A familiar pattern has returned to the computer industry: Memory prices are down and computing power is up.
It`s the kind of news that brings smiles to an industry that relied on these two fundamentals to achieve its amazing growth through the late 1970s and much of the `80s, when it wasn`t unusual to see 20 percent sales increases in a single year.
But as it approaches the `90s, the industry finds itself in an entirely different situation than it did during those innocent early years, when new, distinct products were popping up left and right and sold despite their flaws.
Now the customer is smarter and more sophisticated. Standardization has made it easier for competitors to imitate products made by the larger firms, giving users wider choices and better prices.
In the personal computer market in particular, incompatibility is not nearly the problem it was when the industry was young. Many items in a PC system, from disk drives to interfaces, are now common commodities.
For all these reasons, said Kenneth Flamm, computer industry analyst for the Brookings Institution, a think tank based in Washington, D.C., ”This will be a very competitive year in the industry.”
The U.S. Commerce Department predicts a 10 percent rise in shipments of various products made by the U.S. computer industry this year, a sign that the industry is maturing. Total shipments are expected to be in the neighborhood of $67 billion.
Tim Miles, a senior industry analyst for the Commerce Department`s International Trade Administration, said this represents a slippage from last year`s 14 percent increase in shipments, when computer demand was strong despite a shortage of memory chips.
While the PC market in 1989 is expected to be strong, he said, sales of mainframes, still the workhorse among industries with large computing needs, are expected to be softer because of a slowing economy.
Many analysts believe that 1989 will bring more industry consolidation because of the slowing market and because industry giants want to achieve greater economies of scale in order to compete with growing foreign competition.
But Flamm said he doesn`t buy the consolidation theory, because standardization is enabling more companies to get into the market.
Charlotte Walker, vice president of County NatWest USA, a New York-based brokerage firm, said consolidation already is a fact among firms dealing with older mainframe and mini-computer technology. Firms that are trying to support these older technologies while trying to keep up with the new technologies are finding it difficult to juggle both, she said.
But Flamm said this phenomenon merely gives new players an opportunity to compete against the ”big boys,” because the smaller firms do not have the baggage of trying to support older technology.
The softening of memory prices came about because the Japanese stepped up production, apparently because of pressure from the White House, Flamm said. The shortage traces back to 1987, when President Reagan imposed trade sanctions against Japan for dumping cheaply priced chips into the U.S. Not long after, Japan scaled back on chip production, causing memory prices to rise.
But with so much chip production now overseas, it is still unclear whether the declines in memory prices that accompanied the industry`s fantastic growth in its early years will be resumed.
”The question is whether we will see a return to the historical rate of decline in memory prices or whether we are in a new ball game,” Flamm said. ”I think we`re in a new ball game. I wouldn`t count on supercheap memory prices.”
Flamm said that, according to his government contacts, Japan has told the Bush administration that Japanese chip producers will not flood the market with semiconductors called DRAMs, for dynamic random access memory.
The implication is that while the Japanese will continue to cooperate and meet demand in the U.S., they will not play ball to the point they drive prices down dramatically, to their detriment.
Miles, of the Commerce Department, said that another factor could bring the DRAM shortage back: the race to build the first four-megabit memory chip. He said that if the Japanese manage to beat U.S. producers to the punch, the chip shortage could resume. He added that the four-megabit DRAM could be developed this year, and ”the big question is who will be first.” U.S. producers are essentially down to three: Micron Technology Inc., Boise, Idaho; Intel Corp., Hillsborough, Ore.; and Texas Instruments Inc., Austin, Tex.
The memory-price issue worked to the detriment of Apple Computer Inc., the Cupertino, Calif.-based firm that suffered losses because it bet that memory prices would stay high.
According to Andrew McVeigh, a semiconductor industry analyst for Rodman & Renshaw Capital Group Inc., a Chicago-based brokerage firm, Apple stockpiled the higher-priced chips to make sure it had enough to meet demand for its 1988 product line. But time worked to its disadvantage, because chip prices began to fall as Japanese production rose.
”Apple got stuck with some high-priced inventory, but I think their losses were temporary,” McVeigh said. ”I don`t think they`re in any trouble at all.”
Chip prices are critical to the thousands of corporate and individual customers wondering when to buy computer systems and software that eats up large chunks of memory. McVeigh said the high prices last summer were an exception and that prices should resume the historical decline experienced during the industry`s early years.
Ironically, one solution to the memory shortage issue may come in computer design. The leading edge of design is in the so-called RISC (for reduced instruction set computer) microprocessor pioneered by several major chip producers, including Motorola Inc. in Schaumburg and Intel.
The RISC microprocessor represents a major simplification in the way data are processed inside a computer, therefore enhancing the speed of the microprocessor several fold.
The catch is that it relies more heavily on software than do current-generation chips for performing applications, said Millard H. Phelps, analyst for Hambrecht & Quist, a San Francisco-based investment banking company. Development of this new kind of microprocessor seems destined to get computer firms to put more emphasis on software development than they have in the past.
Intel electrified the industry when it announced its new RISC i860 (also known as the N-10) chip on Feb. 27. The chip is so fast, perhaps 10 times speedier than previous ones, that it can perform many of the same duties as a supercomputer, the company said.
”In a marketplace that is speed-oriented, this can be very powerful,”
said Jay Samstag, an analyst covering the semiconductor industry for Duff & Phelps Inc., a Chicago-based firm that conducts market research and rates the credit of corporations.
For many scientific and academic uses and some business uses, the power of supercomputing can be brought to the desktop level, Hambrecht & Quist`s Phelps said. Computers incorporating the chip would feature advanced three-dimensional graphics more vivid and flexible than anything now on the market and enable more advanced design to be performed at the desktop.
The key is that development of the chip represents a further decline in the price of computing for highly sophisticated uses. It`s estimated by experts such as Phelps that a machine with the i860 would cost from $10,000 to $15,000, a mere drop in the bucket to the millions a supercomputer costs or a supermicrocomputer, which runs into the hundreds of thousands of dollars.
Flamm noted that Intel doesn`t have a corner on the market. Motorola, Sun Microsystems Inc. of Mountain View, Calif., and International Business Machines Corp. of Armonk, N.Y., among others, are developing and testing RISC- based machines. But the i860 chip already has excited manufacturers, with IBM announcing plans to use it in connection with a high-performance graphics card for its PS/2 computers.
The new Intel chip will ”allow artificial intelligence at the desktop,” Rodman & Renshaw`s McVeigh said. ”You will be able to build in expert systems. Hospitals, for example, may be able to use computers to not only keep records of various medicines a patient is receiving, but also to do cross-checking to see if any of the medicines are incompatible.”
Another positive side of the RISC technology is that it may spur the industry to fill a big gap-software development.
Customers long have criticized the industry for focusing too much on hardware development while neglecting software. Indeed, the potential of chips that are now being phased out-such as Intel`s 8088 and 80286-has never been fully realized because the industry is leaping ahead to the next chip before fully supporting the last.
”But this is changing,” Miles said. ”Many companies are putting more money into software packages. Look at IBM. It recognizes that hardware has become a commodity, and it has to make it in the future off software and services. They are not alone. I find more and more firms expanding their software operations.”
One reason is Japanese competition. Miles said Japanese firms, to make up for a general weakness in their software development, have been buying U.S. software firms left and right in order to close the gap. ”We`re still ahead of the Japanese in software,” he said.
Walker, of County NatWest USA, said, ”Software (development) has always lagged behind hardware.” She added that the new chip ”is going to create a bigger gap. It`ll create an opportunity for new players, too.”
She predicted that the software industry will experience solid growth of 15 to 18 percent in 1989.
The RISC chip will run on the Unix operating system, another feather in the cap for a system originated by American Telephone & Telegraph Co. ”I think it`s a very big boost for Unix,” Walker said.
