The days of wine and roses for electronics buyers will continue through much of 2002 as component availability remains high and prices fall through at least the first half of the year. However, some leadtimes will stretch and prices won't fall as much as they did in 2001. There won't be much change until high component inventories finally get worked off.

The bad news is that supply headaches for buyers will likely occur in the fourth quarter as end equipment demand picks up and the effects of consolidation of the supply base and supplier production cutbacks kick in.

However, for the first three quarters of the year buyers can expect ample supply for most semiconductors including DRAM, flash, microprocessors, logic and digital signal processors. There shouldn't be any availability problems for passives, connectors and electromechanical devices even though inventory levels have dropped down and suppliers have cut back production.

Component demand won't likely pick up in earnest until the third quarter and buying conditions will continue to favor buyers. Three to five percent quarterly price declines are still likely for many semiconductors for the first half of the year and passives will likely drop 5-8% for the year. Because demand is not likely to pick up until the second half, many analysts, component suppliers and distributors are predicting a flat to down year.

"Our prediction is that 2002 will be unhealthy for the entire year," says Jim Sogas, vice president North American sales for Elpida Memory, Santa Clara, Calif. He notes that historically DRAM bit shipments grow 60% per year. In 2002 bit shipments grew only in the 30-40% range. "Even if DRAM came back to 60% bit growth, which we think is highly unlikely for 2002, we are still oversupplied between inventory and pure production capability," he says. "We think we will see a balance not until the first half of 2003. After that there will be shortages and we will get back to normal growth rates," he says.

Some distributors are also less than bullish about 2002.

"Business is about the same. The bottom appeared to be the fourth quarter of 2001 and the indications are that it's not going to get any worse," says Tom Pitera, president of distributor Pioneer Standard's Industrial Electronics Division. "But we haven't see any signs of improvement. This year could be flat or down slightly. It's hard to tell at this point in time.

He says he expects the communications segment to be weak in 2002 and contract manufacturing should be sluggish until the second half at least.

Industry analysts don't agree on whether or not recovery will take hold in 2002. Semico Research is bullish and forecasts a 19% growth rate for the semiconductor industry. However, Cahners In-Stat forecasts a 6% decline and IC Insights forecasts 1% growth for the chip industry. The Semiconductor Industry Association sees about a 6% growth rate.

The odds are there will be modest single-digit growth for the chip industry which means that supply conditions will favor buyers for most of the year. However, all good things must come to an end and the buyers' market that purchasers have reveled in will wind down in the fourth quarter with some parts shortages likely. Prices will also likely bump up in Q4 and a full-scale sellers' market is likely in 2003.

With that in mind, in 2002 electronics purchasers would be wise to qualify an additional supplier or two for critical components for the equipment their companies make. Flash, DSP and tantalum capacitors are likely candidates for shortages if the economy recovers and demand for equipment picks up. Longer term DRAM buyers may have more than their share of supply challenges because of consolidation.

While buyers have enjoyed low prices for DRAMS for several years, bargain basement tags have had a major impact on the DRAM supply base. Consider this: in 1995 the DRAM market amounted to $40 billion. In 2000 it was $28 billion, and in 2001 it was just $11 billion, according to semiconductor market researcher IC Insights. The decline was not due to a decrease in bit demand, but because of oversupply, which drove down prices and revenue for DRAM manufacturers. As a result, many companies exited the DRAM business or have merged operations. In the past several years, Hyundai and LG Semicon merged operations and formed Hynix. Now Micron plans to acquire Hynix. Micron had previously acquired one of Toshiba's major DRAM fabs and Toshiba plans to get out of the DRAM business. NEC and Hitachi merged to forge Elpida.

If Micron acquires Hynix there will be four companies—Samsung, Micron, Elpida and Infineon—which own 80% of the DRAM market. That may be healthy for the DRAM industry as the remaining key players should be able to turn a decent profit in the coming years because of less competition. In addition, bit growth will likely continue and DRAM will be used in more electronic equipment such as cell phones and consumer electronics. However, for buyers it will mean fewer supply choices and higher prices in the coming years.

For this year, however, DRAM buyers won't have any sourcing problems even though prices have risen and leadtimes stretched recently.

Purchasing Magazine's monthly survey of buying conditions shows that leadtimes for many DRAMS increased from four weeks in October to six weeks in December. At the same time spot market prices, which were at rock bottom, increased by about 10% for some parts.

However, that trend won't continue through the year, according to Jim Feldhan, president of Semico Research. "We believe there is ample capacity out there. If you look at DRAM, average prices will increase this year because higher density parts will ship," he says. Higher-density DRAMS carry higher price tags.

"However, if you look at individual densities, the prices will still decline, but not at the same rate as last year," says Feldhan. "Prices should drop 5-10% per quarter. In 2003 and 2004 there will be shortages and prices will go back up."

Much of the price decline will be with 256 Mb parts, says Feldhan. "If you look at 128 Mb and 256 Mb parts, we see price parity between those two by mid-year. We are seeing falling prices on 256 meg through the year and 128 meg parts will stabilize," he says.

Beside the transition to higher density 256 DRAMS, there will also be a transition from synchronous to double data rate (DDR) architectures. A lack of chipset support for DDR had been an issue, but now there are several manufacturers of DDR chipsets including Intel, which originally had only supported Rambus DRAM.

Feldhan says there may be some shortages of DDR parts if the transition to DDR occurs faster than expected. "There will be spot shortages, but they won't be long-term," he says.

For the year, most DRAMS shipped will be synchronous, but in the fourth quarter most DRAMS will be DDR, according to Feldhan. In 2003 most DRAMS shipped overall will be DDR.

For 2002, buyers can expect to pay a 5% premium for DDR DRAM over synchronous. However, RDRAMS will continue to cost about two times what synchronous DRAMS cost.

Because of the cost, RDRAMS will be relegated to high-end machines for the most part. DDR will start off the year in mid-range computers, but move into lower cost machines by the end of the year, according to Feldhan.

"If Dell sells PCs for $699, there isn't a lot of margin to pay hefty premiums for memory," he says. "There will be a place for RDRAM, but it won't be in the mainstream."

DRAM will be in ample supply for most of the year, as will other components despite a pickup in demand for end equipment.

For example, computer shipments are expected to rise from 118 million units in 2001 to 125 million in 2002. Cell phone handset shipments will also rise from about 395 million units in 2001 to 425 million this year. However, much of that growth will not occur until the fourth quarter. The increase in end equipment demand coupled with falling component inventories and scaling back of component production could mean some supply problems for buyers in the fourth quarter. However, fortunately the real supply migraines won't be felt until 2003.

While computers traditionally drive semiconductor sales, other end equipment segments are growing in importance to component suppliers. The boom in cell phone shipments in 1999 and 2000 resulted in shortages of flash memory and passive components. Another boom in cell phones will likely begin in the fourth quarter when Web-enabled 2.5G phones ship in the U.S. If a cell phone boom occurs, there will be greater demand for higher density flash parts and DSPs as well as capacitors and resistors.

Buyers should keep a close eye on the automotive industry and its supply chain. Vehicles are being equipped with more safety and entertainment systems, which tend to require a variety of semiconductors such as microprocessors, DSPs, flash memory.

"Automotive has not been a big portion of the semiconductor industry, but they are adding a lot of features that have a lot of semiconductor content such as global position satellite (GPS) systems," says Feldhan. "Those systems used to be in high-end cars but are now moving down to less expensive vehicles. GPS has a DVD drive, an RF section, a high resolution screen, memory and a high end microprocessor," he says.

While equipment such as computers, cell phones and networking equipment are large consumers of semiconductors, they are also big users of passives, connectors and electromechanical devices. Last year, many end equipment segments had negative growth and there was an excessive amount of passives in inventory at end equipment manufacturers and contract manufacturers facilities.

Sean Wood, an analyst at iSuppli, says that the inventory in the field was between 20-26 weeks in 2001. "Now the inventory in the field is at normal levels which is four to six weeks, less than 10 anyway."

However, the inventory at suppliers is harder to pinpoint.

"Suppliers are still building inventory because they are trying to run their factories but are having a tough time trying to figure out what the forecasts are going to be going forward," he says. Wood estimates that suppliers have about 18 weeks of inventories at their factories. "That is manageable for them. It's better to have inventories on their shelves than to have it out there. They can get an order and ship it and have control over it. Before, the orders might have been coming into distribution," he says.

Despite inventories being reduced from last year, buyers can expect passive tags to decline through most of the year. However, low prices and weak demand will create some future supply problems for buyers because most major passive manufacturers are losing money and may decide to stop making certain products.

"Of the four major North American component manufacturers, only one is making profit," says Wood. The one is Kemet. The others are losing in excess of $25 million a quarter, he says.

"If you look at the cash position of each of them, some of them are going to run out of money," says Wood. "Right now it is cheaper for them not to turn the lights on and go to work then it is for them to go to work. The cost of running the factory is killing them. Some of them are losing $30 million a quarter and they don't have $30 million in the bank.

"They can't hold out for long," he says.

Some suppliers will start to pull out of certain product segments and focus on core competencies. "They may make one or two parts which are profitable and drop the unprofitable components," says Wood. "Someone will likely drop out by the third quarter and there will be tightness of supply by the fourth quarter. The guys who are losing the most money are the tantalum cap manufacturers," says Wood.

That may be hard for electronics purchasers to believe because two years ago there was a severe shortage of tantalum capacitors and buyers had to scramble to find them and had to pay through the nose.

However, that component shortage was caused by a shortage of tantalum powder. "Tantalum powder was the weak link in the supply chain, but there was a of lot powder capacity added last year. Capacity for powder went up 20%. It will go up again this year. So for 2002 and 2003 there will be no shortage of powder," he says.

That would seem to be good news for buyers, of course, but there could still be a shortage of tantalum capacitors late in 2002 and in 2003 because tantalum capacitor manufacturers are losing a lot of money and someone will likely drop out.

"A supplier will say he is not making that part any more because he can't make money on it. All of a sudden there will be some buyers who have a hard time sourcing tantalum caps," says Wood. "Supply will tighten and prices will rise."

However, there is a chance that tantalum capacitor demand won't be as strong as in past years even if cell phone shipments skyrocket in 2002. Reason: After the last tantalum capacitor shortage, many cell phone manufacturers designed out as many tantalum caps as they could from their phones. Many used ceramic capacitors. Some newly designed phones will be using niobium capacitors.

"They (niobium caps) are not direct drop-in replacements, but they play in the same space," says Wood. Tantalums can provide high capacitance and low ESR (effective series resistance). Niobium capacitors don't perform as well as tantalums, but they are less expensive. Ceramics can deliver high capacitance, but there are ESR issues.

"Anyone who needs to find a substitute for tantalum is going to be able to do that very easily," he says.

While prices of semiconductors and passives will decline in 2002, so will connector tags.

Connector price erosion is usually in the 2-4% range, but last year it was double that, says connector industry researcher Ron Bishop. He expects price to fall about 5-7% in 2002.

"We still have an overcapacity situation, high inventories and not a tremendous amount of demand. That means lower prices."

It also means there will be consolidation in the connector industry which declined about 20% in 2001 from $30.5 billion in 2001.

"I think something will happen with one or two of the top 10," says Bishop. "There is a tremendous amount of price pressure to consolidate at the low end with small and mid-sized companies.

But consolidation is nothing new to the connector industry. In recent years, connector industry leader Tyco alone has acquired AMP, MA/COM, Thomas and Betts and Augat. Now Tyco wants to break out its electronics sector as a separate business.

Consolidation coupled with an economic recovery, however, could mean longer leadtimes and higher prices. Because connector business declined, "everybody has gotten conservative," says Bishop.

"Companies have gone through a period of having too much inventory, too much capacity. They are not going to build inventories or manufacturing capacity. They have adopted a wait and see attitude."