Nvidia chief executive Jen-Hsun Huang is on a mission to get graphics chips into everything from handheld computers to smart phones. In a dinner with reporters on Monday night, the head of the Santa Clara, Calif.-based graphics chip maker said that the company is going to invest heavily in research and development and announce some interesting hires (see announcement story) soon. We last interviewed him in August.
Nvidia had a tough 2008 because of the recession and the resurgence of rival Advanced Micro Devices in graphics. Huang says he’s planning ahead for the new landscape that will emerge when the recession subsides. He expects, for instance, that low-cost Netbooks will become the norm and that gadgets will need to have battery life lasting for days. Holding up an Ion platform, which couples an Intel low-cost Atom processor with an Nvidia integrated graphics chip set, he said his company is looking to determine “what is the soul of the new PC.” With Ion, Huang said he is prepared for the future of the computer industry. But first, he has to deal with Intel. Here are some excerpts from the conversation.
Q: [What’s your take on the PC industry?]
A: A lot of people are asking when the economy will recover. The more pertinent question is what will the industry look like when it does. My guess is that the industry will not resemble the industry we had when the recession began. Some things we know for sure: The average selling price of the personal computer is nose-diving; the fastest growing segment of the PC industry is what people call Netbooks. It’s the not because the industry wants a new product category, but because it is the cheapest PC. I think therein lies some of the insights about what the PC industry will look like after the recovery.
And now we are starting to see the increased tension between the CPU (central processing unit, such as microprocessors made by Intel) and the GPU (graphics processing unit made by Nvidia). But we think it is a bigger story than that. It’s what is the soul of a new PC? We are investing in digital computing, GPU computing (using the graphics chip for non-graphics computing tasks), and mobile computing. The perfect example of all of those investments showed up in our platform we announced recently. It generated an enormous amount of buzz and an enormous amount of controversy. It’s our Ion platform. It’s a two-chip computer, with a CPU and a GPU. It’s so striking as a metaphor for the tension between the two chips and the shifting of the soul of the new machine.
Q: To play devil’s advocate, everyone thinks that Netbooks are the future. Why aren’t Netbooks just a passing fad?
A: We believe that the Netbook is growing because it is really cheap. It’s a cheap PC with Windows XP and an Intel microprocessor. You can give a Netbook to a teenager as a first PC. You can take it on the road with you, for accessing the web. The Netbook is not a new category. It’s just a cheaper PC. It’s like when Compaq launched the $1,000 PC a decade ago. People said it was dysfunctional. The first one was, but the next version was a lot better. The $999 price never went back up. It just got cheaper and cheaper. With Netbooks, it’s clear the $399 PC is here. It will just get better and better.
Q: Is there a place for a $250, full-keyboard device that uses ARM processors?
A: I think there is, but I think the price is less than $199. It has a full keyboard, it’s thin and it runs for a long time on batteries. It has a one-watt microprocessor. I believe this whole segment will have versions like the iPod Touch, the Sony PlayStation Portable and the Nintendo DS. I believe mobile internet devices will access the internet and come in a lot of different form factors, like phones, notepads and game players.
Q: How long can this go on? You just keep making them cheaper and how can you make money on it?
A: You’re going to get free PCs from your cable service. There is real innovation to be done here. I think the consumer would be really well served with a computer, a keyboard, a display and a battery life that last for two or three days. You use it to access the internet. What’s not to like about that? That segment of the market will happen.
Q: The iPhone at $199 is a fine experience. What’s the usefulness of these Netbooks? About 40 percent of them are getting returned.
A: That’s because people expect them to be like a PC. It doesn’t behave like a good PC. The components are getting better. A good LCD display is about $10. A notebook computer with Atom consumes about 20 watts. An iPhone is about half a watt. The Ion platform uses 15 watts when it’s in use, and maybe three watts when it’s not.
Q: Is Intel coming around to accepting that its Atom processor will be used with your Ion platform?
A: That’s the big controversy now. Everybody keeps saying that Intel won’t allow OEMs (computer makers) to use Ion. I’m still trying to figure out what that means. We’ve heard those kinds of things. The customers are trying to figure it out too.
Q: They could get in trouble with the new Obama Federal Trade Commission.
A: You can’t tell your customers what they are allowed to do.
Q: Are they offering an outrageously low bundling deal of Atom and Intel’s own chip sets (so that if you buy them together, you get a big discount) that locks out the Ion platform?
A: I’m not an antitrust lawyer. But if you offer a product A and bundle it with a product B and price it in a way that people can’t succeed without the bundle, then that’s a problem. The product bundling prices we have heard from the market seem very alarming. If you buy product A and B, the price is lower than if you bought product A alone. That seems like a weird bundle pricing strategy. (Intel denies this claim) We don’t know anything directly. It’s worth some investigative reporting.
Q: Your argument is that Ion is going to make Intel’s Atom much more useful?
A: Yes. We think Intel’s Atom is a fabulous processor. It’s small, low-power. Why not give it a companion processor from Nvidia that brings out its full glory. We hope Intel will see the wisdom of letting great innovative products reach the market. They have never done anything like what is being talked about now, which is bundling in such a way to make it impossible to sell a product.
Q: If you got Apple as a customer for Ion, would that prove that Intel is not standing in your way?
A: (Pause) I’m not sure that it is a good example. Apple gets what Apple wants.
Q: Your concern is more about the rest of the market?
A: Yes. The question is whether the rest of the market could build products like Apple. Apple innovates. It’s hard to push Apple around. They don’t need Intel’s advertising funds. They are pretty good at doing their own ads and their ads are funny.
Q: You said the industry would look a lot different after the recession. Who will be gone?
A: That’s harder to predict. I’m going to invest a lot more in R&D this year than I did last year — by a lot. I will be unique that way. I believe in the work we are doing with GPU computing, digital computing and Ion and mobile devices. The reason to have cash in the bank is to weather storms. This is an extraordinary storm. But we will invest in those things. For a lot of companies, the concern is that maybe all anybody wants after the recession is something like Ion. I am prepared for that outcome. Not every company will be happy about that. The CPU sells for less than $20 in this platform. But it’s a realistic world. If you know a $99 DVD player works fine, you will not buy a $499 DVD player.
Q: Are you saying Intel has tough times ahead?
A: It will be tough for all of us. Intel is a smart company and they will figure out a way to adapt. I see there is a 10-watt or 20-watt power device that you can use to do things you can’t do with 1 watt. These two kinds of devices are not likely to meet for a while. It’s maybe close to ten years, though they will meet one day closer to one watt. There is a decade’s worth of opportunity to move from a 60-watt world, where we are today, to a 10-watt world and then a 1-watt world. The 1-watt world is the smartphone, mobile internet device, the MP3 music player. They will all be able to access the web.
Q: There is only one proof point for that. It’s the iPhone. From one company. Why will these devices succeed across the board?
A: There is the Google Android. The Palm Pre. A few points makes a line.
Q: We saw this before. Palm was going to change the world. The only thing that changed was Palm.
A: I know what I can built with 1 watt. That’s the advantage I have. I built Tegra two years ago. The next Tegra is coming. Imagine what it’s going to do.
Q: I don’t have doubts about the chip industry’s capability. I have doubts about all of the others – the software makers, the box makers. Is your chip too far ahead of the curve?
A: What does Moore’s Law stand for? Whatever is possible will happen. If it’s possible for your competitors to do something, you assume they will. It is possible to build an extraordinary computer with 1 watt that is as wonderful as any computer you have ever seen. It’s in the next two or three years.
Q: Why don’t you do all of it? Create your own version of Atom?
A: We could, but my thinking is Atom exists. Why redo it? I should build the things that Intel doesn’t build. We could build great products for the world. Why would I build DRAM (dynamic random access memory chips) if Samsung already does? Nvidia has already avoided it. Look at Adobe. They don’t build operating systems like Apple or Microsoft. Adobe does just fine on top of the operating systems.
Q: But Intel is moving into graphics and AMD is doing a combination CPU-GPU chip.
A: That’s OK. That’s no different from your great shampoo with conditioner. The universal device is not a bad thing for some. But has Coca-Cola come up with another great brand besides Coke? Has Intel done great outside of microprocessors? We need to find a way to be extraordinary at what we do. It’s possible we don’t get integrated well. But most of the scenarios show there will be just two chips. I will have one of them.
Q: You don’t think it will be one chip?
A: I don’t think so. Intel has slipped its schedule on its (code-named) Havendale single-chip chip set. The problem is that CPUs and GPUs are both progressing. You can’t catch up on both with just one that tries to do both. And why not try to do it at the low end? AMD’s problem is that they are putting the high-end CPU with the high-end GPU. Who’s going to buy that?
Q: Microsoft, for the Xbox 720?
A: That’s cold. (laughs). That’s cold.
Q: Can you be more specific on R&D spending going up this year?
A: We are investing in GPUs. We started to invest in GPU computing, or using the GPU for parallel processing. We are investing a lot more in mobile than we ever used to. It’s moving fast. We invest in things like Ion and its successors, and Tegra. We are excited about our work with stereoscopic 3D glasses for the home. You get full resolution and you can see it in the home with televisions and displays. That brings the visual computing experience to the next level. Now we have to bring a new look to the world. During the recession, it’s challenging. But it won’t last forever.
Q: Are you picking up new talent?
A: Yes. It’s so exciting we’re going to announce it. It’s a great time to recruit.
Q: The Ion platform is targeted at Netbooks now. Who else is it targeted at?
A: I think the next-generation game console looks like this (holds up small Ion motherboard). Embedded computers look like this. Netbooks look like this. And the next-generation PC looks like this. The platform costs could be pretty inexpensive over time.
Q: There isn’t a big chip company being formed that dabbles in everything. Do you think the era of big chip companies that do lots of things is over?
A: That’s a great question. The days of the general-purpose merchant semiconductor company are gone. If the essence of the company is making chips of all kinds, that makes no sense. That’s gone. That’s like the consumer electronics industry. What does your company do? We make consumer electronics. It made sense for a while, but not anymore.
Q: What’s your view of AMD?
A: They couldn’t have survived without the Abu Dhabi deal. What else would Abi Dhabi have purchased? They don’t want the microprocessor business. They don’t understand it. They want a manufacturing business that requires an enormous amount of capital. That’s what they have. And very long-range goals. They have to be good at capital investment, process technology, and have an appetite to invest long term. They are smart investors. It’s a brilliant deal that saved the company. It would have vaporized by now.
Q: Are there any Ion-based notebooks announced yet?
A: I cannot announce our customers products. We just announced it a month ago. Design cycles are typically less than a year.
Q: If you were writing the ad copy for the first Ion notebook, what would you say about it?
A: Amazing PC at $399. It is amazing, and its value proposition is that it is inexpensive at a time when the world has no money.
Q: Will Windows 7 make it into Netbooks?
A: Yes. Windows 7 will go all the way down into the cheap PCs. Vista didn’t.
Q: What is the next-generation user interface?
A: I may sound old fashioned. I want a MacBook Air with battery life that lasts forever for $199. I like that big keyboard.
Q: AMD gained share in 2008. How will you sock it to them in 2009?
A: We are miles ahead of them in GPU computing. That’s using the GPU for non-graphics processing. It has passed the tipping point and it is now the plan of record for all of the major computer companies.
Q: Jeffrey Katzenberg is giving effusive testimonials for Intel’s Larrabee graphics chip.
A: You and I don’t know what Larrabee can do. It doesn’t exist. If you pay for Katzenberg’s compute farm, he will tell you your lawnmower is great for making movies. There are no applications that run on Larrabee today. You have to recompile them.