[Engineering Feature]
Key Companies Shake Up This Year's Top Employers
It’s a tough time for the industry, but several companies have grown and even thrived despite the recession. A closer look beyond the numbers shows how they achieved their success.
A further risk is that almost 36% of 2008 sales were made up by its top five customers, such as Apple and HP. The company keeps driving this percentage downward. Yet if Apple decided to use a different chip supplier for its iPod, for example, Broadcom could face a significant problem. Broadcom supplies a key component of the iPod, a video-multimedia processor chip, making around $8 for each iPod—that’s significant, considering iPod sales figures.
ALTERA TAKES PLD COMMAND Altera Corp. saw a similar leap forward, catapulting 57 places from its 86th finish in 2007 to 29 in 2008. In 1984, the San Jose company invented the world’s first reprogrammable logic device (PLD), which is an IC that customers can program to perform multiple functions.
This contrasts with ASICs, which can only be programmed once by the manufacturer to perform only one function. The added adaptability makes PLDs more valuable, since they can be used multiple times for different projects. However, PLDs are slower than custom-designed silicon solutions.
Altera’s PLDs are standard ICs that allow customers to program and personalize the application of the chip to provide market differentiation. The company serves more than 13,000 customers in four primary market segments: communications, industrial, consumer, and computer and storage.
Its main products are the Cyclone, ArriaGX, and Stratix series of FPGAs; the MAX series of complex programmable logic devices (CPLDs); the Hardcopy series of structured ASICs; and the Quartus II software (Fig. 2). The HardCopy process transitions the FPGA design, once finalized, to a form that is not alterable to reduce design security risks.
The Stratix families are larger, faster devices, with more features than the Cyclone families. They also cost more. Stratix IV FPGAs made Altera the first PLD company to ship devices at the advanced 40-nm node. The GX parts contain dedicated high-speed serial transceivers. The Cyclone and Arria families comprise lower-cost, smaller devices meant for less demanding applications.
Leading to this success, Altera has done a very good job of managing its financials while continuing to gain market share and improve its return on equity, which is a tough balancing act.
While scoring a five or higher in many ranking categories, Altera showed a two-point or more improvement in employee growth, sales growth, pretax income growth, pretax income margin improvement, and ratio of closing stock price to high stock price, offset by a fourpoint loss in year-over-year change in patents issued.
The programmable logic market, both FPGAs and PLDs, has gross margins in excess of 60%. It’s controlled by Altera and its competitor Xilinx, the FPGA founder and market-share leader. The two companies concede a small market share to Lattice Semiconductor, Actel, QuickLogic, Atmel, and Cypress Semiconductor. Lattice represents less than 10% of the market. FPGA makers Actel and QuickLogic sell to a lower-end market segment that Altera mostly does not address.
In broader terms, Altera competes with ASIC, structured ASIC, and zero mask-charge ASIC companies like eASIC. In recent times, FPGAs and structured ASICs have become powerful enough to compete head-to-head with DSP devices, microcontrollers, and virtually every other embedded product. Moore’s Law and improving software tools are rapidly expanding potential markets for FPGAs.
Altera and Xilinx have a combined market share of over 80%, with Altera somewhere around 35%, so it may have more room for growth. While both offer PLDs, their products are different enough that their chips can’t be used together in the same device. Engineers using these chips must invest time to learn proprietary software tools used to program the PLDs, creating very real switching costs. As a result, it is difficult for Altera to attract customers from Xilinx and vice versa, so the competition is in the design phase.
Both companies make chips that can be programmed to perform any function you want. Basically, they make “blank slate” chips that can be configured to implement any logic function that you can imagine. In fact, Intel uses huge boards made of tens of these chips to ensure that their logic works before they get their CPUs fabricated into silicon. Smaller ASIC companies have become smarter over the last decade and make prototypes of their chips (using programmable logic) before they start mass production.
Attracted by the high margins, many companies have tried to compete with the “Big 2” such as Intel, Motorola, AT&T, and AMD, but all have failed. Although the business is profitable, it is also mature. If you try to increase market share, you most likely will decrease your margins.
Altera has also benefited from a shift from ASICs to PLDs. Historically, most of the chips in the semiconductor industry have been ASICs because of their lower per-unit costs (assuming substantial production), while PLDs were used more for prototyping, where the reprogrammability of the chips is most valuable. The ongoing shift toward PLDs is being driven by several factors.
First, the cost of PLDs is falling due to technological development associated with their manufacturing. Next, electronics have shorter life cycles, so the buyers using PLDs see less benefit in customizing ASICs. The shorter life cycle reduces the long-term benefit of ASICs (cheaper per unit), and the developmental cost to ASICs exceeds this benefit in the short term.
Third, more advanced chip manufacturing technology adds a premium on ASICs. The complicated designs increase the chance of failure during production, making ASICs cost more. Finally, the PLD market is still relatively small compared to the ASIC market, so PLD market growth needs to continue for Altera to keep growing profitably.
Reprints
