking cartel. That would make a lot less money and make people more productive if the supply of credit was always stable or decreasing.

 

[Anarchist420]

Wow. So where's the incentive for intel to spend eleventy billion dollars designing new cpus if Bill Gates could just hire a couple of their top engineers and steal the entire design with zero consequences? What you propose would do more than just run a few publishing houses out of business, it would create, um, well, anarchy. You have a very appropriate name.

It may not take them elenty billion dollars to design new CPUs. Don't they have a really expensive CEO and an expensive legal team?

 

[TemjinGold]

They wouldn't be working for free if they were creative enough. And it doesn't always take 100s of millions of dollars to develop something. In fact, it rarely does. A lof of research is tax payer funded, so they waste it since it's not the researcher's money.

Really? Intel uses money from tax payers? Where are you getting that from? So let's make this simpler for you:

You spend a lot of money (let's say only $1B to humor you.) WITH IP protection, your great ideas may recover this money in 20 years. WITHOUT IP protection, you will sell products "for a few days" before your idea is stolen and you need to "get more creative." Here's the problem: How much did you make in those "few days"? Because "getting more creative" isn't free. It requires you to sink more money in than you already have. Do you know what else you can do with $1B instead of throwing it into this ridiculous gamble? Do you now see why no one would ever risk large sums of money if there was no IP?

If you really don't understand still, go start a GPU company. Please do.

 

[TemjinGold]

It may not take them elenty billion dollars to design new CPUs. Don't they have a really expensive CEO and an expensive legal team?

Are you suggesting you want to start a company with no CEO and no legal team? Your $6B won't last more than a couple of lawsuits in that case...

 

[bononos]

They wouldn't be working for free if they were creative enough. And it doesn't always take 100s of millions of dollars to develop something. In fact, it rarely does. A lof of research is tax payer funded, so they waste it since it's not the researcher's money.

Bur why is agriculture monopolized? Because of the government of course. The wealthiest 10% of agriculture firms get >70% of subsidies, ...... safety. Since when does the FDA care about our safety?

As for finance instruments, that's because of the state created banking cartel. That would make a lot less money and make people more productive if the supply of credit was always stable or decreasing.

You're taking the extreme position by stating that if govt doesn't work as well as it should then we should ideally have no govt at all and let the free market do all the work by deregulating agriculture/ banking/ finance/ everything. But the thought of cleaning up govt is anathema to neoliberal types and Austrians. If bleeding a vein doesn't cure the disease then bleed 2 veins and then go on to 3-4-5 if that doesn't work.

You didn't respond to the problem of IP infringements in real world examples above except to appeal to neoliberal axioms that the free market will cure everything.

You fail to understand that any work in the modern gpu industry is not executed single handedly like IronMan making his heart-type-thing, the myth of heroic CEO leadership plays into the neoliberal/conservative theme of capitalists 'entrepreneurs' running everything when in fact big corps work (and run better) like worker ants - to repeat my earlier post about knowledge and work embodied in machines and finished products vs the pre-industrial order where skilled workers were the indispensible machines which Austrian theories might make sense.

 

[TemjinGold]

OP: To answer your original question, the GPU industry is so expensive to enter because the world doesn't work the way you think it does. I think that's the simplest answer anyone can give you.

To be fair, I know you wouldn't hesitate to take a lower wage for the greater good and you wouldn't mind getting your ideas stolen for the overall betterment and advancement of technology. However, do realize that you are in the very small minority with such views as far as the whole world is concerned and such views require adoption by the vast majority to become successful.

Most people cannot afford to adopt your views because they have families to feed and livelihoods to live. They cannot take the chance that "everything will just work out." That's why IPs exist and that's why it is difficult to compete with AMD/nVidia.

If you do start a GPU company, you cannot legally hire AMD/nVidia's current people, so you are limited to hiring people who do not work for either. Add to the fact that you plan on paying them frugally and you will not find anyone competent. Also, if you've never run a large company before, know that the cost of labor doesn't end at the grunts. You need competent managers, directors, etc. all the way up to the CEO or your company will fall flat before it starts. Trust me on this one: The myth that management does nothing but take a big paycheck is very wrong. I know because I am one and I can tell you we do an insane amount of work and carry an insane burden behind the scenes that the grunts never even realize exists. Great management is the difference between stellar work and horrible work. Now throw in PR folks, HR folks, a legal team, an accounting team, and you begin to see that your $1B doesn't even come close to covering salaries for a few years, let alone all the other stuff you plan to spend on.

The last problem you face is that if you do put your ragtag team together today, you need to start work not on something that beats a GTX590/HD6990, you need to start work on something that will compete well with whatever AMD/nVidia will throw out in 2017 or so. Your ragtag team has not worked for AMD/nVidia recently, so they have no experience in current tech design, much less what it would take to compete in the future.

Now you need a way to pay people and keep things running for years while your design gets created. Using your model of no IPs, you would be lucky to sell 1 million cards at $200 each (realistically, you will sell a lot less even with IPs because you are new) in your few days. That's $200 million GROSS (not profit). Even if you spent $1B (which is really underestimating), you are sitting on a huge loss (we haven't even thrown in packaging costs, marketing, etc.). Your plan now is to "go get more creative to stay in the game." That means throwing more money into the pit before you can come up with something to sell for another $200 million tops.

Do you see why this is a losing proposition and why most people would not bother?

I know you fancy yourself as altruistic for the betterment of society. I hate to break it to you but humankind as a whole is not. We are a selfish being who primarily looks out for ourselves. Until that changes, which it never will, your vision of the world will never come to being.

 

[Anarchist420]

Really? Intel uses money from tax payers? Where are you getting that from? So let's make this simpler for you:

You spend a lot of money (let's say only $1B to humor you.) WITH IP protection, your great ideas may recover this money in 20 years. WITHOUT IP protection, you will sell products "for a few days" before your idea is stolen and you need to "get more creative." Here's the problem: How much did you make in those "few days"? Because "getting more creative" isn't free. It requires you to sink more money in than you already have. Do you know what else you can do with $1B instead of throwing it into this ridiculous gamble? Do you now see why no one would ever risk large sums of money if there was no IP?

If you really don't understand still, go start a GPU company. Please do.

Thanks for the kind reply
Sorry for the misunderstanding on the tax payer money comment. I meant that a lot of research in general, like that done at Harvard I didn't mean to imply that Intel necessarily took tax payer money.

I'm not an expert on the subject, so it could very well take a lot of money, although it's impossible for myself to accurately know what percentage of the difficulty comes from the market vs. legislated barriers to entry unless I were to actually try it, but as you pointed out, I couldn't because I just wouldn't be motivated and aggressive enough. The main reason I tend to think that entering the GPU market takes not as much money as it does now, is because of all the competition there was in the 90s and that nvidia actually benefitted from using 3dfx's IP (which I applaud nvidia for and they possibly wouldn't be around today if they hadn't used the idea of dual texture units).

I agree that the heroic CEO is rare (at least in America), and that the one who isn't aggressive and isn't driven to make money is the one who usually doesn't do as well. I wouldn't be a good CEO because I don't have the qualities of a typical American CEO as you pointed out I do think that in a free market, the CEO would be a position in which less aggressive people would prosper more than now, but I'll probably never know for sure.

 

[Ben90]

The main reason I tend to think that entering the GPU market takes not as much money as it does now, is because of all the competition there was in the 90s

The main reason I tend to think that entering the GPU market takes as much money as it does now, is because of how little competition there is in the 2010s

 

[TemjinGold]

OP: You are right in that it was much easier to enter the market in the old days. Here's a simple analogy to help you see why that isn't the case anymore:

If you were thinking about entering the weapons industry in prehistoric times, the cost of entry is tiny because you would be making clubs and stones. If you were to try to get in today though, it would be a bit more expensive...