A Near-Duopoly Within the Semiconductor Space

The semiconductor industry is a vast, $300 billion landscape filled with a plethora of companies of all different sizes. All compete in various market segments, from computer processors and memory chips, to power management devices and chips that control the touchscreen on your smartphone. Dig a little deeper, and you'll find two of our favorite firms,  Altera  and  Xilinx , both of which reside in the programmable logic device, or PLD, chip space. While the PLD segment only accounts for just over 1% of total chip industry revenue, Altera and Xilinx essentially operate in a duopoly, thanks to high barriers to entry in the market.

What's a Programmable Logic Device?
PLDs are chips that can be programmed so the circuitries in the device can be configured by electronics makers for specific applications. This is different from the more conventional high-volume applications specific integrated circuits, or ASICs, which are created to handle specific functions and cannot be reconfigured. We usually find ASICs in our everyday electronic gadgets. The trade-off between PLDs and ASICs is cost. Although PLDs are programmable, they are larger in size and may have transistors that are left unused, making them more costly on a per-unit basis for electronics manufacturers. ASICs are significantly cheaper to manufacture in mass volume, but have much higher upfront startup costs. The choice between a PLD and an ASIC boils down to whether the electronics maker will see enough volume for its products to bear the higher up-front costs associated with designing an ASIC. As a result, PLDs are usually found in lower-volume electronics, such as those in industrial, automotive, military and communications infrastructure applications.

Within the PLD segment, we estimate that Xilinx holds about 50% share, while Altera has about 35% of the market. Although there are many smaller players in the space, such as Lattice Semiconductor (LSCC) and Actel (now part of Microsemi ), Xilinx and Altera have always dominated the PLD space because of their economic moats. The two firms benefit from high switching costs, as engineers tend to be trained and become well-versed on either programming Xilinx or Altera chips. Given the familiarity that engineers gain with the software and design tools associated with each platform, they tend to stick with a specific firm's products over time.

Tailwinds for the PLD Space
Altera and Xilinx are poised to benefit from several favorable trends. As semiconductor technologies continue to advance, the initial setup costs of creating cutting-edge ASIC chips has risen exponentially. This has gradually lifted the volume threshold necessary for ASICs to be cost effective, and in turn, has increased opportunities for the PLD market. One notable example of a new opportunity for PLDs is the high-definition 3DTV space, where Xilinx, in particular, has earned some new design wins in TVs that previously would have never considered PLDs to be a viable solution. Altera estimates that although the PLD market in 2009 was only about $3 billion in size, $30 billion of the $82 billion ASIC market is now accessible to PLD makers. As a result, Altera and Xilinx have historically outgrown the semiconductor industry, and we expect the two firms to continue to do so in the future.

More importantly, we believe that the two names provide indirect plays for the rapidly burgeoning smartphone market. The proliferation of smartphones has, and will continue to, dramatically increase the amount of data that is being transported on wireless networks, requiring carriers to expand and upgrade their infrastructures, which contain a lot of PLDs. In fact, the two firms held up better than other chipmakers during the severe cyclical downturn in the semiconductor industry in the 2008-2009 timeframe, as China's buildout of its latest 3G wireless network helped boost demand for base stations, and thus, PLDs. Xilinx and Altera are also poised to benefit from the global deployments of next-generation 4G networks, which will be capable of handling higher smartphone data traffic. Communications-related PLD sales accounted for 47% of Xilinx's and 44% of Altera's revenue in 2009, respectively. According to Altera, the firm gets double the dollar content of PLDs in a 3G base station versus a 2G base station, while a 4G base station is expected to contain three times the PLD dollar content than that of a 2G base station.

Near-term Outlook
In recent quarters, Xilinx and Altera have been boosted by a robust rebound in the semiconductor industry, following the cyclical downturn during 2008 and 2009. Sales growth of PLD chips has been driven by strong communications-related chip demand, thanks to global wireless infrastructure buildouts, as well as healthy demand from the industrial segment.

The firms' latest results were solid. For the September quarter, Xilinx reported revenue of $620 million, which represented a 4% sequential increase, and a 49% increase from the same quarter a year ago. The firm saw solid demand across all end markets, communications equipment in particular, and generated an operating margin of 36% for the quarter. Altera also had an impressive September quarter, as sales were $527 million, up 12% sequentially, and up 84% from the year-ago quarter. Similar to Xilinx, Altera saw healthy PLD demand from the telecom equipment sector, and saw growth across all customer segments. Altera's profitability was spectacular, posting a jaw-dropping 45% operating margin.

Despite their robust results, the two firms gave divergent outlooks for the December quarter. Although many chipmakers across a broad range of industries and segments expect flat or modest sales declines next quarter, Altera expects 3%-6% revenue growth, again driven by vigorous demand from communications customers. Xilinx's forecast is more in line with the rest of the semiconductor industry, as the firm expects sales to be down 0%-4%. We suspect that the divergence is, in part, because Xilinx has relatively greater exposure than Altera to the 3DTV market, which has seen lackluster demand in recent months. Also, communications-related PLD sales can be somewhat lumpy from quarter to quarter, based on customer inventory corrections and the timing of new projects.

Xilinx and Altera are presently trading around our fair value estimates for the firms. Nonetheless, there have been growing fears that excess chip inventories have built up along the electronics supply chain. This could lead to a slowdown in PLD purchases down the road. If such a scenario were to occur, buying opportunities could arise for Altera or Xilinx, because stock prices of semiconductor names tend to fluctuate with business cycles. Given their solid competitive positions within a sweet spot of the chip industry, we believe that investors should keep these two PLD chipmakers on their radar screens for the long-term.