Another Interesting Article:
Interviews
Semiconductors ready to get in the chips
Analyst Danny Lam says the cyclical chip industry is on the verge of a boom with potential growth of more than 20% this year. Telecom suppliers like STMicroelectronics will benefit, but not Intel.
By Eneida Guzman
It's a little known fact that before Wall Street's big-name stock analysts make recommendations, they often check in with their own behind-the-scenes industry gurus. For semiconductor industry analysts, that often means talking to Danny Lam.
Lam is a director at the Delaware-based consulting firm Fisher-Holstein and a semiconductor industry specialist. He claims some of Wall Street's most respected analysts, as well as Silicon Valley's most profitable chip makers, as clients. He relies on sophisticated, proprietary analytical tools for everything from measuring a chip company's production to estimating its future profitability -- often sifting through mounds of minutiae too time-consuming for his fast-track clients.
In an Investor interview, the "analysts' analyst" predicted that chips are headed for a two- to three-year boom, but one that may not benefit the sector's traditional darlings.
The chip industry hasn't enjoyed the prolonged bull market cycle enjoyed by the rest of the technology sector. Yet its products are at the center of almost every major innovation. How do you view this business?
We are traditionally a boom-and-bust industry, and I think we are right now hanging on the verge of a major boom. The industry traditionally evolves and changes its stripes every cycle. During the last decade, the PC drove the boom to a large extent. I think we're going to see the drivers change. PCs will always be there; they will still be important. But the next major driver is probably going to come out of telecommunications. So it's going to be a very different driver. A different group of players is going to benefit. The metrics of measurement in the last round focused on companies like Intel. That's going to change.
Chips abysmal in 1998
After an abysmal performance last year, the semiconductor industry is expected to grow by 11% in 1999. What groups within the sector do you expect to lead the recovery?
The DSPs, or digital signal processors, are going gangbusters. Microprocessors have already recovered. Memory is recovering slowly, but it is gradually turning around. I'd like to address the 11% number. That's the most pessimistic number. I visualize a range of scenarios for growth in the sector, including one scenario where there's a possibility of growth in excess of 20%.
Tell us more about that more optimistic scenario. And where within the range of scenarios do you think we're most likely to fall?
We measure the market in U.S. dollars, which leads to wild swings in growth numbers as exchange rates of major semiconductor producer and consumer countries shift. The optimistic foreign-exchange scenario will see the Japanese yen remain at close to 110 to the U.S. dollar. At that exchange rate, the market only has to grow marginally to be nearly assured of 11% growth in U.S. dollar terms. Under such circumstances, a stable or moderately growing Japanese semiconductor market, combined with no major outbreaks of "Asian" or "Latin American" or "European" or "Russian" flu, would result in growth numbers in the teens for 1999.
The metrics of measurement in the last round focused on companies like Intel. That's going to change.
It seems that all of the demand drivers that are in place this year for the chip industry -- namely telecom devices and the Internet -- were in place last year as well. Why the turnaround in stock performance?
Last year was unusual in that we had a double-dip. The market peaked in 1995. It slumped in 1996, which everybody sort of expected. It came back a bit in 1997, and then the Asian crisis literally just gutted 1998. So 1998 was a terrible year because a lot of people had burned up their reserves preparing for what appeared to be recovery in 1997, only to have their legs cut off from under them. This year it is very clear that the recovery is for real. We have now seen at least two solid quarters with the microprocessors. We have seen key segments like DSPs come on strong, and even memory prices have risen a little bit from their very depressed levels as of July last year.
Are the oversupply problem and the Asian economic crisis over for semiconductors?
Again, you have to break down the story. It is over in the sense that Intel (INTC) does not have a problem with oversupply. Similarly, Texas Instruments (TXN) does not have a problem with oversupply in DSPs. However there is a lot of supply in the memory market, particularly in the older generation devices.
Obsolescence should gradually take care of that.
Correct. And sometime later this year we're going to transition to yet another new memory technology called Direct Rambus.
Details
--------------------------------------------------------------------------------
Company Report
Earnings Estimates
Research Wizard
Tell us about Rambus.
Rambus (RMBS) is a company that has developed a proprietary architecture for memory called Direct Rambus. It provides much faster access to memory, solving one of the key bottlenecks to faster overall systems performance. However, the company does not manufacture memory chips; it licenses the technology to chip companies who pay them a royalty. Rambus is more expensive to manufacture than existing memories, and chip makers also have to pay a royalty. Memory chip makers, after losing billions in the last few years, are not in the mood to invest in a new product which costs them more to make, requires substantial royalties, and may end up only having a small market because it is much more expensive than traditional SDRAM.
Who's leading the way for the new memory standard?
There are not many chip makers that are committed to it right now. Intel is scrambling to try to bring people on board. It has committed to that standard in the sense that it has designed microprocessors on motherboards around that memory. So, in effect, they have to have it, unless they're going to seriously change their product road map. I'd say Micron Technology (MU) has the strongest commitment to Rambus. And Intel is close behind.
Details
--------------------------------------------------------------------------------
Company Report
Earnings Estimates
Research Wizard
Intel's Strategy
Please expand on the reasons Intel and Micron are so committed. What do they have to gain from the success of this Rambus technology?
Intel is strongly committed because they view Rambus as an integral part of their strategy to control the architecture of the PC and protect the dominance of the Intel microprocessor on the desktop. By backing Rambus, they have created yet another expensive hurdle for AMD and Cyrix to jump over in cloning the x86 architecture. AMD would have to clone both the x86 processor and a compatible chip set to support Direct Rambus memory. While this does not stop them, it slows them down in a world where time is money. If Rambus fails to gain broad acceptance, Intel will have to resort to other defensive measures.
You mentioned that demand drivers would shift from PCs to the telecommunications industry. PCs now make up about half of the end-use demand for semiconductors. What will that number look like in the near future?
The computer segment, which includes PCs, was about 54% of the overall market in 1998, according to IC Insights. PCs account for roughly 40% of that 54%. It would have been higher if memory prices had not been in the dumps. Telecom is 19% of demand now, rising to 22% of the market by 2003. However, the key is to recognize that, while the computer market appears to be big, vast chunks of that market consist of commodity parts like memories that make slim profit margins.
Demand for wireless products is expected to fuel this boom for quite some time.
Yes, and wireless devices are not limited to just cellular phones. We have digital cellular products such as personal communications services, like CDMA and GSM digital cell phones, digital pagers, and probably in a year or two, wireless Internet devices are going to take off. We are now at a point where wireless prices have fallen so much that they are beginning to nibble at the wired infrastructure. Prices are beginning to be comparable between carrying a wired phone vs. a wireless phone. That's going to be a big shock to the traditional telecom players.
You're not a telecom analyst, but who benefits under that scenario?
I track telecommunications businesses because they are big demand drivers for semiconductors. The telecommunications company to be reckoned with is Nokia (NOK.A) Wireless. According to the Financial Times, Nokia sold 37.4 million handsets in 1998, representing a 22.9% market share. Motorola (MOT) is now No. 2. It sold 32.3 million units and has a 19.8% market share. Ericsson (ERICY) has now fallen to third place with 23.8 million units sold.
Details
--------------------------------------------------------------------------------
Company Report
Earnings Estimates
Research Wizard
You study companies that are strongly positioned to ride future booms in semiconductors. What makes sense in here?
We're looking at suppliers to the telecommunications companies, like STMicroelectronics (STM). They are a European company and one of the largest suppliers to Nokia.
Nokia by another name?
So STM is really a play on Nokia?
It's a play on Nokia, but STM is a very diversified group. They also have a large business in smart cards.
Fundamentally, what do you like about the company?
Growth estimates for the company for 1999 are about 8.5%, increasing to approximately 20% for 2000. And you're paying 32 times trailing earnings. It's not expensive.
Details
--------------------------------------------------------------------------------
Company Report
Earnings Estimates
Research Wizard
You still like Taiwan Semiconductor Manufacturing?
Yes, Taiwan Semiconductor Manufacturing (TSM) is a company that is going to outperform not so much because it's linked to telecoms, but because of the nature of the foundry business. The foundry business goes through considerably steeper up-and-down swings than the rest of the semiconductor industry. The reason is because during slumps the higher-end companies like the Intels tend to enter the foundry business. They'll make other companies' chips just to use up capacity. Similarly, if a memory maker has no business it will become an instant contract manufacturer. But what happens when demand tightens up is that these companies just as quickly exit the business. So this is a company that can go from 60% capacity utilization to 90% capacity utilization in a matter of a couple of months. In my judgment, that's going to happen sometime within this year, most likely in the second half of the year when demand tightens up.
You're a big supporter of Motorola. Future earnings growth looks bleak for them. What's your thinking there?
Motorola is a bit of a long shot and I know opinions on this company are very mixed right now. But I look at it this way; Motorola is a company that has stumbled very badly. They're generally regarded as having missed the market in digital and wireless, and now they are scrambling to catch up. But, if you think about who the company really is, you are getting a semiconductor company -- that's roughly 28% of its business -- bundled in with a wireless cellular phone company plus pagers and all these other little businesses on the side. At Motorola's present valuations, it is like getting the semiconductor business for free if we assume that Motorola will regain its footing in the digital wireless telecommunications business.
You're referring to Iridium World Communications and its satellite wireless systems?
Precisely. Iridium (IRIDF) was the first company to invest in it. Right now, in terms of hand-held sets, the numbers are running a little behind schedule. But the fact remains that they are the only company that has a working system that is actually deployed.
Details
--------------------------------------------------------------------------------
Company Report
Earnings Estimates
Research Wizard
How to judge Motorola
So what specifically should investors watch for with Motorola?
Watch very carefully their market-share numbers in the digital cellular infrastructure and handset business. Make sure that the semiconductor products sector does not make the mistake of not having sufficient capacity to meet demand during the next semiconductor boom. Specifically, watch for them to have solid foundry announcements within the next 60 days to look after their capacity needs later in the year.
You're expecting big things to happen in the chip market this year, yet you believe that the world's largest chip player, Intel, is fully valued. Why?
What we know today is that PCs in the United States are in roughly 50% of homes. There are very few places I could think of to put more PCs in North America. There's good market potential in Europe. There is fantastic market potential in China and India, when those markets begin to mature. But in aggregate, I see unit PC growth as not having any prospect of going to the 30% level that would justify Intel as an above-average performer. So, in a nutshell, I don't think Intel offers great growth potential.
I could not have envisioned PCs hitting 50% of homes if it wasn't for the Internet. I could see 25%. But not 50%.
Europe is the second-largest market in the world for semiconductors. When will the promise of widespread PC penetration materialize there?
We looked at this issue very carefully last year. It was very clear that the issue of PC adoption rates in Europe being much slower than North America is a structural problem. It's not an easy problem to solve. It involves dealing with things ranging from reasonably priced Internet access to social issues. Most countries in Europe do not have flat-rate local calls. So you start drilling down, and you realize suddenly that these are very complicated structural issues that lead to great barriers of entry in rapid technology adoption. Why are the phone companies like that? You find, oh, it's a national monopoly. Those are very hard to change.
How do you view the effect that the Internet industry has had on the semiconductor business?
It has given the industry a breath of fresh air. It has given it life that I could not have foreseen five years ago.
Specifically?
I could not have envisioned PCs hitting 50% of homes if it wasn't for the Internet. I could see 25%. But not 50%.
Can you share with us anything about the prospects for the future of the chip business that most of us are overlooking?
Expect at least two to three good years. But then don't be surprised if we have a bust. It has happened before, and it will probably happen again. The later you get into the boom, the more likely you are to be in a bust. Don't hesitate to sell at or near the peak.
|
