SI
SI
discoversearch

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor.  We ask that you disable ad blocking while on Silicon Investor in the best interests of our community.  If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.
Technology Stocks : Micron Only Forum -- Ignore unavailable to you. Want to Upgrade?


To: Trey McAtee who wrote (39546)10/1/1998 11:30:00 PM
From: DJBEINO  Respond to of 53903
 
Is that light at the end of the tunnel?
By Will Wade and J. Robert Lineback

SAN JOSE -- The semiconductor industry is beginning to think it sees the start of a turnaround, but it is still too shell-shocked to believe it. Market observers say it's still too early to declare a bona fide recovery is in the making, but they do see a clear rebound from the depth of the slump last July.

"The tea leaves seem to be indicating there was some nascent strength in the market during the third quarter," observed Silicon Valley analyst Ed Henderson of Henderson Ventures in Los Altos, Calif. "We have seen some pickup in the printed-circuit board demand, for example. Semiconductor orders have picked up two months in a row. Some of this could be seasonal, but we seem to be heading toward a slow recovery," he said.

The next piece of evidence that people are looking for now is a move to higher average selling prices for semiconductors. This is particularly true for DRAM prices, which have declined dramatically for the past three years because production capacity outstripped growing worldwide demand.

Now, major DRAM makers -- led by NEC Corp. and Micron Technology Inc. -- have started to test the recovery waters by increasing the prices of 64-Mbit memory prices by about 10% after several vendors began scaling back production this past summer.

"That's a marketing experiment," pointed out Michael L. Hackworth, chairman and CEO of Cirrus Logic Inc. in Fremont, Calif. "If the pricing [increase] holds, it means that capacity utilization has bottomed out," said Hackworth. "The second half is going to be better than we had feared it would be."

But like many Silicon Valley executives, Hackworth is now waiting to see if the improvements are just seasonal or whether the rebound has staying power. He currently believes the improvements will lead to a mid-1999 recovery.

The third quarter may have ended the downturn, but "the question is when will we see the beginning of the upturn?" asked Jean-Philippe Dauvin, vice president and chief economist at ST Microelectronics in Paris. But he is optimistic and believes that all of the conditions appear to be falling in place for a modest upturn in the fourth quarter.

"If so, it will be the first [sequential] increase we've seen in six quarters," said Dauvin, who was just named president of the World Semiconductor Trade Statistics (WSTS) organization that keeps the official market numbers for most industry trade associations.

The battered chip makers and their capital equipment suppliers could certainly use some good news after 1998 turned out to be such a big disappointment. First, the Asian financial crisis turned out to have a much bigger impact on the chip industry than expected at the start of the year. Chip vendors also saw some of their markets change in 1998 with such shifts as a boom in low-cost personal computers, which drove down chip prices.

As a final shock, an acceleration of device shrinks -- aimed at cutting costs -- ended up dumping more products onto an already flooded market, according to industry observers.

One troubling trend in the 1998 slump is weak unit shipment growth, said Jim Feldhan, president of Semico Research Corp. in Phoenix. "We are expecting only 1.3% unit growth this year, which is very weak considering the fact that prices are down," the analyst noted. "Usually, when prices drop, unit shipments go up."

Slow end-equipment market growth and chip integration have reduced the chip unit growth this year, he said. IC volumes will increase by about 8% in 1998 but could be in the 20% growth range in 2000.

But coming up with industry forecasts for next year is proving to be a trickier job than usual. ST Microelectronics' Dauvin said he was still finalizing his own 1999 forecast at the end of September because he wanted more economic data from several key countries, in particular the United States and Germany, which is in the process of installing a new government under chancellor-elect Gerhard Schroeder of the Social Democrat party.

The European chip executive is maintaining a cautious stance at this point. "Right now, it looks like a 5% to 8% global growth year compared to -3% to -5% in 1997," said Dauvin, who believes a full-blown chip recovery will begin late in the second quarter next year after a somewhat flat first quarter.

The annual wave of fall forecasts did start, however, and Dataquest, at the end of September, was the first to issue its prediction. But the uncertain San Jose market researcher wasn't taking any chances, and issued two different forecasts: one based on the assumption of a 29% recovery in DRAM revenues, and the other leaving out the erratic memory chip segment altogether.

If such a solid rebound in DRAMs does occur -- that would be rising from $14.3 billion this year to $18.5 billion in 1999 -- Dataquest expected that worldwide chip revenues will grow about 12%, from $139 billion in 1998 to $155 billion next year. For this year, the market researcher predicted that total global semiconductor revenues would fall by 6%. If DRAMs are taken out of the forecast, semiconductor revenues will reach $24 billion in 1999 -- about a 10% increase, Dataquest predicted.

Another problem in making 1999 forecasts now is trying to figure out how much of an impact consolidation in DRAMs will have on capacity and memory pricing next year. "This is the big question in 1999," said Dauvin, referring to the merger and acquisition activities that could leave the industry with four DRAM suppliers controlling two-thirds of the market (see story in the Sept. 15 publication). "I continue to think the overcapacity in DRAMs stands at 25%," he said. "I am convinced that the DRAM industry is going through a consolidation." But he believed the impact of that consolidation will not likely be felt until the third quarter of 1999.

When DRAM markets do recover, the impact will be felt on other IC segments, according to analyst Bill McClean, president of IC Insights Inc. in Scottsdale, Ariz.

"An indirect relationship between the health of DRAMs and other IC segments has emerged," he maintained. "If the DRAM business picks up, it takes away some of the incentive for memory makers to target system-on-chip products and logic. Samsung, for example, has been pushing harder to offer foundry services," McClean noted.

Dauvin also believes healthy DRAM suppliers will also stay out of other memory markets, such as flash. "The DRAM consolidation will certainly help many semiconductor segments in the coming years," he added.

But for now, many chip makers are still adjusting themselves for slow growth and reduced revenue streams after a rocky nine months in 1998. The latest round of layoff announcements in late September is coming at a time when some executives are beginning to become more optimistic. Cirrus Logic, for example, announced in late September that it was laying off 500 workers and restructuring its fab capacity in a move to return to a full-fabless model.

"Confidence is improving," said Cirrus chairman and CEO Hackworth, who remains cautious. "We see the light at the end of the tunnel. We worried at first that it was a train, and now we are pretty sure it is not."

But there won't be any quick action by chip makers to rehire laid-off personnel or restart postponed or canceled fab projects, even if conditions continue to improve, industry analysts believed. "Everyone has decided to adjust themselves for the worst," said analyst McClean of IC Insights. "If things pick up a little in the fourth quarter, they probably can make adjustment with the people they have, but if the recovery continues, we'll see hiring again. If the markets do take off, we'd have a classic case of mistiming because companies would have to rehire all the people they were walking to the door."

semibiznews.com



To: Trey McAtee who wrote (39546)10/1/1998 11:53:00 PM
From: DJBEINO  Respond to of 53903
 
Trying to make Korea's 'shotgun marriage' work
By Jack Robertson

SEOUL -- Nothing is tougher to execute well than a corporate merger. Even the best-planned corporate marriages between companies that are seemingly compatible with one another often blow up. So where does that leave the "shotgun marriage" now being forced on Hyundai Electronics Industries Co. and LG Semicon Co. by the South Korean government?

The outlook certainly doesn't look good for the new combine, which is expected to generate enough DRAM sales to rank it near the top of the global heap. The strong-willed owners of the two companies have been squabbling for weeks over which one will control the new semiconductor giant. Even if they can agree on who will run the new company, industry observers predict that melding the disparate managements is going to be a rocky affair that's going to take a lot of time.

Sorting out the pecking order isn't going to be easy, agreed Ilsuk Han, analyst for ING Barings Securities in Seoul. Concluded International Data Corp.: "Hyundai and LG will have large vested interests in the new company that don't [see] eye to eye. Each [faction] will be fighting for [its] stake and it will take some time to work it all out."

But that's just the beginning, because the merged company then will have to merge two wildly divergent approaches to chip design and production processes, and marketing, R&D, and capital investment strategies.

LG Semicon scored an early jump on the PC-100 64-megabit fast SDRAM market by improving its existing i-line lithography processes to handle quarter-micron feature sizes. This proved to be a faster way to get to market than was Hyundai's strategy, which was to jump headlong into the next-generation deep-UV steppers.

LG's conservative approach was defended by Michael Sporer, technical marketing manager for LG Semicon America in San Jose. "LG manufacturing technology does not require the most advanced lithography processes," he said. "We can build 64-Mbit DRAMs on 0.30 micron processes in the smallest die size [using 0.30-design rules]," Sporer continued. "Indeed, LG has been imminently successful in the market with this approach and is now shipping the largest number of PC-100 64-Mbit SDRAMs in the world, with a 35% market share."

Hyundai, in contrast, is committed to deep-UV and DRAM die shrinks to get more and more chips from a wafer, which is the cost-cutting strategy of most other major competitors. Survival in the DRAM industry will depend on successful die shrinks and extending the 248-nanometer wavelength processing down to 0.18-micron feature sizes, according to Kye-Hwan Oh, Hyundai's executive vice president for semiconductors (see story in the June 15 publication).

LG currently sells more heavily into the spot market than does Hyundai which concentrates primarily on OEM contracts, analysts said. Each strategy has its own strengths and weaknesses, but the merged chip operation will have to be able to patch up any conflicts between the two in supply channels.

Considerable cost-savings could be achieved from the merger by consolidating their massive manufacturing operations, analysts said. This would mean a down-sizing of facilities -- there are now eight Hyundai fabs and seven LG semiconductor plants - and their staffs. But most observers expect both sides to fight vigorously to prevent any paring of their operations.

One approach, however, that might work would be to close three aging 5-inch and 6-inch Hyundai fabs and two 6-inch LG fabs. Analysts believed that the two chip makers were already planning before the merger came up to get rid of these plants in the near future.

Another potential battleground is in the United Kingdom. What will happen to Hyundai's postponed Scotland plant and LG's Wales fab that also is on hold? And which fab would get the green light if the new combine decided later on to expand in Britain?

Another problem coming out of the merger is that each chip maker uses a different tool set in its fabs. That has to be harmonized, observers said, if the merged operation is to achieve efficient production economies of scale. In the critical area of steppers, Hyundai has standardized on ASM Lithography's deep-UV tools, while LG predominately uses Canon Inc. i-line systems.

Micron Technology Inc., which is facing the same kind of a problem in its deal to acquire the DRAM business of Texas Instruments Inc., will replace TI's Canon steppers with ASML equipment, its stepper of choice. Micron, however, has the benefit of using deeply-discounted TI financing in the upgrade. By contrast, any Hyundai-LG chip combine will start out cash-short.

Perhaps the biggest challenge the new Korean chip enterprise will face is to cope with the mountains of debt that it is likely to inherit from the parent companies. LG presently carries a debt load of more than $5 billion, which is almost all related to semiconductors, and has a 5-to-1 debt-to-equity ratio. Hyundai is carrying $8 billion in debt, which represents a whopping 9.3 debt-to-equity ratio. Its debt load also relates to telecom and computer operations.

Analysts said that both companies are asking for some kind of financial bailout from the government as the price for merging their chip operations. ING Barings' Han said the two chip firms were asking the Korean government and banks to convert much of their debt into equity in the new company. That could put the new combine in a lot better position to finance R&D and capital spending, he said.

That kind of deal might help to get the new semiconductor company out of its financial quagmire, but would be certain to unleash a strong protest from the U.S. chip industry. The U.S. Semiconductor Industry Association, as well as Micron president Steve Appleton, has pledged to fight any further Korean government bailouts of the chip chaebols.

The Korean chip merger also would give the U.S. Commerce Department a headache when it tries to levy new DRAM dumping duties, which in August set different amounts for each firm -- 9.28% penalty tariffs for LG Semicon and 3.9% for Hyundai. There is little or no precedent for adjusting such dumping penalties.

The merged firm could escape much of these penalty duties by selling DRAMs in the U.S. made at Hyundai's new fab in Eugene, Ore. These chips would be exempt from the dumping tariff, which could motivate the new Korean chip combine to produce as many DRAMs as possible in Oregon for the U.S. market.

semibiznews.com