By Thomas W. Smith, CFA A highlight of the July earnings season for the semiconductor industry was positive news about Intel's (INTC) gross margins. The world's largest chipmaker delivered a second-quarter gross margin of 50.9%, which was above our estimate for 50% and the midpoint of its guided range of 50%. This followed Intel's first-quarter gross margin of 52%, which was at the high end of its guidance.
Altogether, we at S&P think Intel did well in hitting its gross margin targets during the first half of 2003, which was a seasonally slow period pockmarked by a price increase in flash memory that cost market share, the Iraq conflict, the SARS epidemic, and mediocre PC sales. Of course, Intel made some of its own luck by introducing in March its Centrino technology that enables wireless connectivity for laptops, which helped steer PC customers toward notebooks with these high-margin chips.
The news about gross margins for the rest of 2003 was the bigger upside surprise, in our view. Guidance for gross margins for both the third quarter and for all 2003 was 54% -- above the 50% and 51% we had modeled, respectively. To make 54% for the year, Intel's fourth-quarter gross margin needs to be 57% or better, and we have modeled 58%.
NO PC SURGE YET. While the jump may appear big, unit-volume gains from cycle-low levels can often dramatically improve wafer fab efficiency. Gross margins for chipmakers can also be affected by product mix, average selling prices, inventory valuations, and the timing of startup costs for new product lines.
Interestingly, Intel reports no surge in PC sales immediately ahead, although management has faith that sooner or later customers will need to replace aging computers bought in 1999 for Y2K compliance. Confirming the no-surge-yet thinking, the company left its plans for capital spending unchanged at $3.5 billion to $3.9 billion. Of course, whenever PC sales do finally accelerate, Intel should benefit, as will the rest of the chip industry.
Intel's ability to improve gross margins at today's modest revenue levels boosts our confidence in its ability to earn substantial profits as the industry cycles up. We project an expansion that should endure into 2006, based on rising U.S. gross domestic product, the need to replace aging tech gear, and the desire to upgrade equipment to gain the advantages of wireless connectivity.
HAZY FUTURE. After Intel's second-quarter earnings report, we raised our earnings-per-share estimates to 67 cents for 2003 and to 85 cents for 2004, and upgraded the shares to 5 STARS (buy). With Intel trading near $24 on Aug. 7, the price-earnings ratio on our 2004 estimate is 28, which is at premium to the S&P 500-stock index but in line with peer companies.
Looking into the hazy future, we estimate EPS for 2005 of $1.05. Our 12-month stock-price target of $32 is based partly on applying a p-e of 30 (near the current forward p-e of 28) to our $1.05 estimate for 2005, and partly on discounted cash-flow analysis, which indicates an intrinsic value near $32. Our DCF model's assumptions include annualized growth of cash flows over 15 years at 13.5%, then annual growth to perpetuity at 4%, and a weighted-average cost of capital of 11%.
The potential price appreciation from $24 to our target is about 33%. Risks to achieving that level include the pace of the economy, fluctuations in chip pricing, competition, execution of product launches, wafer fab efficiencies, and valuation levels for semiconductor stocks.
HIGHER FORECASTS. Our other 5-STAR buys among semiconductor stocks are Texas Instruments (TXN), Analog Devices (ADI), and Vishay Intertechnology (VSH). Analog Devices reports results Aug. 14 for its fiscal third quarter. The ADI conference call should be interesting because it's for a July quarter that doesn't include results for the month of April, which we think suffered some of the worst effects of the SARS epidemic on electronics orders.
Supporting our notion that the semiconductor industry expansion is slowly gathering speed is the Aug. 2 release from the Semiconductor Industry Assn. announcing that second-quarter worldwide chip sales rose 10.4% year-to-year and 3.2% from the first quarter. Sales for the first half of 2003 were up 11.8% year-to-year, which is in line with our forecast for 12% industry growth for all 2003 and above the SIA's 10.1% growth forecast.
Intriguingly, the SIA raised its forecast for third-quarter sales growth to a 5% to 8% increase over the second quarter. The midpoint of that range, 6.5%, is higher than the SIA's prior forecast of 5.8%. Despite the increase for the third quarter, the association didn't raise its forecast for the whole year, reasoning that the outlook for the fourth quarter was too hazy to merit a change from its 3.7% forecast.
We believe that market observers who had been expecting global chip sales for 2003 to come in at 10% or less might need to increase their estimates as the year evolves.
SIGNS OF AN UPTURN. This is an interesting time in the semiconductor cycle. It has been about two years since the bottom for industry revenues in the second half of 2001, yet perhaps still two or more years away from the next top. Many companies have finished their cutbacks and are at low levels of profitability, but revenue levels are modest and rising only gradually. Even if one believes that the industry will strengthen toward a cycle top for sales in 2006, little sign of a sales surge is apparent in the third or fourth quarters this year, and the first quarter of 2004 is apt to face seasonal weakness.
In a classic semiconductor cycle, a point is generally reached when chip demand begins to outpace supply, and the whole industry seems to shift into high gear. Order lead times start to lengthen, average selling prices start to rise, capacity utilization starts moving above 90%, capital spending starts to increase, and inventories suddenly seem too low, so orders pick up not just to meet current demand but also to replenish inventories at higher levels.
We believe that some of the conditions for a surge are already in place, but not all. Many of the companies we talk to often describe channel inventories as lean. Investment in new capacity has been low for many quarters, although we sense that a lot of mothballed equipment exists that could be turned on before new plants would be built. Capacity utilization generally is at the mid-80% level, not quite high enough to generate big excitement, but not too far from the tipping point.
OVERREACTING INVESTORS. Average-selling prices remain modest in most chip categories, with sporadic upticks. Again, interesting, but not at the tipping point. Order lead times remain quite short. When they begin to stretch out, it'll be a prime signal that industry activity is picking up. We're not there quite yet.
When the turning point to higher demand is reached, investors are apt to react in a buying frenzy, as fundamentals suddenly look a lot better. While we expect something like this scenario to play out in this cycle, we think the surge is likely to be in mid-2004 or later, given the large amount of productive capacity in mothballs.
Still, it might occur within the 12-month investment horizon that we use for our STARS recommendation system. Accordingly, we maintain a positive outlook on the semiconductor industry for the next 12 months. A caveat might be that many investors are watching chip cycles closely and tend to push the shares up on false starts during the gradual portion of the recovery. This lends volatility and risk to investing in the group. Analyst Smith follows semiconductor stocks for Standard & Poor's