Over the last six months, Altera (ALTR) has been a bit of a bull run, appreciating by 24.9%. As impressive as this was, one must not forget that the NASDAQ gained 33.3% over the same time period. In this article, I will run you through my DCF model on Altera and then triangulate the result with an exit multiple calculation and a review of the fundamentals compared to Xilinx (XLNX) and Texas Instruments (TXN). I find that the firm is meaningfully undervalued, while Xilinx is overvalued.
First, let's begin with an assumption about the top-line. Altera finished FY2011 with $2.1B in revenue, which represented a 5.6% gain off of the preceding year. I model revenues trending from 10.5% to 8% over the next 6 years.
Moving onto the cost-side of the equation, there are several items to consider: operating expenses, capital expenditures and taxes. I model cost of goods sold as 30% of revenue versus 13.4%, 14.5%, and 1% for SG&A, R&D, and capex, respectively. Taxes are estimated at 18% of adjusted EBIT (ie. excluding non-cash depreciation charges to keep this a pure operating model).
We then need to subtract out net increases in working capital. I expect this to hover around 1% of revenue over the projected time period.
Taking a perpetual growth rate of 2.5% and discounting backwards by a WACC of 10% yields a fair value figure of $49.12, implying 27.4% upside. My sensitivity analysis (ie. considering a variety of different WACC and perpetual growth rate inputs) yields a standard deviation of $5.90. This represents only 15.3% of the stock value - thus the DCF model is reliable in the sense that it is not very sensitive.
All of this falls within the context of strong performance:
Despite the headwinds of the second half, 2011 was a good year following a phenomenal 2010. We saw revenue increase nearly 6%, with another year of share gains versus our primary competitor.
Gross margin at 70.4% was stronger than our 70% guidance. As the year progressed, we successfully stepped up our R&D spend to a level more in line with our long-term business model as we began the introduction of our 28-nm FPGAs and initiated a series of strategic initiatives with a longer-term payoff. We were disciplined though. OpEx at $605 million was $5 million under the guidance for the year. Our profitability remains the best in the PLD industry.
From a multiples perspective, Altera does not appear attractive at first glance. It trades at a respective 16.4x and 17.4x past and forward earnings. On the higher-end, Xilinx trades at a respective 17.3x and 18.4x past and forward earnings. Texas Instruments trades at 17.4x past earnings but only 13.7x forward earnings.
Consensus estimates for Texas Instruments' EPS forecast that it will decline by 20.4% to $1.76 in 2012 and then grow by 34.7% and 20.3% in the following two years. Assuming a multiple of 16x and a conservative 2013 EPS of $2.33, the rough intrinsic value of the stock is $37.28, implying 14.4% upside. 37 of the 41 revisions to EPS estimates have fallen for a net change of 5.8%. Many investors are fearful about poor end market demand among semiconductors, but I believe this only sets the foundation for higher risk-adjusted returns.
Consensus estimates for Xilinx's EPS forecast that it will decline by 20.1% to $1.91 in 2012 and then grow by 3.1% and 20.3% in the following two years. Assuming a multiple of 16x and a conservative 2013 EPS of $1.95, the rough intrinsic value of the stock is $31.20, implying meaningful downside. If end market demand in technology is to remain weak, Xilinx is thus the most vulnerable of the three to fall.
Disclosure: I have no positions in any stocks mentioned, and no plans to initiate any positions within the next 72 hours.
Disclaimer: The distributor of this research report is not a licensed investment adviser or broker-dealer. Investors are cautioned to perform their own due diligence. We seek business relationships with all of the firms in our coverage, but research covered in this note is independent.