On Mercury news, there's a long interview with nVidia's CEO, Jen-Hsun Huang. You might remember him from E3 2005 where he told us more about the RSX in ps3. It doens't give much info about the rsx, but he does say that rsx has been in production for quite some time now. All in all, an interesting read imo 
. I highlighted the questions and some important quotes.
link: http://blogs.mercurynews.com/aei/2006/07/nvidia_ceo_an_e.html
. I highlighted the questions and some important quotes.
What are some things to know about you?
My priorities are my family and my company and that's it. My perfect day is spending time with just my family, deciding what to cook for dinner, going grocery shopping, cooking dinner and opening a bottle of wine. My daughter goes downstairs and picks out a bottle of wine. She has wonderful taste. We all cook dinner and eat in front of a movie. I'm incredibly proud of my kids. I'm glad they are wonderful people and I'm glad they can do something -- as black belts -- beyond anything I've been able to do. They can be so good at their karate. The most important thing to know about a CEO is you can't be successful unless your family and in particular your wife wants you to be successful.
You came to the U.S. as an immigrant at age nine from Taiwan. You had a harder start in life than a lot of people. How much do you think about that experience today?
Part of being successful is having the ability to make sacrifices. To endure a lot of hardship for something you believe in and having a lot of patience. Those are all elements of being successful. When you are an immigrant to this country, you have to endure some hardship. When I came to the United States, my older brother -- who was ten years old -- and I were the only Chinese people Kentucky had ever seen, it seemed. I know we were the only Chinese people that Oneida, Kentucky, had ever seen. We lived in Tacoma, Wash., for a few months. Even then, we looked like foreigners. You have to endure some amount of individualism, be self confident, and be comfortable with yourself. Growing up as an immigrant, It was not easy back in the good old days.
Does that make you a different kind of CEO?
Sacrifice, enduring challenges, seeing obstacles as an opportunity. Being able to deal with adversity. Those are characteristics that all CEOs need to have. I think I handle those things very easily and very comfortably. I make sacrifices easily. I'm impatient about doing great things and building a great company.
What makes it to the list of hardships you have to deal with now?
Our company is doing really well. We are building great products. You know I've always believed that graphics is one of the most important technologies for this decade. It is very evident that the graphics processing unit is becoming important in all kinds of devices whether it's game consoles or cell phones and what not. More and more of the innovation we do is at the intersection of computing technology and consumer applications. More of those applications relate to multimedia and to graphics. The growth is there. Our strategic relevance is higher than ever before. We are organized better than ever before. We have strong leadership. I don't know if we have an immediate adversity. We have challenges. How do you go from a $3 billion company to a $15 billion company, where your employees are spread out all over the world, where you're building products for every industry? From PCs to consumer boxes, to automobiles. We're building technologies for a lot of different markets and different industries. Scaling the company is a big challenge. It is a cultural, management and leadership challenge.
It seems like where you have gotten to has almost been accidental. Do you look at it that way? Will the graphics chip always be in the personal computer? Will all of the chips in the computer merge into one chip? For a decade, you've been able to grab more than your fair share of computing power inside the PC. How accidental is this?
Serendipity plays a role in all successful companies. We had more than our fair share of serendipity. Of course, people say that serendipity comes to those who work the hardest and are the smartest. We work incredibly hard. But I would say that the question about whether graphics being integrated into other chips roots itself in the idea that graphics is good enough. We were confronted with that question with GeForce 256 in 1999. That was on par with the Reality Engine supercomputer that Silicon Graphics used to sell. The fill rate, the performance, the throughput of the chips was amazing. In 1999, we were confronted with that question. I thought we made one of the most important decisions for not only our company but the entire industry. That was going to programmability. As you know, investing in programmability, which ultimately ended up in GeForce FX, made the graphics chip far more programmable. But the cost was so high. It was quite a burden on our company for a couple of years. I would calculate that the amount of profitability lost on GeForce FX was $200 million a year for two years. It was almost $500 million in lost profit. That was a half a billion dollar risk we took going to a completely programmable model. Our vision is simple. Fixed function graphics accelerators will limit the long term growth of the industry. There are so
many things you can do with a fixed function device. If I give you a Ginsu knife, and that was all you could do, what kind of artistic things can you do with it? Going to an infinitely programmable device, we made graphics essentially into an infinite palette for artistic expression. Look at all of the applications that we are seeing now. It's possible to use this graphics processor to express all different kinds of graphics, from toon shading to all kinds of film quality shading. You can do X-Ray shading and all kinds of amazing looks you couldn't do before. We invented, if you will, the paint brush. And now computer programmers can use this computer paint brush to articulate anything. That opened the horizon for us to really grow into what we are today. The GPUs are becoming increasingly complex, increasingly flexible. There is so much more that the software programmers want us to capture.
Is there a collision you can foresee with the Intel microprocessor because you made the move to programmability?
Programmability has different types. There are scalar programs. That uses a scalar microprocessor with a flow of instructions and it fetches instructions out of a cache. It processes data in a data-dependent way. That sort of programming is what microprocessors are really wonderful at. We are not very good at that kind of processing. Our processors are adept at processing large amounts of data that have less dependency. Our processor is more akin to a stream processor. The types of architectures are radically different. Just as the CPU can run DSP programs, a DSP is much better at running DSP programs. There are different types of programming models, whether it is signal processing for baseband, or voice. There are scalar processors. There are image processors for enormously large data sets which is what a GPU does.
If Moore's Law gives you unlimited numbers of transistors, a few years from now you could put a CPU in the corner of your chip, and Intel could put a graphics chip in their CPUs. Why or why not would this happen?
There is integration at two levels. There is the unification of processing models. There is the CPU and the GPU, combined together in a unified processor model. I think the latter is very unlikely. Although on balance, transistors are free, we are challenged because most of the opportunities require low power. So you have to have efficient programming. It is far more efficient to run a program written for CPUs on a CPU, and it's far more efficient to run a program for GPUs on a GPU. There is the issue of power efficiency and cost efficiency. Brute force is not a very good option. There is the second approach of combining two processors onto one chip. In some markets, that would happen. For example, integrated graphics combines two chips into one where the technology is not very demanding. The market requirements are much slower in commercial, corporate desktops and others that require very little graphics. But if the graphics technology is a defining part of that system, whether it is a game console or high-end PC or workstations, the two devices innovate at different rhythms. There is no reason the two devices want to merge into one in that case. In fact, combining them into one makes it very difficult to combine two modern cores into the same substrate on the same schedule. There, what causes the two to move apart is not difference in programming models but differences in market requirements and rhythms. By putting it in one
chip, you end up getting the worst of both worlds.
So you are stuck with Intel and they are stuck with you?
Our primary focus is to advance the GPU. In the PC market, we focus on the marketplaces that value the the contributions we make. Workstations, the gaming platforms, multimedia machines, media centers. That's where graphics matters. We don't participate in all parts of the PC industry. We focus on market niches where we can add value. In the other markets, we are connected to the PowerPC/Cell microprocessor in the PlayStation 3. In handhelds, we are connected to SH microprocessors, Arm microprocessors. For embedded applications, we are connected to PowerPCs and what not. Our focus is the GPU. With respect to the CPU, we try to be as agnostic as we can. We focus on the marketplace.
As your company becomes bigger, is it harder to find the growth in graphics alone?
Someday, that is going to happen. But it's just not happening right now. The thing we talked about earlier, more than ever, graphics processing is more important in all of these digital devices. As the number of digital devices explodes, we find ourselves relevant in a lot of new markets that we were never part of before. You could also think of the importance of GPUs by just thinking about the number of the liquid crystal displays around the world and the quality of these displays. That's what is driving the consumption of GPUs.
Do you think you picked the right horse in the video game war this time?
You can't build chips for all the game consoles. That's not possible. They would all like a slightly different style from the others. Difference is important. The same chip company would have difficulty designing chips for the different styles. It's also so high stakes that you need to focus. No one has enough extraneous resources around to build chips for all the game consoles. You have to build one or so at a time. In a lot of ways, they also pick you. Sony picked us and Microsoft didn't. It's not so much we don't pick the horses. I don't think that working with Sony is wrong. There is no way that is going to be wrong. There are many wonderful things that Sony did. I'm excited that they made Blu-ray high-definition storage as a standard part of the PlayStation 3 platform. The first PlayStation had a CD-ROM drive. The PlayStation 2 had DVD. It makes no sense for the PlayStation 3 to use DVDs. To postpone it by a few months so they could include Blu-Ray was a master stroke. When that comes out, it's going to look so much more advanced than last-generation game consoles. I think that was a wonderful call on their part.
When you look back on your relationship with Microsoft on the Xbox, did it serve your purpose of getting into the game console business?
I always felt it was inevitable we would work on consoles. We invest $750 million a year in R&D in graphics processing. No other company invests that much in graphics processing today. This is such an incredibly deep technology and there is so much more to do. It makes sense that in the long-term we would work on game consoles as well. The others can't keep up with the R&D that we do. That part makes perfect sense to me.
Are the graphics in games as good as movies now?
We are far from movies. People use "Toy Story'' as a standard. But Toy Story is an animation. We know that every movie is 3-D these days. "Star Wars Episode III: Revenge of the Sith'' was all 3-D except for the human characters. The actors were standing in front of blue screens the whole time. People didn't realize they were watching an animated movie the entire time. With "Superman Returns,'' you can't really put a camera on a person and have him fly through a metropolis. That entire movie was animated. It was one big computer-generated movie with a guy in tights in front of a blue screen. You look at that imagery, and you know we are nowhere near that level of imagery.
Where do you want to see graphics go?
I would like to see it go in a couple of directions. I would like it to be easier to access. Games are so hard. They're intimidating to get into. The controls are pretty complicated. These are complicated games with complex story lines. I would like to see simpler, episodic games where they could let you play for 30 minutes at a time. They could show it to you 30 minutes at a time and you could play a little bit of it for a long period of time. And if you want to make episode games that have more storytelling, you're going to have to make the games more realistic. Right now the graphics is simply not at a level where you could use it to tell human drama or some kind of a deep story without facial articulation and good environmental effects to pull off the story telling. I wouldn't mind it if the business model of games was more episodic and I wouldn't mind it if video games were more realistic.
Are photorealistic games just around the corner?
We are a good solid 10 years away from photorealism. In the next several years, we will still just be learning to do the basics of film, like motion blur, depth of field -- all of that stuff alone chews up a lot of graphics processing. We're pretty excited about moving to high-dynamic range where the color system has the fidelity of what we see in real life. The images don't seem realistic yet. Articulating a human form and human animation, the subtlties of humans and nature, are still quite a ways away for us.
What is exciting about putting graphics into cell phones?
Today, a cell phone is a phone. When the PC industry first started out, it had a keyboard and a monitor. People asked why do you need a $3,000 device to replace a typewriter. And so, clearly, the PC is much more than a typewriter. In the long term, people will discover the cell phone is much more than a phone. We will discover it is one of our most important personal computing devices. We just compute with it in a different way than we compute with a computer or a game console. But it's a computer nonetheless. And these computers are going to be amazing. The video is going to have graphics and video. It's going to have all of the same virtues that you have in your PC today in a smaller form factor. The use will be different. It will be a lot more services and a lot less applications. Shorter enjoyment. The applications will be applets. The experience will be just as rich. All of that kind of innovation related to multimedia is what we are about.
How are you betting with your engineering team these days?
That's a good question. If you are spending three quarters of a billion dollars in R&D, you're investing that much. Our managers are really fund managers who are investing three quarters of a billion dollars each year. If the return on investment is not better than what the market can do by itself, the company is not going to do well over time. Where are we betting? We are betting a lot on the cell phone, believing that multimedia will be an important part of these devices. With GPUs, we are betting a huge amount on programmability. GPUs are going to become more programmable, easier to write programs, and support much more complex programs. For graphics, of course, not for running a spreadsheet.
What do you think of Intel's troubles these days, given how big it has become? What lesson is there for you as your company becomes bigger?
I don't understand their internal challenges. From outside the company, all great companies that have a winning hand will run into the same thing. Pentium 4 had a wonderful architecture and was an incredible market success. But the obvious thing to do with it was to make it a better Pentium 4. They took a good idea and kept scaling it. And over time, it stopped scaling. There are many products that eventually don't scale indefinitely. What they experienced was no different than what any other successful company would experience. You play the same hand over and over and eventually it stops working for you. In their case, power got in the way and the movement toward portability got in the way.
What do you observe on the sidelines of the Intel-AMD competition?
Intel is large and they can build an entire platform. AMD's advantage is that they are small, nimble and very focused. They can exploit the seams that a large company leaves behind. That's what they did. They exploited the fact that the enthusiast and video game market for PCs was still very large and performance still mattered. They collaborated with people that can help them build entire platforms like us.
Can you exploit the competition between them?
We can support both processors and always have. I don't know that we exploit the competition between them. We have plenty of competition ourselves. The advantage of Intel is they can innovate across the entire platform. The advantage of AMD is they have an open platform and they co-opt the entire platform.
What games are you playing?
I'm waiting for Call of Duty 3. I'm not playing anything right now. I'm playing Nvidia.
It's Nvidia 3.0.
What was Nvidia 1.0?
Nvidia 1.0 was building 3-D graphics. Creating the consumer 3-D market. That was Nvidia 1.0. Discovering and creating strategies that made it possible for us to leave the other start-ups behind. Nvidia 2.0 was expanding into other segments of the PC market. We focused intensely on the desktop PC market. We moved into notebooks and workstations. It ended when we got into the original Xbox. It was still a PC architecture, but it was for a completely different market. Nvidia 3.0 is about taking the GPU into completely different markets. From PCs to consumer electronics to handsets. Expanding it into markets where graphics matters.
What will Windows Vista do for you?
If you look at the simple math of it, the number of GPUs that can run Windows Vista well, DX 9 and beyond, and the number of integrated graphics chips that are currently shipping, is far less than 50 percent of the market. Let's say it is 25 percent of the market. But we know in two years, Vista will be 100 percent of the market. So what that says is we have to from where we are today, where integrated graphics is not capable of running Windows Vista Premium, to a point where Vista will be 100 percent of the market. That will prompt us to think there will be a surge in demand for GPUs. The question is how long that surge will last: two years or four years? It's hard to say. One of the secondary dynamics of Windows Vista is that very shortly after Windows Vista launches, you'll see all sorts of applications that have gone completely 3-D. Excel is going to go 3-D. Everything goes 3-D. Then there will be more consumption of GPU capabilities. That will prompt a second wave of consumption for GPUs.
Is it going to be a tough road to get there?
Whether it is GPUs or integrated graphics, it's not entirely clear. The more the operating system uses graphics, the better it is for us. In the beginning we'll see a surge of demand for GPUs, but then after that we will see a surge in demand for integrated graphics. We'll participate in both markets.
Doesn't it make sense for Intel to go deeper into graphics with stand-alone GPUs if Vista is going to do that for the market? And doesn't it make sense for you to go deeper into integrated graphics chip sets?
I think so. I don't know what their plans are. You will have to ask them. There is a market for GPUs obviously. This is a market we are intimate with and we're very good at. The R&D for GPUs is no longer insignificant. Although we are a much smaller company, the number of engineers who are world class at building world-class GPUs is not insignificant. Whatever decision they make is going to be a big one.
And for you to go in the other direction?
Well we build integrated graphics for AMD today. We constantly evaluate integrated graphics
for Intel. We need to focus on where we can add value.
What do you think of the health of Silicon Valley?
Jobs are great. Too great. It seems like Silicon Valley went through a slump and it is doing fabulously. Restaurants are getting filled up again. Not as crazy as it once was during the Internet boom. For a while there, restaurant owners were bringing in more high-end foods because these entreprenuers wanted the best possible cuisines. It's much more normal now. We are more mature about the Internet boom. There is still a lot of innovation going on. There are still a lot of start-ups. That's great to see.
How badly do your kids beat you in games?
Completely badly. It doesn't matter what game we play. Both kids beat me. It's particularly embarrassing when Madison beats me.
What is their game of choice?
The first-person shooter, Counter-Strike Source. Madison and Spencer both like Source.
For a while, the cadence of graphics chips was to move at twice the speed of Moore's Law and get a new chip out every six months. Is that the schedule still?
The high-end GPU still comes out every six months. We announced the 7900 GTX a year ago and six months later the GX2. You do have to offer a new way of bringing graphics chip to the market every six months and we haven't seen that change. We have done that in a multitude of ways. We have done that with SLI. We have done it with SLI on a single card we call GX 2. Between the combination of all these components, we still increase performance by a factor of two every six months.
What does SLI mean to the market where you can get more graphics chips into each PC?
SLI is probably one of the most important innovations that we have brought in the last several years. It's an enormous brand now. It's a community. People who have GPUs with SLI, motherboards with SLI, PCs with SLI. You can go on eBay and buy SLI stuff. It's a community and an ecosystem unto itself now. The thing that resonates with the market on several fronts. You could never get enough graphics performance. These large 30-inch displays with 2500 x 1900 resolution. That's a lot of pixels. That's four million pixels. You're going to need a pretty hefty GPU to drive it. Those four-million pixel displays are beautiful. You have these high-resolution screens and you never have enough performance. The new generation of gamers that really require self-expression, like people buying the Toyota Scion and modding it. This includes women. Car modders. Myspace. It's about self-expression. What SLI tapped into was to allow people to upgrade and configure their PC to their personalities.
How do you look at these high-end gamer PC companies that are selling a $5,000 custom paint job for a PC? Voodoo can sell a $10,000 personal computer?
The people they sell to are enthusiasts. They aren't gamers. They are buying the Mercedes SL 50 AMGs. It is an after-market modification. There are all kinds of car companies building after-market enhancements. That is what these enthusiast PC companies are doing. It suggests that the PC will follow home electronics in that way. There is enthusiast hi-fi equipment and there are radios and stereos. There are people out there who want the best and can pay for the best. That market for PCs is developing.
What's the secret to marketing to those folks and hanging onto them at the same time you are trying to go mainstream?
Nvidia has, if you look at our company and contrast it with Via or SiS, arguably ATI, we are the company that people who want to build high image, high impact, high brand come to. If you want to build the most leading-edge workstation or media center or notebook or game PC, or PlayStation 3, or a 3G cell phone from Motorola. They see us as the technology company that can help them create something very new, high image, high impact, and high brand. That is what Nvidia is. We do that through innovation and R&D. But we have to bring new ideas. The invention of the programmable shader is one of those ideas. The invention of SLI is one of those ideas. The invention of PureVideo HD is one of those ideas. You have to take risks and you have to be willing to make mistakes and build something pretty amazing.
How do you look at ATI?
They are a formidable competitor. They have wonderful people. It's the same way I've seen them before.
Given how the Xbox 360 turned out, did you have any regret about not winning the graphics chip for that console?
Not at all. We could not afford to build the graphics for the 360. Our most important asset is our people. If we use our people on a project where the economic return is not good enough, and there are other projects we could be working on, then we're going to lose money. We were a lot smaller company than ATI at the time. Maybe ATI could afford it and we couldn't. I know I couldn't afford it. I would love to build it. I just can't afford it.
ATI is excited about unified shaders. If you pull back, how do you see if your people are making the right decisions?
For each one of our generations, we need to have a vision of what we want to do. It costs hundreds of millions of dollars to come up with a new architecture. I don't know what they've built for next generation. It comes down to a different system vision for what we are trying to achieve.
Are you manufacturing the RSX for Sony now?
It is in production. It has been for some time.
Do you have a PlayStation 3 in your home?
Not yet. I hope I get one of the first ones.
Does it look like it's on schedule?
Sony hasn't changed their schedule. I think that's the most important thing. I thought it was a master stroke that they did.
Everyone criticized them for falling behind and having a high price and costs as high as $900. Why was it a master stroke?
PlayStation 2 was launched seven years ago in Japan at about $399. If you use inflation, it's the same price, approximately. The important thing is you cannot announce a game console for the next ten years and not have Blu-Ray. It's an impossible scenario. I think they got that perspective right. The moment we put those consoles together it's going to be very clear. If I'm going to buy a next-generation game console, I'm going to buy a console with next-generation media. It's going to last 10 years.
Two out of three of the players have bet you don't need it. Nintendo has bet you don't even need HD for the next five years.
Nintendo's perspective has always been different. The platforms that are being built now are not just game consoles. You use it for all kinds of other kinds of applications. In the case of Nintendo, they wanted to build a game console. They built a wonderful game device. Their focus is games and enjoyment. They will be myopically focused on that. I think that is terrific. Their perspective is different from the other two. If I'm going to buy a next-generation game console, I'm going to want next-generation media.
Do you think Microsoft planned for a five-year cycle and that Sony planned for a ten-year cycle?
I'm not sure how Microsoft is going to do in this transition. They are clever and they will figure out a way. I'll make a prediction that Xbox 360 can't possibly be a DVD-only device by Christmas of next year.
They will modify it?
I don't know how they will do it. But I just can't imagine going to a store and saying that this console has a Blu-Ray and this one has DVD. Remember Dreamcast?
Do you think they will go further than the HD-DVD accessory that they have planned?
I don't know how they will do that. Then they will have two platforms. You have the Xbox 360 with DVD. That is what Xbox 360 will mean. If you want to play HD, you have to buy an accesory? If that's the case, I'll buy an HD player. I know Microsoft will come up with clever ideas. These are the challenges. Every Xbox 360 they make in advance of that decision makes that decision that much harder. You could have two Xbox 360s. It's a very tough strategic challenge. Sega had the same strategic challenge when they launched ahead of Sony on PlayStation 2. The executives at Sega are very smart and the Dreamcast was a very good machine.
My priorities are my family and my company and that's it. My perfect day is spending time with just my family, deciding what to cook for dinner, going grocery shopping, cooking dinner and opening a bottle of wine. My daughter goes downstairs and picks out a bottle of wine. She has wonderful taste. We all cook dinner and eat in front of a movie. I'm incredibly proud of my kids. I'm glad they are wonderful people and I'm glad they can do something -- as black belts -- beyond anything I've been able to do. They can be so good at their karate. The most important thing to know about a CEO is you can't be successful unless your family and in particular your wife wants you to be successful.
You came to the U.S. as an immigrant at age nine from Taiwan. You had a harder start in life than a lot of people. How much do you think about that experience today?
Part of being successful is having the ability to make sacrifices. To endure a lot of hardship for something you believe in and having a lot of patience. Those are all elements of being successful. When you are an immigrant to this country, you have to endure some hardship. When I came to the United States, my older brother -- who was ten years old -- and I were the only Chinese people Kentucky had ever seen, it seemed. I know we were the only Chinese people that Oneida, Kentucky, had ever seen. We lived in Tacoma, Wash., for a few months. Even then, we looked like foreigners. You have to endure some amount of individualism, be self confident, and be comfortable with yourself. Growing up as an immigrant, It was not easy back in the good old days.
Does that make you a different kind of CEO?
Sacrifice, enduring challenges, seeing obstacles as an opportunity. Being able to deal with adversity. Those are characteristics that all CEOs need to have. I think I handle those things very easily and very comfortably. I make sacrifices easily. I'm impatient about doing great things and building a great company.
What makes it to the list of hardships you have to deal with now?
Our company is doing really well. We are building great products. You know I've always believed that graphics is one of the most important technologies for this decade. It is very evident that the graphics processing unit is becoming important in all kinds of devices whether it's game consoles or cell phones and what not. More and more of the innovation we do is at the intersection of computing technology and consumer applications. More of those applications relate to multimedia and to graphics. The growth is there. Our strategic relevance is higher than ever before. We are organized better than ever before. We have strong leadership. I don't know if we have an immediate adversity. We have challenges. How do you go from a $3 billion company to a $15 billion company, where your employees are spread out all over the world, where you're building products for every industry? From PCs to consumer boxes, to automobiles. We're building technologies for a lot of different markets and different industries. Scaling the company is a big challenge. It is a cultural, management and leadership challenge.
It seems like where you have gotten to has almost been accidental. Do you look at it that way? Will the graphics chip always be in the personal computer? Will all of the chips in the computer merge into one chip? For a decade, you've been able to grab more than your fair share of computing power inside the PC. How accidental is this?
Serendipity plays a role in all successful companies. We had more than our fair share of serendipity. Of course, people say that serendipity comes to those who work the hardest and are the smartest. We work incredibly hard. But I would say that the question about whether graphics being integrated into other chips roots itself in the idea that graphics is good enough. We were confronted with that question with GeForce 256 in 1999. That was on par with the Reality Engine supercomputer that Silicon Graphics used to sell. The fill rate, the performance, the throughput of the chips was amazing. In 1999, we were confronted with that question. I thought we made one of the most important decisions for not only our company but the entire industry. That was going to programmability. As you know, investing in programmability, which ultimately ended up in GeForce FX, made the graphics chip far more programmable. But the cost was so high. It was quite a burden on our company for a couple of years. I would calculate that the amount of profitability lost on GeForce FX was $200 million a year for two years. It was almost $500 million in lost profit. That was a half a billion dollar risk we took going to a completely programmable model. Our vision is simple. Fixed function graphics accelerators will limit the long term growth of the industry. There are so
many things you can do with a fixed function device. If I give you a Ginsu knife, and that was all you could do, what kind of artistic things can you do with it? Going to an infinitely programmable device, we made graphics essentially into an infinite palette for artistic expression. Look at all of the applications that we are seeing now. It's possible to use this graphics processor to express all different kinds of graphics, from toon shading to all kinds of film quality shading. You can do X-Ray shading and all kinds of amazing looks you couldn't do before. We invented, if you will, the paint brush. And now computer programmers can use this computer paint brush to articulate anything. That opened the horizon for us to really grow into what we are today. The GPUs are becoming increasingly complex, increasingly flexible. There is so much more that the software programmers want us to capture.
Is there a collision you can foresee with the Intel microprocessor because you made the move to programmability?
Programmability has different types. There are scalar programs. That uses a scalar microprocessor with a flow of instructions and it fetches instructions out of a cache. It processes data in a data-dependent way. That sort of programming is what microprocessors are really wonderful at. We are not very good at that kind of processing. Our processors are adept at processing large amounts of data that have less dependency. Our processor is more akin to a stream processor. The types of architectures are radically different. Just as the CPU can run DSP programs, a DSP is much better at running DSP programs. There are different types of programming models, whether it is signal processing for baseband, or voice. There are scalar processors. There are image processors for enormously large data sets which is what a GPU does.
If Moore's Law gives you unlimited numbers of transistors, a few years from now you could put a CPU in the corner of your chip, and Intel could put a graphics chip in their CPUs. Why or why not would this happen?
There is integration at two levels. There is the unification of processing models. There is the CPU and the GPU, combined together in a unified processor model. I think the latter is very unlikely. Although on balance, transistors are free, we are challenged because most of the opportunities require low power. So you have to have efficient programming. It is far more efficient to run a program written for CPUs on a CPU, and it's far more efficient to run a program for GPUs on a GPU. There is the issue of power efficiency and cost efficiency. Brute force is not a very good option. There is the second approach of combining two processors onto one chip. In some markets, that would happen. For example, integrated graphics combines two chips into one where the technology is not very demanding. The market requirements are much slower in commercial, corporate desktops and others that require very little graphics. But if the graphics technology is a defining part of that system, whether it is a game console or high-end PC or workstations, the two devices innovate at different rhythms. There is no reason the two devices want to merge into one in that case. In fact, combining them into one makes it very difficult to combine two modern cores into the same substrate on the same schedule. There, what causes the two to move apart is not difference in programming models but differences in market requirements and rhythms. By putting it in one
chip, you end up getting the worst of both worlds.
So you are stuck with Intel and they are stuck with you?
Our primary focus is to advance the GPU. In the PC market, we focus on the marketplaces that value the the contributions we make. Workstations, the gaming platforms, multimedia machines, media centers. That's where graphics matters. We don't participate in all parts of the PC industry. We focus on market niches where we can add value. In the other markets, we are connected to the PowerPC/Cell microprocessor in the PlayStation 3. In handhelds, we are connected to SH microprocessors, Arm microprocessors. For embedded applications, we are connected to PowerPCs and what not. Our focus is the GPU. With respect to the CPU, we try to be as agnostic as we can. We focus on the marketplace.
As your company becomes bigger, is it harder to find the growth in graphics alone?
Someday, that is going to happen. But it's just not happening right now. The thing we talked about earlier, more than ever, graphics processing is more important in all of these digital devices. As the number of digital devices explodes, we find ourselves relevant in a lot of new markets that we were never part of before. You could also think of the importance of GPUs by just thinking about the number of the liquid crystal displays around the world and the quality of these displays. That's what is driving the consumption of GPUs.
Do you think you picked the right horse in the video game war this time?
You can't build chips for all the game consoles. That's not possible. They would all like a slightly different style from the others. Difference is important. The same chip company would have difficulty designing chips for the different styles. It's also so high stakes that you need to focus. No one has enough extraneous resources around to build chips for all the game consoles. You have to build one or so at a time. In a lot of ways, they also pick you. Sony picked us and Microsoft didn't. It's not so much we don't pick the horses. I don't think that working with Sony is wrong. There is no way that is going to be wrong. There are many wonderful things that Sony did. I'm excited that they made Blu-ray high-definition storage as a standard part of the PlayStation 3 platform. The first PlayStation had a CD-ROM drive. The PlayStation 2 had DVD. It makes no sense for the PlayStation 3 to use DVDs. To postpone it by a few months so they could include Blu-Ray was a master stroke. When that comes out, it's going to look so much more advanced than last-generation game consoles. I think that was a wonderful call on their part.
When you look back on your relationship with Microsoft on the Xbox, did it serve your purpose of getting into the game console business?
I always felt it was inevitable we would work on consoles. We invest $750 million a year in R&D in graphics processing. No other company invests that much in graphics processing today. This is such an incredibly deep technology and there is so much more to do. It makes sense that in the long-term we would work on game consoles as well. The others can't keep up with the R&D that we do. That part makes perfect sense to me.
Are the graphics in games as good as movies now?
We are far from movies. People use "Toy Story'' as a standard. But Toy Story is an animation. We know that every movie is 3-D these days. "Star Wars Episode III: Revenge of the Sith'' was all 3-D except for the human characters. The actors were standing in front of blue screens the whole time. People didn't realize they were watching an animated movie the entire time. With "Superman Returns,'' you can't really put a camera on a person and have him fly through a metropolis. That entire movie was animated. It was one big computer-generated movie with a guy in tights in front of a blue screen. You look at that imagery, and you know we are nowhere near that level of imagery.
Where do you want to see graphics go?
I would like to see it go in a couple of directions. I would like it to be easier to access. Games are so hard. They're intimidating to get into. The controls are pretty complicated. These are complicated games with complex story lines. I would like to see simpler, episodic games where they could let you play for 30 minutes at a time. They could show it to you 30 minutes at a time and you could play a little bit of it for a long period of time. And if you want to make episode games that have more storytelling, you're going to have to make the games more realistic. Right now the graphics is simply not at a level where you could use it to tell human drama or some kind of a deep story without facial articulation and good environmental effects to pull off the story telling. I wouldn't mind it if the business model of games was more episodic and I wouldn't mind it if video games were more realistic.
Are photorealistic games just around the corner?
We are a good solid 10 years away from photorealism. In the next several years, we will still just be learning to do the basics of film, like motion blur, depth of field -- all of that stuff alone chews up a lot of graphics processing. We're pretty excited about moving to high-dynamic range where the color system has the fidelity of what we see in real life. The images don't seem realistic yet. Articulating a human form and human animation, the subtlties of humans and nature, are still quite a ways away for us.
What is exciting about putting graphics into cell phones?
Today, a cell phone is a phone. When the PC industry first started out, it had a keyboard and a monitor. People asked why do you need a $3,000 device to replace a typewriter. And so, clearly, the PC is much more than a typewriter. In the long term, people will discover the cell phone is much more than a phone. We will discover it is one of our most important personal computing devices. We just compute with it in a different way than we compute with a computer or a game console. But it's a computer nonetheless. And these computers are going to be amazing. The video is going to have graphics and video. It's going to have all of the same virtues that you have in your PC today in a smaller form factor. The use will be different. It will be a lot more services and a lot less applications. Shorter enjoyment. The applications will be applets. The experience will be just as rich. All of that kind of innovation related to multimedia is what we are about.
How are you betting with your engineering team these days?
That's a good question. If you are spending three quarters of a billion dollars in R&D, you're investing that much. Our managers are really fund managers who are investing three quarters of a billion dollars each year. If the return on investment is not better than what the market can do by itself, the company is not going to do well over time. Where are we betting? We are betting a lot on the cell phone, believing that multimedia will be an important part of these devices. With GPUs, we are betting a huge amount on programmability. GPUs are going to become more programmable, easier to write programs, and support much more complex programs. For graphics, of course, not for running a spreadsheet.
What do you think of Intel's troubles these days, given how big it has become? What lesson is there for you as your company becomes bigger?
I don't understand their internal challenges. From outside the company, all great companies that have a winning hand will run into the same thing. Pentium 4 had a wonderful architecture and was an incredible market success. But the obvious thing to do with it was to make it a better Pentium 4. They took a good idea and kept scaling it. And over time, it stopped scaling. There are many products that eventually don't scale indefinitely. What they experienced was no different than what any other successful company would experience. You play the same hand over and over and eventually it stops working for you. In their case, power got in the way and the movement toward portability got in the way.
What do you observe on the sidelines of the Intel-AMD competition?
Intel is large and they can build an entire platform. AMD's advantage is that they are small, nimble and very focused. They can exploit the seams that a large company leaves behind. That's what they did. They exploited the fact that the enthusiast and video game market for PCs was still very large and performance still mattered. They collaborated with people that can help them build entire platforms like us.
Can you exploit the competition between them?
We can support both processors and always have. I don't know that we exploit the competition between them. We have plenty of competition ourselves. The advantage of Intel is they can innovate across the entire platform. The advantage of AMD is they have an open platform and they co-opt the entire platform.
What games are you playing?
I'm waiting for Call of Duty 3. I'm not playing anything right now. I'm playing Nvidia.
It's Nvidia 3.0.
What was Nvidia 1.0?
Nvidia 1.0 was building 3-D graphics. Creating the consumer 3-D market. That was Nvidia 1.0. Discovering and creating strategies that made it possible for us to leave the other start-ups behind. Nvidia 2.0 was expanding into other segments of the PC market. We focused intensely on the desktop PC market. We moved into notebooks and workstations. It ended when we got into the original Xbox. It was still a PC architecture, but it was for a completely different market. Nvidia 3.0 is about taking the GPU into completely different markets. From PCs to consumer electronics to handsets. Expanding it into markets where graphics matters.
What will Windows Vista do for you?
If you look at the simple math of it, the number of GPUs that can run Windows Vista well, DX 9 and beyond, and the number of integrated graphics chips that are currently shipping, is far less than 50 percent of the market. Let's say it is 25 percent of the market. But we know in two years, Vista will be 100 percent of the market. So what that says is we have to from where we are today, where integrated graphics is not capable of running Windows Vista Premium, to a point where Vista will be 100 percent of the market. That will prompt us to think there will be a surge in demand for GPUs. The question is how long that surge will last: two years or four years? It's hard to say. One of the secondary dynamics of Windows Vista is that very shortly after Windows Vista launches, you'll see all sorts of applications that have gone completely 3-D. Excel is going to go 3-D. Everything goes 3-D. Then there will be more consumption of GPU capabilities. That will prompt a second wave of consumption for GPUs.
Is it going to be a tough road to get there?
Whether it is GPUs or integrated graphics, it's not entirely clear. The more the operating system uses graphics, the better it is for us. In the beginning we'll see a surge of demand for GPUs, but then after that we will see a surge in demand for integrated graphics. We'll participate in both markets.
Doesn't it make sense for Intel to go deeper into graphics with stand-alone GPUs if Vista is going to do that for the market? And doesn't it make sense for you to go deeper into integrated graphics chip sets?
I think so. I don't know what their plans are. You will have to ask them. There is a market for GPUs obviously. This is a market we are intimate with and we're very good at. The R&D for GPUs is no longer insignificant. Although we are a much smaller company, the number of engineers who are world class at building world-class GPUs is not insignificant. Whatever decision they make is going to be a big one.
And for you to go in the other direction?
Well we build integrated graphics for AMD today. We constantly evaluate integrated graphics
for Intel. We need to focus on where we can add value.
What do you think of the health of Silicon Valley?
Jobs are great. Too great. It seems like Silicon Valley went through a slump and it is doing fabulously. Restaurants are getting filled up again. Not as crazy as it once was during the Internet boom. For a while there, restaurant owners were bringing in more high-end foods because these entreprenuers wanted the best possible cuisines. It's much more normal now. We are more mature about the Internet boom. There is still a lot of innovation going on. There are still a lot of start-ups. That's great to see.
How badly do your kids beat you in games?
Completely badly. It doesn't matter what game we play. Both kids beat me. It's particularly embarrassing when Madison beats me.
What is their game of choice?
The first-person shooter, Counter-Strike Source. Madison and Spencer both like Source.
For a while, the cadence of graphics chips was to move at twice the speed of Moore's Law and get a new chip out every six months. Is that the schedule still?
The high-end GPU still comes out every six months. We announced the 7900 GTX a year ago and six months later the GX2. You do have to offer a new way of bringing graphics chip to the market every six months and we haven't seen that change. We have done that in a multitude of ways. We have done that with SLI. We have done it with SLI on a single card we call GX 2. Between the combination of all these components, we still increase performance by a factor of two every six months.
What does SLI mean to the market where you can get more graphics chips into each PC?
SLI is probably one of the most important innovations that we have brought in the last several years. It's an enormous brand now. It's a community. People who have GPUs with SLI, motherboards with SLI, PCs with SLI. You can go on eBay and buy SLI stuff. It's a community and an ecosystem unto itself now. The thing that resonates with the market on several fronts. You could never get enough graphics performance. These large 30-inch displays with 2500 x 1900 resolution. That's a lot of pixels. That's four million pixels. You're going to need a pretty hefty GPU to drive it. Those four-million pixel displays are beautiful. You have these high-resolution screens and you never have enough performance. The new generation of gamers that really require self-expression, like people buying the Toyota Scion and modding it. This includes women. Car modders. Myspace. It's about self-expression. What SLI tapped into was to allow people to upgrade and configure their PC to their personalities.
How do you look at these high-end gamer PC companies that are selling a $5,000 custom paint job for a PC? Voodoo can sell a $10,000 personal computer?
The people they sell to are enthusiasts. They aren't gamers. They are buying the Mercedes SL 50 AMGs. It is an after-market modification. There are all kinds of car companies building after-market enhancements. That is what these enthusiast PC companies are doing. It suggests that the PC will follow home electronics in that way. There is enthusiast hi-fi equipment and there are radios and stereos. There are people out there who want the best and can pay for the best. That market for PCs is developing.
What's the secret to marketing to those folks and hanging onto them at the same time you are trying to go mainstream?
Nvidia has, if you look at our company and contrast it with Via or SiS, arguably ATI, we are the company that people who want to build high image, high impact, high brand come to. If you want to build the most leading-edge workstation or media center or notebook or game PC, or PlayStation 3, or a 3G cell phone from Motorola. They see us as the technology company that can help them create something very new, high image, high impact, and high brand. That is what Nvidia is. We do that through innovation and R&D. But we have to bring new ideas. The invention of the programmable shader is one of those ideas. The invention of SLI is one of those ideas. The invention of PureVideo HD is one of those ideas. You have to take risks and you have to be willing to make mistakes and build something pretty amazing.
How do you look at ATI?
They are a formidable competitor. They have wonderful people. It's the same way I've seen them before.
Given how the Xbox 360 turned out, did you have any regret about not winning the graphics chip for that console?
Not at all. We could not afford to build the graphics for the 360. Our most important asset is our people. If we use our people on a project where the economic return is not good enough, and there are other projects we could be working on, then we're going to lose money. We were a lot smaller company than ATI at the time. Maybe ATI could afford it and we couldn't. I know I couldn't afford it. I would love to build it. I just can't afford it.
ATI is excited about unified shaders. If you pull back, how do you see if your people are making the right decisions?
For each one of our generations, we need to have a vision of what we want to do. It costs hundreds of millions of dollars to come up with a new architecture. I don't know what they've built for next generation. It comes down to a different system vision for what we are trying to achieve.
Are you manufacturing the RSX for Sony now?
It is in production. It has been for some time.
Do you have a PlayStation 3 in your home?
Not yet. I hope I get one of the first ones.
Does it look like it's on schedule?
Sony hasn't changed their schedule. I think that's the most important thing. I thought it was a master stroke that they did.
Everyone criticized them for falling behind and having a high price and costs as high as $900. Why was it a master stroke?
PlayStation 2 was launched seven years ago in Japan at about $399. If you use inflation, it's the same price, approximately. The important thing is you cannot announce a game console for the next ten years and not have Blu-Ray. It's an impossible scenario. I think they got that perspective right. The moment we put those consoles together it's going to be very clear. If I'm going to buy a next-generation game console, I'm going to buy a console with next-generation media. It's going to last 10 years.
Two out of three of the players have bet you don't need it. Nintendo has bet you don't even need HD for the next five years.
Nintendo's perspective has always been different. The platforms that are being built now are not just game consoles. You use it for all kinds of other kinds of applications. In the case of Nintendo, they wanted to build a game console. They built a wonderful game device. Their focus is games and enjoyment. They will be myopically focused on that. I think that is terrific. Their perspective is different from the other two. If I'm going to buy a next-generation game console, I'm going to want next-generation media.
Do you think Microsoft planned for a five-year cycle and that Sony planned for a ten-year cycle?
I'm not sure how Microsoft is going to do in this transition. They are clever and they will figure out a way. I'll make a prediction that Xbox 360 can't possibly be a DVD-only device by Christmas of next year.
They will modify it?
I don't know how they will do it. But I just can't imagine going to a store and saying that this console has a Blu-Ray and this one has DVD. Remember Dreamcast?
Do you think they will go further than the HD-DVD accessory that they have planned?
I don't know how they will do that. Then they will have two platforms. You have the Xbox 360 with DVD. That is what Xbox 360 will mean. If you want to play HD, you have to buy an accesory? If that's the case, I'll buy an HD player. I know Microsoft will come up with clever ideas. These are the challenges. Every Xbox 360 they make in advance of that decision makes that decision that much harder. You could have two Xbox 360s. It's a very tough strategic challenge. Sega had the same strategic challenge when they launched ahead of Sony on PlayStation 2. The executives at Sega are very smart and the Dreamcast was a very good machine.
