Here are the chip industry's Top 10 news stories in 2000:
1. The biggest story of the year had to be "boom and bust" in the chip and chip equipment markets.
2000 started out as a heckuva year for both the chip makers and the production equipment suppliers. But in the chip business, a bust follows every boom. And for nearly all of the market and financial analysts, and most of the industry itself, this boom was expected to continue through 2001 and even into 2002.
Markets kept getting better as the year went along. Many chip analysts revised their forecasts upward at least third times in 2000. A panel of equipment analysts in mid-summer agreed that wafer fab investments in 2000 would be 80% higher than last year. Just weeks earlier, they were talking 60-to-70% higher. These analysts also agreed that the growth in chip sales could exceed 45% in 2000. Two months ago, the always optimistic Semiconductor Industries Association issued a 31% forecast.
Chip analysts and researchers who gathered for the Robertson Stephens meeting in July in "Baghdad by the Bay" were their usual bullish selves and pooh-poohed any predictions of an industry downturn. Most of them went along with Sue Billat, Robertson Stephens analyst who "just didn't see any signs of a downturn yet." She predicted the "case is very strong that we will have a very healthy year in 2001 and probably a good 2002 as well."
Earlier in July, Salomon Smith Barney analyst John Joseph upset all the optimists by pointing out how many fabs were under construction and by declaring that the industry had reached the peak of its upturn and was poised to begin its periodic descent. Most of the analysts challenged Joseph's conclusion at the time, but it turned out that he was more right than wrong.
Robertson Stephens' Billat led the attack on Joseph by declaring that demand will remain strong enough to soak up all of the new capacity. "We are still in the first half of a major up cycle," she said. And as far as DRAMs were concerned, the current shortage could last through the end of 2002, figured Jim Handy, Dataquest's memory analyst. At the time, I said that these chip analysts were all "whistling Dixie."
By November, things seemed to be getting a little messy. Big chip customers like Gateway and Micron were reporting lousy PC sales for the important Thanksgiving weekend, but wafer production was going through the roof.
Wafer starts at fabs worldwide took a sudden and somewhat unexpected jump in the third quarter with starts shooting up 7% over the second quarter. This increase came at a difficult time for chip makers, who were trying to deal with cutbacks in new orders by major customers and inventory adjustments in key end-user markets. If this continues in the fourth quarter, "we're looking at wafer starts increasing 25% this year," cautioned analyst Bill McClean, president of IC Insights.
That would be a disaster as well as a "record for wafer-start capacity increases," he says. In 1995, the last boom year for chip growth, wafer starts increased 22%.
Fabs were now running at their highest capacity utilization rate since the mid-1990s. In the third quarter, they were operating at 96.4% of their installed capacity, up from 95% in the second quarter and 90.9% in the year-ago quarter. The worst was yet to come. While capital spending by chip makers was at an all-time high in 2000, climbing nearly 80% from last year, most of that investment in fabs and new equipment won't be felt until next year at the earliest.
This is causing increased concern about 200l. McClean has cut back his forecast for capital spending in 2001 to an increase of 10% or less. "We will see capacity spending turned off pretty quickly at the first real sign of weakness," he predicted. And may have happened.
Intel suddenly delayed the opening of its new, $2 billion Irish fab by one year, worrying analysts. The microprocessor giant maintained it did this so that it could upgrade the Leixlip, county Kildare, plant from a 200-mm to a next-generation 300-mm wafer fab. But some industry watchers are convinced it was the company's response to slowing sales of MPU chips.
Then the entire industry started seeing sales coming in lower than expected. Chip makers didn't seem to be doing accurate market forecasting, because a dramatic slowdown in business from many of their OEM customers caught most of them by surprise. By early December, nearly all of the big U.S. chip makers were rushing to chop back sales forecasts they had made just a few weeks earlier.
Intel was the biggest surprise. The world's leading chip maker blamed a "large cancellation" of orders by customers and lackluster chip demand for revising its forecast. It now expects fourth-quarter revenue to be flat from the previous quarter; earlier it had projected quarter-to-quarter growth of 4% to 8%.
2. Mergers and acquisitions made more industry news in 2000 than they had in years.
Major semiconductor companies were rushing to buy smaller chip companies that sold products in markets they wanted to expand into. And prices went through the roof.
It was hard to get used to seeing these relatively small chip makers going for billions of dollars in stock. Applied Micro Circuits acquired MMC Networks for $4.5 billion in stock--second largest deal ever in the chip business after Texas Instrument's $6.1 billion acquisition of Burr-Brown earlier in the year. A week after TI purchased Burr-Brown and Alantro Communications, it acquired Dot Wireless for $475 million in stock in a move aimed at bringing more strength to its third-generation (3G) cellular phones and wireless devices.
Cisco Systems demonstrated that an excellent way to grow a company was by swapping stock to acquire startups and other emerging companies. Now it looks like Broadcom is trying to do the same thing in semiconductors. The Irvine, Calif., company acquired 15 IC makers in just 18 months. It paid $777 million in stock for VisionTech, an Israeli supplier of digital video/audio compression and depression chips, and it paid a whopping $2 billion in stock for SiByte, a relatively new chip maker in the booming network-processor market.
Intel was one chip maker buying companies right and left--especially in the telecom markets. Ziatech, for example, was purchased for $240 million in cash. The privately-held firm employed only 200 workers.
Conexant Systems formerly Rockwell Semiconductor pulled a Cisco Systems by buying startups with technology that it needed. But these chip startups are expensive to buy. It bought tiny, privately held HotRail, a supplier of high-speed switch fabric and backplane channel technology, for stock worth a cool $400 million. Then just a week later, Conexant said it would shell out another $260 million worth of its stock to buy two more tiny chip companies, Novanet Semiconductor, an Israeli fabless chip supplier, and NetPlane Systems, network software company.
PMC-Sierra greatly expanded its offerings in broadband communications and high-speed transmission equipment by making back-to-back acquisitions of startups Extreme Packet Devices and AANetcom. The British Columbia chip company paid a total of $1.3 billion in stock for the two companies.
Then there were the monster deals with companies we had barely heard of. In one surprising deal, Marvell Technology Group expanded its communications chip offerings by acquiring Galileo Technology for $2.7 billion with a B. By acquiring the Israeli-based Galileo, Marvell became one of the world's leading suppliers of switch-chip controllers for Ethernet networks.
There was one amazing acquisition story that unfolded last summer in the chip distribution world. The world's two leading electronics distributors together bought out the third largest and then proceeded to divide it up between them. As a result, the two leaders--Arrow Electronics and Avnet--now account for half of the worldwide distribution of electronic components.
3. This was the year of the foundries. They're pushing harder than ever to become process leaders and major suppliers to integrated device makers (IDMs) as well as the fables chip companies.
Fundamental power shifts are going on now in global chip production. Look at the rapid rate of change in silicon foundries. Taiwan Semiconductor Manufacturing Co., for one, was going all out to become a production-line innovator. It already claims to have more application-specific process technologies than any other third-party manufacturer. The world's largest pure-play silicon foundry claimed it was the first manufacturer to have a production-ready, 0.18-micron CMOS line for making mixed-signal and radio-frequency devices.
Despite the darkening business climate, TSMC said it is going to keep expanding at its current furious pace for at least the next couple of years. Reason for the huge investment is simply that TSMC wants to become the world's largest producer of chips in 2001. In order to claim the top spot, the Taiwan giant originally expected to invest $3.6 billion in 2001 to add wafer-processing capacity.
That will follow an astonishing increase of 167% in capital spending this year. TSMC expects to fabricate 3.4 million 8-inch equivalent wafers this year--or 6% of all the wafers the global industry is expected to turn out. This kind of increase could help the Taiwanese company become the sixth largest chip maker in 2000.
TSMC has even bigger production plans for 2001. It aims to crank out 4.7 million wafers--a walloping 38% increase over 2000--that would put the Taiwanese company in a first-place tie with Toshiba and Hyundai as the world's largest supplier of chips.
To make it easier to succeed with the initial 300-mm lots, TSMC used a mature 0.18-micron process to fabricate the first customer designs. The company is taking a huge gamble with the new technology: It's already putting up two 300-mm volume production facilities in Taiwan. Fab 12, located near TSMC's headquarters in Hsinchu, is scheduled to start production in the fourth quarter of 2001, while Fab 14 in Tainan is slated to begin manufacturing in early 2002.
4. At long last, the next-generation silicon wafer--a 12-inch monster--finally started moving into production.
In April, we said that the giant platters were coming. The off-again, on-again 300-mm wafer was definitely regaining momentum. Finally. Jumping aboard the bandwagon was a European joint venture--STMicroelectronics and Philips Semiconductors. They will spend $700 million to build a pilot line in Crolles, France, to develop a 0.10-micron process for 12-inch diameter substrates. The final tool set will be selected in the summer of 2001, and the first 12-inch wafers will be processed on the pilot line in 2002.
More than a half dozen other chip companies were in the early stages of ramping their first 300-mm fabs. Racing to become the first to put 12-inch wafers into volume production in 2001 were Infineon Technologies and a joint venture of Hitachi and United Microelectronics (UMC).
As many as 22 of the next-generation fabs will be processing silicon by early 2003. The main driving force is lower production costs. "The 300-mm ramp will dominate the industry's efforts in production during the next five years," said Joel Monnier, R&D director at STMicro. While the industry's earlier switch to the 8-inch wafer was a slow process, the 12-inch wafer "is taking off like a flock of birds," said Peter Kücker, who runs the Infineon-Motorola pilot line in Dresden.
Timing is everything in the industry's massive move to 300-mm wafers. In fact, VLSI Research's G. Dan Hutcheson calls it a "bet-your-company decision." Chip makers rushing to build these big factories face major risks in their timing. The trick will be not to hit the market in the middle of the next downturn--whenever that is.
Construction schedules are being driven by the FUD factor (fear, uncertainty, and doubt). "If you miss next August, you can't ramp and get revenue out of a new line by 2002," said Hutcheson. And that, he added, means "you are at risk of having your new 300-mm facility come on line in the middle of the downturn."
Taiwan Semiconductor Manufacturing aims to become one of the first companies to start up 300-mm wafer production. The first production systems will be delivered to TSMC's Fab 6, which will be used to test the new equipment for eventual use in the company's Fab 7 and Fab 12 lines that are now under construction. The two Taiwan fabs will become the giant foundry's main production lines for 300-mm wafers in the next couple of years.
Despite its increasing caution over 2001, Applied Materials is very confident about the sale of 300-mm wafer gear, which has been coming ever since the late '90s. About $6-to-8 billion of the global revenues from wafer fab systems next year will come from 300-mm tools, it estimated. Applied may pull in about $1.8-to-$2.5 billion of that.
By the end of the year, TSMC had started processing 300-mm wafers for customers such as Altera and claimed that it was the world's first silicon foundry to deliver the new generation of wafers to the market.
Chip makers are counting on 12-inch wafers, which have two-and-a-quarter times the surface area of the current 6-inch workhorse, to result in a significant drop in costs as yields go up. If they do, there's no telling what will happen to prices and margins. It could get ugly.
5. China, the global semiconductor industry's "final frontier," started opening up big time.
China has emerged as one of the hottest locations for semiconductor investments as manufacturers scrambled for new plant sites and stronger positions in the huge market for ICs coming. Several major U.S., Japanese, and European chip makers made big investments in 2000 in China, including Motorola's $1.9 billion complex in Tianjin and IBM's $300 million advanced packaging facility in Shanghai.
Motorola complex will include a $1.5 billion 8-inch wafer fab that could end up setting new rules for U.S. controls over technology in China chip plants. U.S. government regulations have limited technology transfers to 0.35-micron design rules in China, but Motorola reportedly has gotten approval for 0.25-micron production in Tianjin.
Startup investments also were made in Shanghai and other locations, including new foundry companies backed by investors from Taiwan. Shanghai's Grace Semiconductor Manufacturing broke ground for its initial 200-mmwafer fab in November. This venture also broke political ground with an investment team headed by the sons of China President Jiang Zemin and Taiwan plastic tycoon Y.C. Wang.
China's political and business leaders also are trying to promote Shanghai and the eastern region of the country as the new "Silicon Harbor." In 2000, a policy directive issued by the State Council of China and implemented by the Shanghai municipal government provided preferential treatment to IC makers. To support the semiconductor expansion, the Chinese government okayed a $1 billion investment to build the country's largest silicon materials plant for polysilicon wafer blanks.
There was good news too out of Washington, D.C., where the House of Representatives voted to establish permanent trade relations with China. Once the Senate approves it, the U.S. chip industry will be closer to gaining greater access to China's $8 billion annual semiconductor market.
A lot of action was going on in 2000 in the China cellular-phone business, as several U.S. companies signed a deal to participate in that soaring market. Motorola was the latest U.S. chip maker to jump in. Motorola is licensing its chip technology to Eastcom (Eastern Communications) and China Integrated Circuit Design Center (CIDC) to go into new handsets for the China market. The three companies will codevelop handsets based on the global system for mobile communications (GSM) digital-cellular standard.
Texas Instruments and Qualcomm already have deals with several OEMs in China to get a bigger piece of the world's fastest growing cellular phone market. The state-run China Ministry of Information Industry figures the Chinese cell phone market will grow from 68 million subscribers in 2000 to nearly 250 million by 2004.
Another U.S. company that expects its China business to boom over the next few years is Applied Materials. CEO Jim Morgan was certainly bullish on China. The leading semiconductor equipment maker wrote less than 10% of its latest quarterly sales in China, but Morgan looked for this business to grow due to China's upcoming entry into the World Trade Organization.
He expects the world's chip makers to set up new plants in China to meet the soaring demand for products going into computers, mobile phones, and Internet-related gear. Applied's sales in China will grow 500% over the next few years from its current annual sales rate of $50-to-$100 million, Morgan said. These sales "should grow to a half billion or more," he predicted. Indeed, China is expected to become the world's second-largest semiconductor market by the end of this decade.
6. The rise and fall at least for now of the Rambus DRAM.
The big news here was that Intel backed off its Rambus-only memory strategy for its next-generation microprocessors and confirmed reports that it was designing a chip set using today's standard memory for its upcoming Pentium 4 microprocessor. The news came as a big surprise to many in the industry.
Rambus is the designer of the Direct Rambus DRAM interface, which has been competing with synchronous DDR memories for the next-generation of higher-speed computer memories.
In July, it became obvious that Intel was shifting its weight from backing only the RDRAM from Rambus to leaving the memory choice up to its MPU customers. Intel made big headlines when CEO Craig Barrett admitted that "we made a big bet on Rambus and it did not work out. In retrospect," he said, "it was a mistake to be dependent on a third party for a technology that gates your performance."
The Rambus "go for broke" drive to dominate the next-generation memory market ran into another roadblock when DRAM vendor Micron Technology filed an antitrust lawsuit against the memory technology provider. The Boise DRAM maker accused Rambus of violating antitrust laws by trying to "coerce exorbitant royalties" from the DRAM industry. Micron claimed Rambus was unduly coercing makers of synchronous DRAM and double-data-rate SDRAM chips as part of a strategy "to file lawsuits against those manufacturers who do not agree to Rambus' non-negotiable license terms."
7. The management change that most surprised the semiconductor industry in 2000 was Hector de J. Ruiz's jump from Motorola to Advanced Micro Devices.
The architect of Motorola's extensive restructuring of its chip business stunned the company when he suddenly bolted and joined AMD as president and chief operating officer in late January. The 54-year-old Ruiz had worked at Motorola for 22 years and had been the president of the company's Semiconductor Products Sector for nearly three years when he decided to join Jerry Sanders.
Some industry observers believed that Ruiz's departure might have taken the steam out of Motorola's plans to shift a significant portion of its chip manufacturing to third-party contractors and joint-venture fabs. Officials denied reshuffling the sector's strategy.
Caught completely by surprise by Ruiz's move, Motorola quickly named Fred Tucker, who was deputy to the company's office of CEO, to replace Ruiz. Eight months later, Motorola turned the reins of its chip business over to veteran executives Fred Shlapak (as sector president) and Bill Walker (as general manager).
Motorola, the world's sixth largest chip company, had been on an aggressive strategy to outsource 30-to-50% of its chip manufacturing, but now it appears to be backing off those targets set under Ruiz's command. Shlapak said Motorola's target for wafer foundry use was now 12-to-15% of its total frontend capacity requirements.
AMD, which had looked hard for a new president who would be a leading candidate to succeed Sanders as CEO, came up with a great package to get Ruiz. The former Motorola exec was promised at least $1.5 million in the event the board does not elect him to replace Jerry by 2002. Ruiz appeared to be the heir apparent when he was hired, but the plan got a lot clearer when AMD filed its annual proxy statement. He is due to receive a lump-sum severance payment equal to two years of his current annual base salary and the vesting of his initial option will be accelerated by two years. Ruiz now earns a $750,000 annual salary and has an option to purchase 1 million shares of AMD stock. Not bad.
8. The "bare knuckles fight" between Intel and AMD for bragging rights to the industry's fastest microprocessor.
It was a battle that AMD ended up winning at least temporarily and one that caused Intel all kinds of angst. My mom always told me: haste makes waste. And that's what some Intel customers said about the chip maker's sudden recall of its new, Pentium III processor. The processor giant had been pushing out new products faster than ever to keep ahead of AMD.
The recall turned out to be an unmitigated public relations disaster--it came on the same day that AMD announced that more than 10 PC makers were now selling systems with its new 1.1-gigahertz processor. "Intel obviously rushed out with the 1.13-gigahertz Pentium III before it was completely ready--just to be able to show something competitive with AMD," said Bert McComas of InQuest Research.
Last year Intel found itself in a rare spot for it to be in. The chip maker's next mainstream processor, the Willamette, looked to be lagging in performance when compared with new versions of AMD's Athlon. So Intel added a higher performing model to its aging Pentium III line to keep up. AMD's new, high-speed desktop Athlon will be approaching 1.5 gigahertz in 2001, so Intel decided to shrink the Pentium III die to boost its frequency to 1.4 gigahertz. "Intel couldn't afford to wait on developing a mainstream desktop Willamette chip, so they've returned to the old tried-and-true Pentium III core as a quick fix," said InQuest's McComas.
Then there was the Pentium 4. When 85% of your business is in just one line of products--and you're about to introduce your first new design there in five years-it is definitely a "bet your company" situation. That's true even if you're the biggest chip maker in the world. That's where Intel stood as it rolled out its long-awaited Pentium 4. "The roll out of the Pentium 4 is critical to the company's whole future," declared Tony Massimini, analyst for Semico Research. The processor business is "still the cash cow for Intel," he pointed out. This story is certain to continue as a big news story in 2001
9. The biggest story in the semiconductor production equipment industry was ASM Lithography agreeing in October to acquire U.S.-based Silicon Valley Group for $1.6 billion in stock.
That put the Dutch company is a strong position to challenge Nikon's long-time grip on the top position in photolithography systems. The surprise move would not only move ASML up in market share but it also would give the it additional sales in photoresist track systems and thermal processing equipment. Owning SVG also would strengthen ASML' presence in U.S. and help to ease American concerns over the company's involvement in the U.S. government-backed next-generation lithography R&D programs.
The sale is now on track to be completed in early 2001. At least one analyst, however, disputes ASML's claim that the acquisition will make it No. 1 in the photolithography market. Robert Castellano of The Information Network said Nikon sold a total of 270 lithography systems in 1999, while the combined ASML/SVG Lithography operations shipped 245 systems last year. In any event, the litho market should be a real horse race now.
10. Jerry Sanders bet the store, now Advanced Micro Devices is winning big time.
Just about every business publication and financial and market analyst has dumped on the AMD CEO in recent years as the chip maker continued to disappoint them. I couldn't help but keep rooting for Advanced Micro Devices through its long trials and tribulations. I've been following AMD for so long that Sanders once told me that he was going to hit a record $20 million in sales that year 1971, I think. AMD has come a long way since then. I've always thought analysts underestimated this company and the evidence in 2000 finally showed that I was right.
It started looking good when the chip maker reported second quarter sales of $1.17 billion, nearly double last year's results. And net income hit $207 million, up from $189 million in 1999, handily beating Wall Street expectations. Making this good news happen was strong revenue growth in AMD's two principal product lines-PC processors and flash memory chips.
Best time for putting on a "dog and pony" show is when you're riding high. AMD and Sanders certainly know this. So last fall, with archrival Intel scrambling to look good after a string of disappointments, AMD laid out its plans to analysts and the press for the coming year.
The Sunnyvale, Calif., chip maker certainly was bullish. Sanders saw sales growing faster than the 22% predicted for the global chip business. By now though he might be revising those forecasts. The chip maker was expecting sales of $4.8 billion and shipments of 38 million PC microprocessors in 2001.
AMD was also more bullish about flash market growth than the SIA. AMD predicted sales this year of $11 billion, up from the SIA's forecast of $10.5 billion. . The chip maker anticipated flash production volumes reaching 350 million units next year. It plans to expand capacity at a 74% rate, well below the 102% demand increase that analysts were then predicting.