Semiconductor Equipment . . . Merrill Lynch upgraded Veeco Instruments to "buy" from "neutral." Brett Hodess said he does not believe the current valuation reflects the company's improving fundamentals, which include increased orders and decreased costs. In addition, the cancellation of the FEI merger earlier in the month removes a potential overhang.
Semiconductors . . . Salomon Smith Barney said in a pre-open note that they continue to be concerned that the steep, "unexpected rush" of HDD demand from early Oct through mid-December may have triggered double-ordering in the HDD component supply chain, and believes that the risk of a HDD IC inventory correction for Marvell Tech has risen (MRVL price target is $13.60). Firm also believes that LSI Logic could be a value trap (target $4), and anticipates that Broadcom could move much lower if anecdotal evidence surfaces about other defections or collapsing employee morale (target $10).
Micron's management held a meeting with a discussion of what has gone wrong over the last twelve months. CEO Steve Appleton's comments boiled down to three key issues. First, it's clear that the aborted merger talks with Hynix took Micron's eye off the ball in terms of Micron's own capacity and spending plans. Second were the mid 2002 line-thinning efforts, which led to delays and lower yields when Micron reloaded its factories. Third, a combination of bad choices and bad luck caused Micron to miss volume sales of 2002's hottest product - 256megabit DDR DRAM. That set the stage for a discussion of where Micron is headed now. Cash burn is unsustainable, although Micron has plenty of options First on our minds was Micron's liquidity situation. In our December comment on Micron's 1st quarter 20Y03 results, Micron's cash and liquid investments had fallen from $990 million to $660 million. However, it's also important to note that Micron drew down trade receivables and increased trade payables during the quarter, generating a $200 million benefit. Although there is nothing wrong with running a balance sheet tightly, the organic (leaving balance sheet management aside) rate of cash burn during the November 2002 quarter was in excess of $500 million. Micron CFO Bill Stover indicated that cash balances have declined by another $100 million since the beginning of the quarter, and also said that the company has added $70 million in equipment financing since the beginning of the quarter. That comes to $170 million in cash burn 2/3 of the way through the February quarter, and expect the February 2003 results to show organic cash burn of between $250 million and $300 million. Bill was unwilling to directly affirm or deny our numbers but presented us with enough data to give us conviction in our analysis. That rate of cash burn is unsustainable, and would likely cause Micron to run out of cash in another 3 to 4 quarters. Given the company's relatively low indebtedness (debt to equity was only 7% as of the November 2002 quarter) the ability to raise additional cash is not in doubt. Indeed, Micron indicated that a financing would likely be in the $500 million to $750 million range, although no details regarding the type of transaction or timing were discussed.
Two weeks of reporting has provided some modest upside surprises to Q4 earnings, particularly Intel and Texas Instruments. Full year
2003 EPS estimates, however, have decreased by about 4% for semi universe coverage in the last month. Despite the mostly modest 4th quarter upsides, some of which resulted in revisions upward to earnings forecasts (Texas Instruments, though not Intel), there has been little follow-through with the stocks. In fact, two weeks ago the SOX was trading 20% higher than it closed on Friday (the SOX close of 283.67 broke technical support at 285). There are a number of reasons for the weak relative performance, despite the somewhat better commentary out of our companies.
Throughout all the volatility of the last couple of months most analysts have retained a Marketweight rating on the semiconductor sector largely because several factors will hold the group from outperforming the broader averages in the next six months, and could under some scenarios present downside risk. These concerns revolve around: likely deceleration in unit growth in the first half, a possible modest buildup in personal computer and wireless handset inventories, lack of pricing traction despite some recovery in units, and valuations that are still not cheap by historical levels.
Deceleration in orders, dollars, and units. Analysts have been talking some months about the unit deceleration that is going on in the semiconductor sector, and how historically that has been a negative for the group. Only once (1991- 1992) did units decelerate (from much more modest highs than were recently set), and most of the semiconductor group performed poorly. The unit peak in early 2000, of course, ushered in the worst of all semiconductor downturns.
Orders, which are indicators of forward-looking conditions, can have an even closer correlation with stock price performance. Before the WSTS stopped reporting semiconductor sector orders (different from semiconductor equipment orders, as reported by SEMI), the change in orders and the change in stock prices had an R of greater than 0.90, which is pretty high in the natural world. Looking at Texas Instruments orders which represent a relatively diversified portfolio, semiconductor orders grew 40% in 3rd quarter, 35% in 4th quarter, will slow to 15% in 1st quarter and will be about flat in Q2 (assuming a book-to-bill ratio of about 1.0 for the next two quarters).
Pricing remains weak. Unit growth has largely recovered in the semiconductor group (WSTS data suggests units peaked at 27% yoy growth in September; Intel reported record unit shipments in December), but prices have not (WSTS pricing was down 3% in November: Intel's prices are off about 20% from the last unit peak in 2000). The most recent evidence of price weakness has been in the DRAM arena, where bellwether DDR 256Mb prices have fallen up to 14% in the past week. As Micron Technology reported last week, unit demand year to date has been weaker than expected. Several months ago analysts started writing about the secular price pressure that dependence on low-income emerging markets will bring about.
Has there been a stealth inventory buildup? One of the most controversial questions with regard to semiconductors relate to customer inventory levels. Virtually all of our companies, including Intel (PCs) and Texas Instruments (cell phones) report their customer inventories are low. For several reasons, a simple comparison with 2000, and even with the spring of 2001, have little value. Distributors and contract manufacturers today simply do not have the strength of balance sheet to over-stock inventory. In addition, a year ago there was the new P4 build, which created excesses. Even so, the primary strength in both the PC market and cellular handset market has been in "white boxes", unbranded systems typically sold in Asia-Pacific (primarily China). The Asia-Pacific market is notorious for speculative inventory building, and the recent weakness of DRAM suggests to us the sector could be in for a modest inventory surprise in February and March. Certainly, recent weakness in DRAM pricing suggests someone is unloading excess product into the market.
Valuations not all that cheap. The longer the semiconductor sector muddles around in a slow-to-no-growth condition, the more likely valuations will grind lower. Analysts do not necessarily believe the semiconductor group must hit previous recession-level lows. Analysts do believe, however, that there re mains some downside risk on a valuation basis, and should the sector see significant weakness this spring, there could be further downside valuation pressure. The horror story, as we all know, is worse with the group still showing 58% downside valuation risk on a price/sales and 41% on a price/book basis. Not that the sector is wildly over-valued, however, on average there is 30% upside potential to the median for our stocks under coverage with regard to price/sales and 114% upside to price/book value median. Median valuations are impacted significantly by several stocks that are probably in secular declines and may take some time to recover to their median values, including Amkor Technology, Cirrus Logic, and Pixelworks. Excluding these three companies, average appreciation to median price/sales is 10% and average appreciation to median price/book value is 87%. Meanwhile, the largest capitalization stocks, like Intel and Texas Instruments, show 30% and 75% downside risk to 10-year, pre-bubble price/sales valuations, respectively.
2020insight.com
Gottfried, the NDX remains over its neckline in a show of relative strength. I expect the market must either rally and soon or we can expect the the BPNDX to begin a rapid decline. We will have to wait a while to begin to figure out bull market data because that will happen in the future.
RtS |
