Intel Corp –ML Upgrade
Pointed in the right direction STRONG BUY
Long Term
BUY Reason for Report: Company Update
Highlights:
• We are upgrading our intermediate-term
rating for Intel, from Buy to Strong Buy, and
establishing a price objective of $42.
• Our 2002 earnings estimate is being raised
from $0.70 to $0.73. Higher gross margin
estimates are responsible for the change – we
have left our revenue estimate unchanged.
• We believe the launch of the 0.13 micron
Northwood puts Intel in the best competitive
position that the company has been in since
1999. We think that the Street is
underestimating the performance that Intel
will get from Northwood this year.
• Our analysis indicates that the PC supply
chain exited the fourth quarter of 2001 with
inventory at five-year lows relative to revenue.
Low inventory should underwrite solid
volume growth in Q1 and Q2 even if demand
is only mediocre.
• Over the long term, questions about Intel’s
ability to sustain growth remain. Our long-term
Buy rating is reiterated.
_ Upgrading intermediate-term rating to strong
buy – our 12-month price objective is $42
We believe that Intel is going to see market share gains,
improving visibility and margin expansion during the next
several quarters, greater than what is reflected in consensus
expectations. Additionally, we believe that pricing
pressure in the enterprise computing business will
increasingly favor Intel, as more and more of the enterprise
computing business moves to IA-32 over the short term
and IA-64 over the longer term. With all of that in mind
we are upgrading our intermediate-term rating on the stock
from Buy (2) to Strong Buy (1). We are establishing an
intermediate-term price objective of $42, or 46x our
forecast 2003 earnings and 58x our 2002 estimate.
_ Room for upside on margin
Pricing pressure during 2002 will be tough. We are
factoring in an overall ASP decline of 10% for Intel,
following the 13% decline that the company endured in
2001. However, Intel appears poised to recapture some of
that pricing with lower product costs during 2002 as the
company transitions to 0.13 micron manufacturing.
Manufacturing process transitions for Intel have
historically had a significant effect on gross margin, and
the move from the 0.18 micron Willamette version of P4 to
the 0.13 micron Northwood version will shrink the size of
the die from 218 mm square to about 140.
The resulting increase in die count per wafer should more
than compensate for the $500 million increase in
depreciation during 2002. We think that Intel’s 51% target
for 2002 gross margin is too conservative – our own
estimate shows 53%, and we believe even that figure has
upside potential. Our 2002 EPS estimate, which we are
moving from $0.70 to $0.73, is being raised on the basis of
higher gross margin expectations.
Table 1: Historic and Forecast Gross Margin
2000 2001 2002E 2003E
Q1 62.6% 51.7% 50.4% 54.5%
Q2 60.4% 47.8% 51.7% 54.8%
Q3 63.9% 45.7% 53.6% 56.4%
Q4 62.9% 51.3% 56.0% 57.6%
Year 62.5% 49.2% 53.1% 55.9%
Source: Company Information, Merrill Lynch Estimates
_ Depleted supply chain means robust unit
growth in 2002
Unit growth for Intel should be robust during 2002 – our
model shows YoY unit growth of 26% YoY. That may
seem high given our PC analyst Steve Fortuna’s forecast of
10% PC unit growth, but there are two important reasons
that we believe the 26% is reasonable. First, Intel’s unit
shipments during 2001 were artificially depressed by
inventory reduction efforts during the first two quarters.
YoY comparisons for all x86 microprocessors were –12%
and –9% for Q101 and Q201 while PC unit growth was
6% and –4%. Such is the math of an inventory drawdown.
The graph below illustrates the point nicely – MPU unit
shipments accelerated before the PC end market picked up
in 1998, and rolled over before the PC market did in 2001.
It is entirely reasonable to believe that microprocessor unit
volumes will recover before PC volumes do, and that is the
assumption built into our forecasts.
Chart 1: Historic and Forecast MPU and PC Unit Shipments
_ PC business exited Q4 with very lean
inventories
The PC supply chain exited Q4 with inventory/sales and
inventory/COGS ratios at the lowest levels in five years,
according to our analysis. Low inventory levels will tend
to further accentuate the gap between microprocessor
shipments and PC unit shipments. We aggregate inventory
and sales data from 21 EMS, OEM, distributor and system
integrator companies to build the analysis in the figure
below, and we now have data from 11 of those companies
on hand. We feel that we have been very conservative in
estimating inventory for companies that have not yet
reported – although all of the companies that have already
reported showed QoQ declines in inventory, we have used
flat QoQ inventory in our estimates for companies that
have not yet reported. The PC supply chain was extremely
depleted at the end of the fourth quarter, which should
support component demand through the first quarter at a
minimum even if demand is weak.
_ The result is improving visibility and
improving volume
All of this translates into improving volume and improving
visibility for Intel. We were somewhat surprised to see
Intel’s stock sell off following its Q4 earnings
announcement and conservative Q1 outlook. With
bookings in hand of $6.7 billion for the first quarter, Intel’s
decision to set the low end of its revenue range at $6.4
billion can only be described as extremely conservative.
We expect low inventories to drive not only volume, but
also visibility, higher for Intel over the next two quarters.
_ Intel should be stabilizing market share this
year
We believe that Intel will be stabilizing market share after
losing ground to AMD during the last part of 2000 and the
first part of 2001. AMD has been gaining market share on
Intel steadily over the last five years, periodic reversals
notwithstanding, but really made headway against Intel
starting in late 1999, as the figure below illustrates. We
expect Intel to come at AMD with a combination of
higher-performing product and aggressive pricing that will
halt some of AMD’s gains over the next two quarters.
_ Don’t underestimate how much headroom
Intel has with P4
The market share argument is an important one, and rests
on our belief that Intel is going to ramp performance for
the 0.13 micron Northwood version of P4 much more
quickly than investors, or Intel’s competitors, appreciate.
Intel has not introduced an entirely new processor core
since 1995, when the P6 architecture first saw the light of
day as the Pentium Pro. Every Intel desktop processor
since the Pro, up to and including the 0.13 micron
Tualatin, has been based on the P6 architecture.
The P7 architecture, or Netburst as it’s been dubbed by
Intel, has been designed with the intention of scaling the
performance of the core over multiple process generations
– the Willamette version of the P4 was just the beginning.
In short, Intel appears to have a lot of performance
headroom in the P4, more than the company is admitting
to. We expect to see processors clocking at better than 3
gigahertz by late in the year, and believe that Intel will be
able to re-open a performance gap between itself and
AMD over the course of the year.
_ Longer term positives include growth in
enterprise computing, early move to 300mm
Over the longer term, we are maintaining our Buy (2)
rating on Intel, one notch below Strong Buy (1). On the
positive side, we think that the commoditization of
enterprise computing that our Merrill Lynch hardware
analyst Steve Milunovich has written about will favor
Intel’s IA32 architecture, which is driving closer and
closer to the data center and other mission-critical
applications. Forecasting the impact that IA-64 will have
is more difficult, but in general we expect IA-64 to move
from zero market share to some level of success in
mission-critical computing.
Intel has also used its enormous resources to reassert its
position as the industry’s most technologically advanced
manufacturer. Moving Intel’s enormous manufacturing
base to 300 mm wafers over the next few years is an
audacious move, which if successful should open a
meaningful manufacturing technology gap between Intel
and most of the industry.
_ Negatives include poor diversification efforts,
pricing pressure
On the negative side, we note that Intel has not done a
good job of diversifying away from its core
microprocessor business. Intel’s efforts to acquire its way
into the communications IC market have largely failed,
with the one notable exception of the IXP series of
network processors. In wireless, although the embedded
X-Scale embedded processor has done well, Intel has had
no competitive impact with its Blackfin DSP. The
majority of Intel’s revenue in wireless is still flash
memory. Other ventures in web hosting and
communications hardware have fizzled as well. Although
the move into enterprise computing represents
diversification of a sort, Intel’s fate is still tied to the
computing market for the foreseeable future.
That represents a challenge when 10% ASP declines are
the norm rather than the exception. Intel’s microprocessor
ASPs were flat or even up for most of the 1990s, but have
declined by 10% or more in two out of the last three years.
Intel may be able to grow revenue sharply off of a bottom
– the next few quarters represent the best juxtaposition of
low inventory and product mix that the company has seen
since 1999. However, it is tough to see how Intel can
grow revenue by more than 10% per year over the long
term.
_ Intermediate-term industry scenario is solid,
but we need a demand recovery at some point
There are also questions to be asked about the
sustainability of the semiconductor rally in general. When
we upgraded our stance on the semiconductor sector last
year from neutral to positive, we pointed towards declining
inventory and bottoming YoY comps as reasons for buying
the sector despite the fact that valuation was expensive.
That has worked for the last quarter and a half, but it will
not work forever.
The figures below illustrate the point – the semiconductor
business has never seen sustained growth in a weak
economic environment. Stocks may snap back as YoY
comparisons stabilize, as we’re showing in Figure 5, but
truly re-accelerating growth beyond the 10% YoY level
requires an economic recovery. With valuations for our
universe of coverage at an average of nearly 62x 2003E
earnings, any leveling of growth will be bad for stock
prices.
_ For now, though, Intel should outperform in all
but the worst possible environment
Whether or not the stock continues to be a Strong Buy
beyond the intermediate-term is dependent on the strength
of any economic upturn, which is hard to forecast. For the
nearer term, however, all the product and margin
indicators are pointed in the right direction for Intel, and
we think that the stock will outperform in anything but a
disastrously weak economy.
_ Valuation – high but reasonable relative to the
sector
Our expectations for Intel’s stock price are based on our
valuation work as well as the demonstrated tendency of the
stock to perform in line with gross margin. The company
is currently trading on 46x our 2002 earnings estimate and
37x our 2003 estimate – revenue multiples are 8x and 7.3x
respectively. That is far from cheap for Intel, which
bottomed at 19x prospective earnings in 1998 and 5x
revenue. It is, however, low relative to the sector in P/E
terms and in line with the sector in revenue terms as the
table on page 6 illustrates.
We also note that companies that have weathered a
downturn in better shape, without going into loss, perform
better early in a cyclical upturn than loss-making
companies even if valuation on peak earnings appears less
reasonable. That has been demonstrably true for analog
semiconductor makers such as Maxim Integrated Products,
which has outperformed the Philadelphia Semiconductor
Index despite high valuation.
_ Correlation between gross margin and stock
price performance
Stock price performance for Intel relative to the S&P500
has correlated well with the direction of Intel’s gross
margin, as the figure below illustrates. That makes sense,
given the fact that Intel’s gross margin is affected by
product mix and revenue. Especially with Intel’s
estimates, and consensus figures, not sufficiently
aggressive, we think that improving gross margin over the
course of the year will continue to be good for the stock. |
