Breakfast In The Valley
NeoMagic: Upwardly Mobile
By Peter Brown
From Page One of Electronic News: October 5, 1998 Issue
San Jose--After a meteoric rise to the top of the notebook PC graphics chip market, NeoMagic is looking for new mountains to climb. The first new targets for the Santa Clara firm: chips for digital versatile disc (DVD) players, and digital cameras.
According to Prakash Agarwal, president and CEO of NeoMagic, instead of competing against the likes of Sony, SGS-Thomson, LSI Logic, C-Cube and others in the home DVD market, the company plans to create DVD offerings for notebook PCs. Meanwhile, digital cameras are mobile by nature so they require low-power, highly integrated semiconductor offerings as well. This is where NeoMagic hopes its embedded DRAM graphics accelerators might find a place.
"If you look at the whole society, we are becoming more and more mobile," said Mr. Agarwal. "This means we need tools to be productive and efficient. At the same time, we saw multimedia coming into the picture, and what's driving the growth in the industry right now is the multimedia use in every (application), which primarily used to be all text.
"Now you see graphics, video, and audio, and that requires more data processing, more data storage, more data capture. That's our focus, to mobilize multimedia. It's very seldom that people do presentations and training at their desk. It's primarily when they travel, they go to the customer site or client site."
At a recent breakfast with the West Coast editorial staff at Electronic News, Mr. Agarwal discussed the company's potential move into consumer electronics, its rise to become the number one supplier of notebook graphics accelerators, and the future of the mobile and notebook markets.
Besides DVD and digital cameras, NeoMagic may be looking into communications products and other consumer ICs that require better mobile graphics, Mr. Agarwal said. "Anywhere we see that performance is required for the mobile product as well as it runs on a battery, we see (that as) an application for NeoMagic's technology or anything that has embedded DRAM," he added.
NeoMagic has made a name for itself in the notebook graphic IC market from a rising startup in 1993 to become the number one supplier in graphics in 4Q97. The company has continued to lead the market throughout 1998, gathering multiple design wins from heavyweight notebook PC makers such as Toshiba, IBM, Dell Computer, Hewlett-Packard, Compaq Computer and others. The company went from $200,000 in revenue in 1995 up to $40 million in 1996 and $124 million in 1997. NeoMagic is projecting it will hit $200 million this year in revenue. The company has approximately a 50 percent market share in graphics for notebooks, according to In-Stat, a market research firm in Scottsdale Ariz., with Intel-owned Chips & Technologies holding the number two position.
Mr. Agarwal said NeoMagic plans to introduce its first consumer electronics IC sometime next year, more than likely in the first half. However, in recent history some companies in the graphics market, such as S3, Chips & Technologies, and Trident Microsystems, have had difficulty moving into other markets, losing focus and neglecting their core products. Mr. Agarwal says NeoMagic will be different because the company is solely focusing on the mobile space and will be utilizing products that it can cross-market in the graphics and consumer space.
EN: How quickly do you think the market will develop for DVD and the digital camera in the portable space?
Mr. Agarwal: "DVD is coming along quite well. Last year they had a snafu because of all the copyright issues. Those are all behind us and quite a few notebooks and desktops are being shipped with DVDs now. So I would say in 1999 you will see them take off. They will start replacing CD-ROMs in all the PCs, including notebooks. And from that point on, the writable DVDs will start appearing. So (in the) year 2000, the year 2001, we will start seeing a lot of writable DVDs. And once that happens, it's going to open up many new applications.
"Digital cameras: In terms of growth, number of units, it's been phenomenal for the last few years. What's driving that growth is, again, the Internet. People want to capture the images and they want to include (them) in their presentations, and they want to share with their friends and family, rather than take the pictures, go to Kinko's, or get it scanned. It's not convenient. There are privacy issues and all of that.
"The same things will happen with the video. People will make their own videos. They will include (them) as part of their presentations and training. They will share with their family and friends without going through the hassles of going to a bureau. They will have the ability to compress it and convert it themselves. So all of those things are coming. Will all that happen in the next two years in a big way? Maybe not. But definitely over three to five years, then there will be big market."
EN: How are you able to succeed when the PC sales have been pretty sluggish over the last couple of quarters. There are a lot of companies that seem to be flat and faltering. You guys are obviously having a better year than last year.
Mr. Agarwal: "Basically, yes. There are three reasons why any company grows. One is, if you come out with a compelling product, you can take the market share away from the competitors, and not necessarily depend on how well the market is growing. So that definitely was the case. In our situation, we have been taking a lot of market share away. We basically went from zero to 45 percent market share in about two years. The market has not grown by that much.
"Second, obviously, whatever the growth in the market is, the notebook market has been growing at about a 15 to 20 percent rate, somewhere around 20 percent. ASPs have been coming down. But in our case, we increased the ASP of our solution, because of our integration. So that also helped us on the revenue side.
"I believe we have still got some more growth left for us in the notebook market and we are still taking market share away from our competitors."
EN: Do you have a particular target for market share?
Mr. Agarwal: "Anything is good. I mean 45 percent is a very good number. I always had done my business plan based on a 20 or 25 percent market share, and so what we have right now is more than what we had anticipated. But I have seen in the past, when I was at Cirrus Logic, we had at one time a 70 percent market share. Chips & Technologies, who was the leader before in the notebook market, had definitely seen 55 to 60 percent market share in the past. So I mean that's possible. (Whether that) will happen or not remains to be seen."
EN: How risky is it, to be in just the one market right now? I mean you obviously talked about moving into digital cameras and DVD, but that was a couple of years away, it sounds like. Now only being in the notebooks, it seems kind of like almost a risky proposition. What if that market bombs out for the company?
Mr. Agarwal: "The market is not going to 'bomb out.' The market will continue to be there. So opportunities are still there. If we had gone and started addressing too many opportunities, you basically end up doing nothing. And some of our competitors have proven that. So the strategy was that first we wanted to develop the embedded DRAM technology, apply that to a large existing market, where the standards are defined, and that was the notebook graphics opportunity, rather than going after Apple PDAs or desktop graphics. I mean we could have gone there with our technology, because it is very compelling in those markets. We probably wouldn't be sitting here, if we had done that, however.
"It has worked out really well, because (notebook graphics) has given us a foundation. It is giving us the R&D dollars to go and leverage the embedded DRAM technology into other areas. Now, if something happened, (if) one of our competitors or some newcomer came out tomorrow with much more compelling technology than NeoMagic, yes, we might be in trouble, but I don't see that in near-term horizon."
EN: Knowing the history of graphics companies, (I think) it seems like there's always one on top for about 18 to 24 months, and then somebody else comes up. How do you keep staying on top?
Mr. Agarwal: "Well, the key thing is innovative new products. And the reason we have been able to get the market share is because of our technology. In the past, what has happened, there was no innovation in the basic technology, the basic semiconductor technology. Everybody was going to the same fab, primarily, and used the same kind of semiconductor processes.
"In our case, what we have done (is to develop) a revolutionary technology, as well as put in a lot of features and performance on-chip. So for competitors to leapfrog us, not only (do) they have to come close to us on the technology side, they (also) have to offer more functions and features than we would be (offering) at that time. So we have raised the barrier much, much higher than it used to be."
EN: Any plans to enter the desktop market or maybe move into other areas?
Mr. Agarwal: "The notebook is the primary focus. We are not going to enter into desktop. That's not the area we want to get into. DVD drives is another market. Some of those drives will go into notebooks. Some of them may go into desktops. Some of them may go to home entertainment systems or whatever. That is one area we are focusing (on). Another is digital cameras. We are looking at some communication opportunities, too. But again, we will do fewer things but on a large scale rather than trying to do too many things."
EN: You mentioned that a lot of your growth was due to taking market share away from your competitors. And one of those is obviously Chips & Technologies, who is now a part of Intel. Are you keeping a watchful eye over there, to make sure that they aren't going to make a comeback?
Mr. Agarwal: "Everybody is, including the Justice Department. Intel is a very successful company. They have a lot of clout in the marketplace, so we definitely are paying attention to what they are doing or what they might be doing. Intel is also a great partner, too, because they're driving, along with Microsoft and a few other guys, interesting new platforms. So we work with them very closely and look for the opportunities where there is a win-win for both companies. So in that sense, you know, they have been more a partner so far than a competitor, but we'll continue to watch."
EN: Do you see room for half a dozen players in the mobile graphics market or just maybe one or two players?
Mr. Agarwal: "I think there is room for about two to three players who have 15 to 20 percent market share and one of the players with probably 40 to 50 percent market share. And this is how it has been for the last seven, eight years. In the past, what companies have not been able to do in graphics has been to diversify very successfully. Those who tried failed miserably. This is where we believe we are different. The reason we are different is that this embedded DRAM technology really gives us a leverage to go out and fulfill our vision. I mean let's take 3Dlabs or 3Dfx for example. What else could they do to diversify their technology? They can't. And if they tried to do that, they basically have to go and do something pretty much from scratch. This is why to companies, those who have tried in the past, it became too much of a burden."
EN: Will there be, or is there, a basic notebook computer, like a sub-$1,000 notebook? Or are they moving toward that?
Mr. Agarwal: "Yes, the notebook guys have no choice. I mean the price disparity between notebook and desktop (has) grown so much. So there is a trend. And probably toward the end of this year, early next year, there should be some machines coming out. Maybe not exactly sub-$1,000 but more like in the $1,400, $1,500 price range. And we had anticipated that and we have done some cost reductions and we'll be able to address that market."
EN: Hopefully that's not going to affect you like maybe it has some of the people in the mainstream market such as S3 and Cirrus.
Mr. Agarwal: "I mean, if that became the big percentage of our revenue, definitely, but I don't see that really being an issue. Actually, on the notebook side, it may be a new market. The reason is that typically where these thousand-dollar desktop PCs have been sold, were to homes and students. Now if there's a thousand-dollar notebook and they start selling in homes and schools, it's a new market. So if it takes off, I think it will be a new market for notebooks. So we'll have to see."
EN: You mentioned that you went from $41 million to $120 million in the span of a year. What kind of growth issues do you face? I mean do you have to go out and just do a mass employment? Or expand your facilities?
Mr. Agarwal: "Our growth, though our revenues have gone up, has been pretty well controlled in terms of head count. Right now we are about 230 people and yes we have been adding building space. The issue really is recruiting, because within the notebook and graphics side we still need to come out with next-generation products. We are also expanding into consumer areas. And it takes people. It takes good people (who are) very difficult to find, especially in this area. The real challenge is to hire employees and then recruit them, retain them, develop them, and grow them. That's one big focus in the company we have. Actually, turnover has not been that bad. It's been, I would say, probably less than 5 percent or so."
EN: Can you give us a status on some of the legal action that has taken place inside the company? In February of 1997, Cirrus Logic sent the company written notice asserting the company's MagicGraph 128 processor infringed on six U.S. patents.
Mr. Agarwal: "That was the second action they took, and then we went through looking at those, and responded to them that we do not infringe, and it has been over a year, and they have not come back to us. There may be some bad blood there and all of that. Any time they see something then they'll go after us, they may pursue that. But right now, there is no legal action pending between (the) two companies. It has been more than a year since they sent us the notice and we responded to them and they have not come back. Now they are very aggressively pursuing their IP strategy and if they decide to do something with their IP against us, we'll respond to it."
EN: There's a second legal action about someone holding a patent in Germany regarding graphics?
Mr. Agarwal: "Yes. There is this guy and he had gotten some patent that's at the system level and so it affects every graphics supplier. The patent is about to expire, actually, and he and his lawyers have filed to get it renewed. Actually this guy took Dell to the court in Italy. I think it got thrown out. So some guys, some OEMs, are making deals just paying him whatever money. Others are just ignoring it. Our position has been that we do not infringe."
EN: You mentioned you're a pioneer in embedding DRAM. How has the major shift in the DRAM market over the past few years affected you? It's become cheaper and cheaper all the time. Has that helped you?
Mr. Agarwal: "Any time that DRAM changes it creates uncertainty and doubt. So obviously, as a supplier to our customers, we have to address that issue as well as (the fact that) all of our chips are manufactured on the DRAM line, in a DRAM process; so we also have to work very closely with our suppliers. When prices go down, the prices of our chips go down, but at the same time our cost goes down because we are primarily manufacturing on the DRAM line. So you could say that that's (an) advantage, because if you are manufacturing on the logic side and logic prices are going down, you are not getting a benefit both on the logic and memory side."
EN: In your opinion, do you think that embedded DRAM has a place in mainstream PCs?
Mr. Agarwal: "I think it does. Some people actually have tried in the past and have not been successful. I believe that there's room there, there's a place there, especially with 3D because 3D is very memory bandwidth-hungry. The reason we're not going into it is not that there's not a play. It's not that compelling. Whereas in the mobile side, it's performance. At the same time, (it's) battery life, and board-space savings. It makes it more compelling. And if we stay focused there, then we'll continue to develop more and more mobile-related technologies, and this is where I believe that the growth of the future will be, more on the mobile side, rather than on the desktop side."
EN: Is the company looking at maybe integrating or embedding any kind of the new high-speed DRAMs that are coming out in the near future, like SLDRAM and Rambus or anything like that?
Mr. Agarwal: "We have looked at Rambus many times. Rambus is solving the same problem that we are, which is the memory bandwidth. But the problem with their technology is it only (optimizes) for performance at the expense that it burns more power and also doesn't help much on the space side. Actually, we are offering more memory bandwidth in our solutions than Rambus is offering with their solutions. Yet it is lower power and it is a single chip and so, in that sense, what we have been able to do is a lot more compelling in the mobile platforms than Rambus technology."
EN: However, Intel has endorsed Rambus, so they've got a lot of weight behind them at this point in the market.
Mr. Agarwal: "Sure. Now where Intel is going to be using them is on the system-memory side. And system memory, which is right now, a typical system is 32 megabytes (MB) to 64MB. It is not an application for embedded DRAM, because it cannot integrate that much memory. I don't think RAMBUS technology will be used in all platforms, especially notebook, because of the power problems."
EN: So you're of the opinion that you're probably just going to stick with your own technology, rather than go with any of them?
Mr. Agarwal: "Yes, because on the other memory like synchronous DRAM we are already getting 3.2 gigabytes (GB) of memory bandwidth with our 256-bit architecture, whereas RAMBUS typically had been about 500 to 600 MB, going up to about 1.6 GB. So I think we are just fine with our technology there."
EN: Could you tell us more about the benefits of having 256-bit technology that seems to be a big deal for NeoMagic?
Mr. Agarwal: "Yes. I mean as we get more multimedia on notebooks, which is 3-D, 2-D, video, and we also are integrating audio in our chips, it basically means you are generating all of the things at the same time trying to get the data from memory, it requires a lot of memory bandwidth. Some of our competitors try to confuse the market: 'Hey, look at our 2-D performance in 64-bit architecture. It is not bad. It is pretty close to NeoMagic's solution. You really don't need 128-bit or 256-bit.'
"The issue here is if you have a four-(lane) highway or a two-(lane) highway, if you're driving in non-rush-hour, it really does not matter whether you have two or four (lanes). But when you are driving in the rush hour, that is when it really makes a big difference. So if you are flying all the multimedia applications, all the multimedia data streams, that is where you really see the big performance boost, because of 128-bit and 256-bit architecture.
"The key thing about 256-bit architecture is that it is the footprint we are establishing right now and as we offer more and more 3-D in our future solutions, the bandwidth of the architecture is there to take advantage of it. We can upgrade people to their next-generation machines and performance. |
