Icwindows Electronics HK CO.,Limited

Vyacheslav Sobolev, DigiTimes.com, Taipei [Monday 5 February 2007]
  Riding on the wings of the golden era of Taiwan's motherboard industry, Elitegroup Computer Systems (ECS) saw its share price soar to over NT$200 by the beginning of 2002, while the company's revenues continued climbing until peaking in November 2002. However, the period between 2003-2005 was a really hard time for ECS, as the company suffered on-year revenue losses for 22 consecutive months. The decline was painful, but now, allied with Tatung, the company is trying to rise back from its ashes like a Phoenix. In addition, with the industry seeing consolidation, ECS is not facing the challenges it saw previously.

  With ECS recovering, does that represent a rebound for the motherboard industry on the whole? And what can we expect in the near future for the company and the industry? DigiTimes.com recently had a chance to sit down with David Liu, ECS vice president and deputy general manager of the mainboard and add-on card business unit, to discuss these and other questions.

  This is part one of a two-part interview. Part II will follow on February 6.

Q: What do you think about the current situation in the motherboard industry as a whole? Is it alive and healthy? Is it growing? Is this business profitable now?

A: The motherboard business is still quite healthy, basically, although many changes happened to the industry over the past several years. I mean not only technologies, such as CPUs or interface buses, but also the way to make boards, in other words - design routines. There are also many changes expected to come in the near future. New CPUs, new sockets, new chipsets, new memory types, new buses. For example, AMD will soon have a new socket called AM2+, and then, it will introduce the AM3 socket for quad-core processors supporting DDR3 memory. On the other hand, Intel said it will a have a serial-bus architecture in 2008. Any change of this sort means an opportunity for motherboard makers to gain more market share, so I believe the future for motherboard makers is bright until, at least, 2010. Generally speaking, Intel and AMD, introducing new technologies and architecture solutions, will continue assigning many new tasks to the motherboard industry, unless they both have a system on chip.

By the way, over the past several years, we heard many times that notebooks will replace desktop PCs, so the market space for desktops will become smaller. However, the desktop market is still growing, and motherboard makers are taking a slice of that pie.

Of course, the business itself has been getting tougher, with profit margins currently slim and expected to be even slimmer. For big guys, it will not be a problem to survive in this situation, but smaller companies may not be able to overcome that challenge. As you see, the number of small players in the motherboard industry has been decreasing. Some of them turn to niche markets, such as industrial PCs, to save their business, while others just move to a different business. Summarizing, I would like to say that the motherboard market will continue growing, but it will mainly benefit top-tier players.

Q: Will quad-core CPUs or the Windows Vista operating system significantly impact the motherboard industry in the near future? Do you see any other factor that may significantly impact the industry?

A: My opinion on quad-core CPUs is that Intel now wants to set up game rules in the market of quad-core CPUs, leveraging its strength in semiconductor manufacturing, to beat the main competitor, AMD. I do not think that the PC market will be mainly driven by quad-core CPUs in 2007, although Intel will likely push them ahead with its plans and roadmaps. As I currently see it, dual-core CPUs will remain the mainstream in the market during the whole year, mainly because prices of quad-core CPUs will remain too high for a greater part of the year.

Regarding Microsoft's new operating system (OS), I expect Vista's impact to be much more noticeable than anything on the hardware side. For the those who want to effectively use the new OS, it will be necessary to upgrade every part of their system, including video (GPU), audio, memory and CPU. It will create demand in the market, which will mean new business opportunities for hardware manufacturers, including motherboard makers.

By the way, with Vista released to end consumers, I believe we will see a spur in the desktop market.

Q: Why?

A: Because Vista requires performance. Notebooks significantly grew in popularity since Intel introduced Centrino, as opposed to desktops using Pentium 4, with its huge power consumption at comparable performance. Now, being based on the same microarchitecture as notebook-use Merom chips, Conroe solves the problem of power consumption, while keeping the ability to provide better performance. I think this will attract users back to desktop PCs. After all, notebooks always have to balance performance and mobility, and although they share the same microarchitecture, desktops are more free to focus on pure performance. Due to that, I would not be surprised if users will choose Core 2 Duo desktops with Vista, while staying with Windows XP on Core 2 Duo notebooks.

I would also like to add here that with LCD monitors becoming mainstream at 17 inches and above, the gap between desktops and notebooks will likely grow even faster. Several years ago, when users started adoption of LCD monitors from 14-inch and 15-inch models, these screen sizes were also available for notebooks. Now, with LCD monitors coming in 17-inch and larger sizes, the user experience, especially the entertainment experience, again becomes something completely different compared to using a notebook.

Q: There is an opinion that less than 10 companies will eventually survive in the motherboard business. Do you agree?

A: I think the number will eventually be five or six, maybe even less. I mean real big players, although the number of motherboard brands can be larger. Maybe up to 8-10 because some market segments can be better served by the same maker with a different brand, so this may help smaller makers survive as brand holders focused on engineering and design work, with manufacturing transferred over to a big player. This could be the way they can still make money, retaining their brands and lowering their operation expenses.

Q: Have you considered maybe that the consolidation process in the motherboard industry is a consequence of motherboards becoming a commodity, making it difficult for a company to differentiate itself from the competition?

A: I agree that in the early years of the PC industry, motherboard makers could put more innovation into their products, and we also had more space to develop new features in cooperation with vendors of semiconductor components. Now, integrated chipsets leave almost no space for that, since they include most of the key functions, which were previously implemented in the form of separate chips on a motherboard. Of course, this makes differentiation harder, and this is actually another reason why smaller players face the problem of survival. If competition is all about cost, big players have a clear advantage over smaller ones.

Q: How do you think recent mergers, such as Nvidia with ULi and AMD with ATI, may affect the industry? Do you already feel any impact? Do you expect more mergers of this sort in the near future?

A: The clear thing is that the two mergers further deepen the integration within the supply chain of key upstream components for the motherboard industry, although even before, the supply chain was deeply consolidated compared to the industry itself. For example, we had and still have only two major CPU suppliers, with VIA playing a role of CPU vendor targeting niche markets rather than the mainstream PC market. With chipsets, graphics and audio solutions, we have similar situations.

There are two different stories behind the two mergers. Nvidia and ULi approached the chipset development in similar ways. For Nvidia, merging was an optimal choice because ULi had a team of experienced designers, while the cost to absorb this team was lower compared to that of a design team with similar experience in the US. The other important thing is that most motherboard makers are Taiwan-based companies, and ULi was in good contact with this community. Now, after the merger, Nvidia can work more closely with Taiwan motherboard makers.

AMD's motivation to acquire ATI was absolutely different. While leveraging CPU performance, AMD gained some success in its fight with Intel.

Intel remains ahead in deploying new manufacturing processes and AMD would need more capital to compete with Intel in production facilities - or to put it simply, more cash to build foundries.

Lagging behind in the adoption of lower process nodes, AMD puts itself in a difficult situation, so they truly need architecture innovations to outpace Intel, offering something that Intel does not have. And if you look at what is currently the most discussed topic for the industry and end users, that would be graphics rather than CPUs, especially in talks about gaming PCs or game consoles, such as the PS3, Wii or Xbox 360. There is no big interest in the CPU used in the PS3, frankly speaking, but its graphics subsystem attracts a lot of attention.

AMD acquired ATI to bring the dual-core concept to a new level: Still two cores, but with different functions. One for general computations, and the other one for graphics. I think that the merge was a signal to Intel, which should maybe accelerate its own graphics development and architecture modification. Otherwise, they may lose a battle in 2008, if a new AMD has a product based on the new concept in the market by that time.

Q: If Intel loses the game, what about Nvidia by that time?

A: Nvidia would be in the same situation, so this is why they try to promote their brand as the only choice for those who want to have high-performance PC graphics. No matter if your system is equipped with a CPU from Intel or AMD, it must have a Nvidia graphics chip inside.

Naturally, after AMD and ATI merged, Nvidia started reallocating internal resources to pay more attention to Intel-related developments. However, in my opinion, this will hardly lead Nvidia to merge with Intel in the near future. I do not expect Nvidia, which is 100% sure its graphics technology is the best in the market, to be open for this sort of negotiation. I even think it was Nvidia that was first approached by AMD before ATI became their target.

Q: What is the current breakdown between AMD and Intel platforms in motherboard shipments by ECS?

A: Motherboards for the AMD platform accounted for 30% of our unit shipments in 2006, while products for Intel-based systems contributed the remaining 70%.

Q: In the first half of 2003, ECS was very close to Asustek in revenues, even sometimes outperforming the competitor in some months. However, starting from the second half of that year, Asustek's revenues soared, while ECS had a very long series of on-year drops in monthly revenues from May 2003 to February 2005. Since March 2005, the company has been only posting on-year monthly revenue growths. Why did ECS' revenues drop that much?

A: ECS actually saw a large amount of success in 2003, dominating the entry-level market and shipping over one million motherboards per month. Why did we drop in revenues from the second half of that year? I think one of important reasons, maybe the most important, was that we lost our R&D vice president Sterling Wu who left ECS the year before and joined Asustek to become ASRock president. Asustek already had a very strong brand name at that time, but their position in the entry-level market was not that good. At ECS, we knew how to make entry-level motherboards, and we also knew where to find customers for them. With one of our top executives, Asustek had the knowledge to enter the market, which was previously inaccessible for them, and they used it to be successful in this market.

Also, I should say that while ECS always relied on its own production, Asustek was more flexible in its manufacturing policy. Actively investing in production facilities in Suzhou, they were not afraid of outsourcing. They leveraged the brand, which was better recognized than the ECS brand, and expanded their business to new areas, such as notebooks and communication hardware. They did many right things that helped to maintain the stable growth. Asustek is actually a very strong competitor, I have to say.

Q: What is your outlook for the company in 2007? How many motherboards does ECS plan to ship this year?

A: I think ECS currently holds the number three spot in the motherboard industry, with the first and the second places respectively occupied by Asustek and Foxconn (the registered trade name of Hon Hai Precision Industry). In terms of branded shipments, I think we are maybe higher than Foxconn because their main output goes to OEM/ODM customers, while the Foxconn brand is still not that noticeable and widely recognized.

In 2007, we will try to maintain the 60:40 ratio between OEM/ODM and branded sales. Our motherboards shipments exceeded 24 million units last year, and the company's target for this year is to ship 30 million motherboards. With our existing production facilities, it will not be a problem to manufacture this quantity, while the actual challenge is how to get them shipped. The motherboard industry is only expected to grow by 6-7% in 2007 in terms of volume shipments, so our target is to outpace the industry by 2X or 3X, grabbing market share from competitors. If we are able to manage this, I believe ECS will either be near Foxconn or even surpass it, becoming the second largest supplier of motherboards.