We are calling Intel (INTC) out for significantly sandbagging its 2013 estimates. I believe that the majority of sell-side analysts are grossly underestimating Intel's full year 2013, and further believe that Intel is being too cautious in its forecast, likely in an attempt to subdue expectations to avoid any disappointment during the year. In this article, we aim to justify a ~$56B revenue estimate for CY 2013, significantly above consensus of $54.02B, which should put the sales right at the high end of the sell-side analyst estimates for the year. The executive summary of the thesis:
- PC chip unit volume returns to CY2011 levels (this is conservative), driven by the "Haswell," principally driving accelerated Ultrabook adoption, and modestly moving the needle in the traditional notebook/desktop PC space. This is quite achievable as it requires only a 1% increase in volumes Y/Y from CY2012
- PC chip ASPs springing back to CY2011 levels (~+6.3% Y/Y from 2012 levels), which should be achievable as "Haswell" brings both GT3 and GT2 graphics, which should allow for Intel to charge a meaningful premium at the higher end for tangible graphics performance increases. Further, "Ultrabook" (ULV) SKUs typically carry a meaningful premium over their non-ULV counterparts due to binning/integration of PCH on-package
- DCG growth seems to be modeled conservatively at +10% exiting CY2013 by the sell-side. We believe this is too conservative, and make the case for 15%.
- While we believe that tablet sales do not become meaningful until the end of 2013, the introduction of "Bay Trail" should drive meaningful share gains in both the Windows 8 and the Android tablet spaces, as we believe that this solution should have a very strong competitive positioning in the market
Below, we more deeply detail the thesis along with attendant upside/downside risks.
PC Unit Volume Growth
A core component to the thesis is that PC unit volumes will spring back to CY2011 levels. We believe that while previous iterations of the Ultrabook specification were uninspiring and offered a poor value proposition over Apple's (AAPL) MacBook Air in the premium "thin and light" category, and further believe that the battery life, form factor differentiation, and most importantly cost were not there to drive significant adoption.
We believe that growth from CY2012 levels occurs with the "Haswell" generation. Products in this generation will require both touch screens (which will drive Windows 8 acceptance) in addition to nine hours of battery life, which is a substantial improvement over the four-six hours that today's "Ivy Bridge" Ultrabooks typically achieve. Further, we believe that "convertible" models at reasonable price points and with good battery life metrics should not only stave off cannibalization by traditional pure-play tablets, but the possibility for "reverse cannibalization" exists, as we believe that there is a significant value proposition with convertible devices.
PC ASPs Should Rebound
In CY2012, ASPs ticked down ~6% in the PC client space, which accounted for the majority of the revenue shortfall during the year. We believe that "Haswell" generation processors will undergo the following puts and takes with respect to ASPs:
- (+) With "Haswell," Intel will be more aggressively differentiating between low end and high end integrated graphics, and will charge a higher price over corresponding "Ivy Bridge" for the models with GT3 (highest end SKU) and keep pricing flat over "Ivy Bridge" for the models utilizing the GT2 SKU
- (+) We believe that "convertible" Ultrabooks will gain significant traction. The processors that sell into these segments command a slight premium over the "standard" 15W ULV parts (+$20/unit), which should have a positive effect on ASPs
- (-) We believe that there will be significant competitive pressure from Advanced Micro Devices (AMD) in the low end of the market. The company is currently very aggressive with pricing and operates its business at ~39% gross margins. While Intel has structural advantages that should allow the company to maintain greater margins at a given price point (in-house fabrication, transistor geometry advantage -- 22nm v.s. 28nm/32nm), competition here that is absent at the high end of the spectrum will serve to detract from blended ASPs
We believe that these factors should lead to a material rebound in PC ASPs, especially as mix shifts more toward convertible/ultrabook models.
Content Sold Into Client Space Will Begin To Materially Increase
Intel has traditionally sold CPUs and chipsets into the PC client space, and often sells integrated LAN/WiFi onto certain higher end client SKUs. We believe that Intel will begin to sell more content into the PC space, and on a longer-term basis, integrate nearly every major component onto a system-on-chip, but in the near term, we believe that the company will start to become more aggressive with bundling NAND flash solutions with its processors and chipsets to the OEMs.
In particular, while Intel has traditionally sold 2.5" consumer solid state drives for standard notebooks and desktops, it has been conspicuously absent from the mSATA SSD game. These thin form factor SSDs are used in Ultrabooks, All-In-One desktops, and embedded designs, which are the higher growth segments of the markets. While we do not believe that traditional hard disk drives will suffer materially as a result of the growth in NAND flash adoption, we do believe that hybrid solutions (Flash as a cache/primary drive, HDD as bulk storage) will dominate the Ultrabook and up landscape over then next several years.
Intel just announced its "Intel SSD 525" model of mSATA SSDs for consumer use, and it is clear that the company's intent is to mirror the success of its DCG business, which according to the company sells $0.50 of non-CPU content for every $1.00 of CPU content sold. We believe that this will help Intel get full leverage of its joint venture with Micron (MU) to produce NAND, and we further believe that Intel's track record of reliable flash drives will lead to nontrivial adoption of its products into Ultrabook products.
By selling $100 - $200 worth of SSD/NAND content into each Ultrabook, Intel stands a strong chance of materially increasing its revenue per Ultrabook.
Data Center Group Should See Better-Than-Expected Growth
Intel estimated "double digit" growth in 2013 for its data center group ("DCG"). Sell side analysts peg the growth at about 10%, which would imply an increase in $1.07B for the segment Y/Y. We believe that this estimate is conservative, and further believe that the company is keeping expectations modest here despite the unprecedented, broad refresh of its product lineup. We go over the key drivers below:
- "Itanium" refresh is heavily discounted. While the "Itanium" has gotten somewhat a bad reputation, the product is still a ~$4B/yr business for the company, likely with strong operating margin given the premium ASPs that the chips command. The firm recently refreshed its Itanium lineup with a complete redesign of the micro-processor core. This represents the first refresh of the product lineup since 2010, which should serve to satisfy pent-up demand for the product. Sales should ramp during 2013
- "Ivy Bridge" refresh is key driver. The "Romley" platform will see a refresh in the 2S/4S space with parts based on the 22nm "Ivy Bridge" architecture. The chips should feature more cores, more cache, and improved RAS features. More interestingly, the big-iron enterprise parts will see a refresh from the 32nm "Westmere" architecture to the 22nm "Ivy Bridge," representing a two generation leap for these customers. We believe that enterprise sales will see both macro-economic improvement as well as a significant product refresh as key drivers for the resumption of growth in this space.
- Xeon Phi is pure upside as a new market segment. The high performance computing ("HPC" industry is among the fastest growing in the data-center. There is no such thing as "good enough" performance -- more is always better. Nvidia (NVDA) pioneered the use of programmable GPUs as highly powerful and power efficient floating point co-processors and has seen significant success in this area. Intel is now entering the fray with a competitive solution. Any share gains are pure upside, and we believe that the mere presence of a large player like Intel will validate the field and lead to much wider spread adoption of such solutions. Margins here should be quite good as there are only two (competitive) players.
- 22nm "Atom" represents pure upside. Intel will be pushing heavily into new areas of the data-center with its micro-server "Avoton" Atom parts built on the new micro-architecture codenamed "Silvermont" and the 22nm LP process. There will also be a variant of this system-on-chip codenamed "Rangeley" for the networking and communications market, which will integrate an on-die cryptography co-processor. The "Atom" family will likely prove to be very important in the more embedded areas of the data-center going forward, as the SoC methodology that the Atom team uses lends itself to multiple, custom-tailored SKUs for the respective market segments.
We believe that this is an unprecedented refresh and extension of the firm's product lineup, and as a result believe that 10% should be used as a pessimistic lower bound on growth in DCG, with 15% Y/Y growth a more realistic target. We also believe that momentum will continue into 2014 as further SoCs built on the new "Atom" core, in addition to a full platform refresh of the Xeon platform to the "Haswell" microarchitecture roll out.
We note that this is Intel's highest operating margin segment (~50%), and as such believe that any upside here will have the most material effect on the net income and, ultimately, EPS estimates. if we are correct, then 15% Y/Y revenue growth implies ~$500M worth of top line upside and, therefore, $250M of operating income upside.
Tablets Unlikely To Move The Needle In 2013, But Will Improve Sentiment
We do not believe that the firm's current "Clover Trail" products are sufficient to drive significant sales growth. While the CPU portion of the product is very strong and competitive with the best ARM (ARMH) solutions shipping today, we do not believe that the graphics processor (licensed from Imagination Technologies) that comes integrated onto the system-on-chip is adequate. It offers a mere 1/15th of the performance of the graphics processor found in Apple's iPad 4. We further believe that this graphics processor was used for time-to-market reasons, as the Windows software stack for that particular GPU (recycled from the latest desktop/netbook iteration of Atom) was already complete.
Going forward (with Bay Trail and beyond), Intel will be using a scaled down version of its HD graphics found in its PC client microprocessors. The GPU in "Bay Trail" should help significantly bolster Intel's competitive position in the mobile SoC GPU space, eliminating the final bottleneck that keeps the chip from powering a high-end tablet product. We believe that Intel's pulling-in of the 22nm Atom SoC codenamed "Bay Trail" into 2H 2013, rather than 1H 2014 as initial leaks suggested, will help to drive sentiment, but we are unsure how much market share Intel can capture.
A positive is that Bay Trail will be supported on both Android and Windows 8, which initial leaks suggested to the contrary. This should significantly widen the SAM for these products and act as a nice hedge against a potential failure of Windows 8 in the mainstream tablet space. We believe that going into 2014, the notion that Intel is not a major player in the tablet space will begin to evaporate.
Coming To An Estimate
We believe that PC client group revenue will rebound to CY2011 levels, which gives us a baseline sales estimate of $54B (+$700M Y/Y). We further believe that while 10% growth in DCG is too conservative and instead model 15% at the midpoint with potential to exceed this target with upside to the 20% level, implying +$1.6B to the top line. We next believe that software and services, which quietly grew 10% Y/Y from $578M to $636M could see continued momentum as next-generation Ultrabooks require some form of anti-virus/security, which could be a positive for McAfee, which could account for +$60M at the top line. Finally, we note that "All Other" (including NAND/SSDs) saw 9% Y/Y growth. This is likely to repeat itself, which adds another $50M to the top line.
We conservatively exclude tablets from this discussion, but it could trigger an upside surprise if "Bay Trail" is shipping in volume by the "Back to School" season.
Summing it all up, we believe that sales should come in at around $55.71B, or ~4.5% year-over-year growth, and significantly ahead of sell-side estimates. This is actually in line with the firm's guidance of "low single digit" growth, but at the high end of it. We believe there is room for further upside from our estimates, but refrain from attempting to bake it in until we have seen the Q1 results and Q2 guide.
Additional disclosure: I own ARMH puts.