COMPANY PROFILE
Intel Corporation (INTC, Financial) is the world’s largest chipmaker by revenue ($43B in 2010) and unit shipments, eclipsing their peers in the semiconductor industry. They design, develop, and manufacture integrated digital technology, primarily integrated circuits (ICs, or aka chips) for original equipment manufacturers, original design manufacturers, personal computer (PC) and network communications products users, and other manufacturers of industrial and communications equipment. Additionally, they develop computing platforms—a combination of hardware and software designed to provide a solution.
Although known for their dominant market share in microprocessors, Intel is transforming from this focus to become a company that can deliver complete solutions consisting of hardware/software platforms and supporting services. As stated in their recent 10-K, their goal is to be “the preeminent computing solutions company that powers the worldwide digital economy.”
Intel was originally founded and incorporated in California in 1968 and later reincorporated in 1989 in Delaware. They are based in Santa Clara, California. Their company stock trades on the NASDAQ under the symbol INTC with a market capitalization of $111B. As of the end of 2010, Intel had 82,500 employees worldwide, of which 55% are employed in the US.
COMPANY HISTORY
Intel was founded in 1968 by three men: chemist and physicist, Gordon Moore (of Moore’s Law fame); physicist and co-inventor of the IC, Robert Noyce; and chemical engineer, Andy Grove. Their company’s original name was NM Electronics, before changing it a year later to Integrated Electronics, or “Intel” for short.
At its start, Intel set itself apart by their ability to produce semiconductors. Their primary products were static random access memory (SRAM) chips. During the 1970s their business grew and they expanded and improved their production processes to allow a wider range of products to be made.
Intel created the first commercially available 4-bit microprocessor, the Intel 4004, in 1971. In 1972, they introduced the first 8-bit processor, the 8008. In 1974, the 8080 came out which was wildly popular in the growing personal computer movement and spawned a variant from Zilog, the Z-80, popularized in Radio Shack’s offering, the TRS-80. Motorola got into the microprocessor game as well with its 6500-series processors being used extensively by Apple in its early models of the Apple, Apple II, and the more powerful 68000-series in the line of Macs.
Intel’s big break came with the introduction of the x86 architecture. The 16-bit 8086, and the 8-bit 8088, became the brains of the IBM PC in the early 1980s. Until then, Intel’s business was dominated by dynamic random access memory (DRAM) production. However, they faced increased competition by Japanese manufacturers reducing profitability of this commodity. Fortunately for Intel, the success of the IBM PC convinced then-CEO Andy Grove to shift their focus and concentrate on microprocessors.
Intel followed this success up with the 80186 and 80286 (1982), the 80386 (1985), and the 80486 (1989). By the end of the decade, the rapidly growing PC market buoyed Intel’s position and they became the primary and most profitable hardware supplier to the PC industry. In 1991, they introduced the “Intel Inside” campaign which linked brand loyalty to consumer selection. By the end of the 1990s, Intel’s Pentium line of processors was a household name.
In the 2000s, microprocessor demand slowed and competitors, mainly AMD, gained significant market share. Intel unsuccessfully attempted to diversify their business beyond chips. During 2005, CEO Paul Otellini, reorganized Intel to refocus the core processor and chipset business on platforms—enterprise, digital home, digital health, and mobility. To regain momentum from lost market share, Intel introduced the “tick-tock” product development model so they could regain their technological lead. This program alternated microarchitecture innovation with process innovation in yearly cycles to push a processor out approximately every two years.
Recently, Intel acquired McAfee, a manufacturer of computer security technology, and Infineon Technologies’ Wireless Solutions business. With the Infineon deal, Intel plans to use this technology in laptops, smart phones, netbooks, tablets, and embedded computers in consumer products.
INDUSTRY INFORMATION
In order to understand Intel’s business model, it’s necessary to spend a few moments understanding the basic nature of the technology.
Transistors, a form of semiconductor, are integral to all modern active electronics and ubiquitous in our everyday lives. The first patent for a transistor was in 1925; however, it wasn’t until Bell Labs’ work in 1947 where the potential of the transistor became apparent. Texas Instruments produced the first silicon transistor in 1954 and ushered in the age of miniaturization. Up until this point, active electronics used vacuum tubes and computers were massive along with their equally huge magnetic-core memory.
Transistors can be packaged as separate, discrete components, or integrated onto integrated circuits (ICs), aka chips, to produce complete electronic circuits. This process is accomplished with semiconductor device fabrication using wafers of silicon in a foundry, or fabrication plant, aka “fab.” An astonishingly low price per transistor can be achieved as transistor densities increase, and microprocessors use a lot of transistors. More transistors generally translates to more powerful processors.
Moore’s Law, made famous by co-founder of Intel, Gordon Moore, predicts that transistor density on ICs would double every two years. With each new generation of microprocessor, the envelope has been pushed further and further outward, generally keeping pace with Moore’s Law. However, physics imposes a lower limit on the size of silicon-chip transistors—below a certain size, they’ll produce excessive heat and have their functioning impaired by quantum effects. Experts predict the transistor density wall will come in about 10-20 years. When this happens, we will have reached the limits of silicon-based processor technology—processor speed and performance will plateau at that point, unless a replacement is developed.
Carbon nanotube technology has the promise to replace silicon chips and allow computer speeds to continue increasing. Nanotube-based transistors, known as carbon nanotube field-effect transistors (CNFETs), have been made that work at room temperature and are capable of digital switching. However, a major obstacle exists—the technology to mass produce them. Additionally, incorporating nanotubes into logic-gate circuits with the required densities to match transistor performance of the 1970s, let alone modern performance standards, hasn’t been demonstrated yet. The future is unclear for the next dominant microprocessor technology.
Additionally, there’s a growing trend to keep in mind—the “cloud.” The cloud is nothing more than the Internet; however, how we use and access it in the future will be changing. With an ever-increasingly connected world with computers and mobile devices, more people and more devices will be accessing the cloud. There’s a growing trend where services and file storage will be offered by the cloud and the consumer will access it, where ever he/she is at, with a cloud-connected device, ie. mobile phone, PC, tablet, etc. In effect, this trend means a transfer of raw computing power to the servers making up the backbone infrastructure of the Internet, or cloud, from the currently powerful desktop computers, as the servers would be the devices managing the requests, executing applications, and crunching data. There’s also a trend for faster and faster data transfer, driven in part by video streaming across the cloud. Not only will end devices need to be fast, but the interconnecting devices and infrastructure will need a build-out to accommodate ever-increasing speeds.
BUSINESS MODEL
Intel designs and manufactures computing and communications components, which include microprocessors, motherboards, chipsets, and wired and wireless connectivity devices. Microprocessors account for 15% of the total semiconductor industry’s revenues, and this field is dominated by Intel and AMD. Intel ships over 80% of the world’s microprocessors.
The company’s vision is to create a seamless continuum of personal computing experiences based on their proprietary architecture. Achieving this vision, in this increasingly interconnected world with mobile devices, will give consumers a portfolio of secure, consistent, and personalized computing experiences across a range of Internet-connected devices. These devices would then be consistent and interoperable amongst themselves for seamless connections and computing capability both locally and in the cloud.
Intel’s goal is to be the preeminent computing solutions company that powers the worldwide digital economy. To achieve this goal they’re transforming from a company with a main focus on the design and production of PC/server semiconductor chips to a computing company that can deliver complete solutions involving hardware, software, and supporting services.
Intel’s operations are broken into three main business segments: the PC Client Group (PCCG), the Data Center Group (DCG), and Other Intel Architecture (IA) that produce 74%, 20%, and 4% of revenue, respectively.
· PCCG: makes microprocessors and related chipsets for the desktop, notebook, and netbook segments. This group additionally makes desktop motherboards and wireless connectivity products.
· DCG: this group’s products includes microprocessors, chipsets, motherboards, and wired connectivity devices used in servers, workstations, storage, and other applications supporting the infrastructure for data centers and cloud computing.
· Other IA: this group is made up Intel’s smaller businesses which include the Embedded and Communications Group (ECG), the Digital Home Group (DHG), the Ultra-Mobility Group (UMG).
o ECG: produces scalable microprocessors and chipsets for various embedded applications in various market segments, such as industrial, medical, and in-vehicle infotainment.
o DHG: makes products used in various consumer electronics that are designed to access and share Internet, broadcast, optical media, and personal content through linked digital devices within a home.
o UMG: offers microprocessors and chipsets for mobile Internet platforms, such as the handheld market segment.
· Other Operating Segments
o NAND Solutions Group: this segment offers NAND flash memory products mainly used in portable memory storage devices, digital cameras, solid-state drives (SSDs), etc. These memory products are made by IM Flash Technologies, LLC (IMFT, Financial).
o Wind River Software Group: develops and licenses software optimization products for devices, including operating systems, to support the needs of consumers in the embedded and mobile market segments.
Intel’s manufacturing operations are mostly integrated. About 61% of their wafer fabrication, including microprocessors and chipsets, was performed in the US at their facilities in Arizona, New Mexico, Oregon, and Massachusetts. The remaining 39% was made outside of the US in plants located in Israel, Ireland, and China. Intel uses third-party foundries to make wafers for certain components, which include networking and communications products. They also use subcontractors to produce board-level products and systems, and buy some communications networking products from vendors in the Asia-Pacific region.
To continue to push the density and performance envelopes, Intel incurs significant start-up costs to ready each factory for each succeeding generation of process technology. Although costly, benefits of migrating to each generation includes higher transistor densities, reduced heat output per transistor, and/or increased quantity of integrated features per chip. The net result is higher performing processors which consume less power and cost less to make.
Intel products flow from manufacturing to assembly and test. These processes are performed at facilities in Malaysia, China, Costa Rica, and Vietnam. Intel augments capacity using subcontractors to assemble certain products, namely chipsets and networking and communications gear.
Intel focuses their R&D on advanced computing technologies, developing new microarchitectures, advancing the silicon manufacturing process technology, delivering the next generation of processors and chipsets, improving platform initiatives, and developing software and tools to support these technologies. They continue to make significant investments to develop systems on chips (SoCs) to enhance growth in areas such as mobile devices, embedded applications, and consumer electronics. They also invest in wireless technologies, graphics, and high-performance computing.
As mentioned previously, Intel operates their technology development on a “tick-tock” schedule which enables them to produce a new processor roughly every two years. Intel has provided leadership in silicon technology over the years, and will be manufacturing products using their new 22nm technology later in 2011. Their leadership in this area has enabled Moore’s Law to become a reality.
Their R&D model is built on a global network of organizations that emphasizes collaboration in identifying and developing new technologies, leading standards setting initiatives, and influencing regulatory policies to speed up the adoption of new technologies. Additionally, their R&D is centrally managed, but de-centrally executed--the different internal business groups perform the R&D, but corporate manages cross-business group initiatives to align and prioritize resources and activities across the groups. Lastly, Intel augments their R&D by investing in companies or entering into agreements with companies that have similar focus areas, such as their joint development agreement with Micron for development of NAND flash memory technologies.
INTELLECTUAL PROPERTY PORTFOLIO
Intel’s intellectual property portfolio includes patents, copyrights, trade secrets, trademarks, and maskwork rights. Interestingly, up until 1984, US law didn’t recognize intellectual property rights related to microprocessor topology (aka microarchitecture), the layout of transistors in a silicon wafer to achieve a design. The Semiconductor Industry Association (SIA) and Intel pursued these rights eventually resulting in the Semiconductor Chip Protection Act of 1984.
Even though patents are integral to Intel’s success over the years, their business isn’t dependent on any single patent. Due to their “tick-tock” development process, their products are typically obsolete before they patents expire. In some cases, products are obsolete before the patent is even granted.
Additionally, the bulk of the software Intel generates and distributes is copyrighted, which includes software (aka firmware) embedded in component-level and system-level products. They also protect details relating to their processes, products, and strategies as trade secrets.
COMPETITION
The semiconductor industry is cyclical, fluid and marked by rapid technological advances and frequent product introductions. As mentioned in the previous section, the life span of their products is typically short. Intel’s ability to compete rests in their ability to out-innovate and improve products and processes faster than competitors. To accomplish this feat requires a massive R&D budget, which Intel has and they’ve successfully out-performed the competition through the years in what would otherwise be a commoditized product. However, Intel is largely dependent on the success of their microprocessor business, which makes up the bulk of their revenue.
Intel’s competitors by product line are:
To expand on the microprocessor segment from above, Intel faces competition with rival architecture designs. IBM (IBM), Sony (SNE), and Toshiba are jointly developing the Cell Broadband Engine Architecture; IBM offers the Power Architecture; Oracle provides the Sun Scalable Processor Architecture (SPARC); and ARM Limited developed the ARM Architecture, which is a 32-bit reduced instruction set computer (RISC) design. ARMS stands for Advanced RISC Machine. Additionally, NVIDIA began developing ARM-based processors to merge with their graphics processors, which offloads some of the traditional workload from a microprocessor to a graphics processor.
ARM processors were originally intended to be used in desktop personal computers; however, this market is now dominated by Intel’s x86 family of processors used extensively in IBM PC-compatible and Apple Macintosh computers. ARM processors are inherently simple and are viable in low-power applications, a feature which has made them dominant in the mobile phone and embedded electronics markets. In 2005, roughly 98% of the mobile phones used at least one ARM processor. As of 2009, about 90% of all embedded 32-bit RISC processors were ARM designs, which are found predominantly in PDAs, mobile phones, digital media/music players, hand-held games, calculators, hard drives, routers, etc. ARM doesn’t manufacture processors. They design, patent, and license the technology intellectual property and offer supporting software and services—a high gross margin enterprise.
Intel’s ability to compete with ARM lies in their ability to design and manufacture high-performance, low-power, processors at competitive prices. The Intel Atom processor is their foray into this market.
Additionally, Intel’s ability to compete and maintain leadership in the PC/server microprocessor space, depends on their talent to continue to out-innovate, and out-market AMD. This’ll require them to continue to push the boundaries of Moore’s Law, of which the ceiling is approaching in the next 10-20 years. Although AMD has shown up Intel occasionally in a single-generation of processor design, they’ve demonstrated a continuous inability over history to upset Intel.
RISKS TO BUSINESS MODEL
Besides the usual risks listed in a 10-K with respect to competition, litigation, margin compression, tax rates, currency exchange, etc., there are a few additional risks from my perspective going forward.
Moore’s Law: as mentioned previously in the Industry Information section, the upper limit for silicon transistor density in an integrated circuit will be coming in the next 10-20 years. This means we can expect another 5 to 10 generations of silicon-based microprocessors. However, unless a new technology appears, microprocessor power and speed will plateau at that point. To date, carbon nano-tube technology may hold promise; however, it’s not viable as a commercially available, mass-produced technology.
Converting semiconductor foundries over to carbon-based methods would obviously be quite expensive for Intel in the beginning. The change would incur a learning curve, as there was for silicon, in order to evolve the technology, increase densities and performance, and lower manufacturing costs. If Intel went to this expense, it would create cannibalization of sales, at least initially, as consumers convert from the old technology to the new.
If Intel doesn’t make the switch, either another firm will create it or the microprocessor industry will stagnate and Intel’s designs will eventually become commodity products. Due to the time required to design and build a foundry, Intel must decide soon when it intends to switch.
Change in business model: Intel’s goal is to be the preeminent computing solutions company that powers the worldwide digital economy. This fact represents a shift in business model from a pure-play microprocessor producer (which represents 75% of their revenue) to a model that’s akin to IBM’s. When companies diversify beyond their area of expertise, the results are usually dismal and they later divest themselves of the non-core business area(s). Additionally, it’s unclear how Intel’s competitive positioning, revenue growth, and earnings growth will fare under this new model.
Although not mentioned in the annual report, I believe this shift is related to the above issue with Moore’s Law.
Mobile Phone Processors: ARM Holdings has a virtual monopoly in the mobile phone processor space—the majority of these devices use processors built and licensed on ARM’s designs. Intel is dipping their toes in the water with their Atom chip, but I believe it’ll be a few years before Intel competes in any significant way with ARM.
However, I believe this aspect is not as dire as it seems—Intel will still benefit even if they never make it to market in this segment. As more and more mobile devices appear and connect to the cloud and demand more data and more services, the cloud build-out will need to occur. This activity would then have beneficial side-effects for Intel as more powerful servers and a higher quantity of servers gets used which will obviously need their microprocessors.
THE $1 PREMISE
In Warren Buffet’s 1983 Owner’s Manual for Berkshire shareholders he stated that earnings retention must, in the long run, deliver at least $1 of market value for every $1 retained in the business. To the extent a business is capable of meeting this test, they’re creating value. Otherwise, earnings retention isn’t being invested wisely and the business, or the management, is value-destructive. It’s a simple and easy acid test to apply.
Upon initial review, it would appear over the last decade Intel hasn’t created shareholder value according to Buffett’s test. The data indicates they’ve destroyed $14 in market value for every $1 retained. Re-examining this metric on the assumption the market cap was still immensely overvalued from the tech bubble of the late 1990s, the table below starts in 2006 which was a cyclical low point in their earnings and covers a sufficiently minimum amount of time:
On this basis, Intel has created shareholder value on the order of $3 in market value for every $1 retained in the business.
CAPITAL STRUCTURE
Intel is capitalized with 5.7B shares of common stock with a current market cap of approximately $120B. They also issued in 2005/2009 junior subordinated convertible debentures of $1.9B, in addition to $128M of bonds from Arizona due in 2035/2037. Their total long-term debt is approximately $2B.
Intel’s liquidity predominantly comes from their cash flow generating abilities. As of the end of 2010, they held $21.5B on the books in cash, cash equivalents, marketable debt, and short-term investments. The Board has an approved, ongoing authorization allowing Intel to borrow up to $3B. Maximum borrowings under this program in 2010 were $150M.
A review of their investments doesn’t reveal anything odd. Their portfolio includes a variety, including government bonds, commercial paper, corporate bonds, bank deposits, marketable equities, asset-backed securities, and money markets. Their marketable equities include: Micron (MU, Financial), Imagination Technologies Group PLC (IMG, Financial), VMWare (VMW, Financial), and Clearwire (CLWR, Financial).
The company has additional non-marketable equity investments which include: a 49% interest in IM Flash Technologies (IMFT) and a 22% interest in IM Flash Singapore (IMFS, Financial); a 45% stake in Numonyx B.V.; and a $1B investment in Clearwire LLC, a wholly owned subsidiary of CLWR. IMFT and IMFS are a joint venture formed with Micron to produce NAND flash memory. In 2008, Intel divested their NOR flash memory business in exchange for their interest in Numonyx. To a smaller extent, the company invested in the SMART Technology (SMT, Financial) IPO later selling one-third of the position for a gain. Lastly, they created an equal-ownership, joint venture with GE in the healthcare industry called Intel-GE Care Innovations, LLC, and was formed by combining GE’s Home Health division and Intel’s Digital Health Group.
Intel’s use of derivatives appears run-of-the-mill and not exotic. They maintain derivatives for hedging against risk is currency exchange, interest rates, commodity prices, and equity markets.
Lastly, Intel has a share buyback program. They’ve reduced the share count from 6.88B shares in 2001 to the current 5.7B shares.
REVENUE, PRE-TAX PROFIT, AND OWNER EARNINGS
A visual review of Intel’s sales, pre-tax profit, net income, and owner earnings curves on a logarithmic graph will quickly illustrate the stability of its operations. Owner earnings are charted along with GAAP accounting earnings (net income), as it is the cash that a business can actually generate for its owners. If there are any accounting shenanigans going on, there’d be a visual divergence in the two curves. Owner earnings is calculated as:
Owner Earnings = Net Income + Depreciation & Amortization + Non-cash charges – Capex
Since owner earnings and net income track each other fairly well, one can conclude Intel is honest in their accounting.
A “wonderful” company with excellent business economics will have straight curves on this chart that parallel each other and rise to the right—a visual indication of excellent management, cost control, and the presence of a competitive advantage. A not-so-wonderful company (ie. those built on commoditized products) will have choppiness in these financial figures, and the curves will be erratic, bouncing up and down.
The semiconductor industry is notoriously cyclical, and many of the chips are essentially commodity products, including microprocessors. As can be seen from the chart above, even though the industry is cyclical, Intel has been able to achieve fairly consistent, “secular-style” revenue growth, in spite of it. However, the cyclicality does affect its underlying cost structure which can be visualized by the pre-tax profit curve. The company’s “tick-tock” processor development model is a major determinant as it causes capital expenditures to increase every couple of years as they ramp up new fabrication processes for their new technology.
Turning to growth rates, the company has enjoyed strong cash flow generation and a massive R&D budget. It’s this budget that allows them to carve a competitive advantage and push processor performance and lower manufacturing costs at a quicker pace than AMD and others that use third-party foundries instead of being vertically-integrated as Intel is. In the last year, owing to a rebounding economy, Intel enjoyed massive net income and owner earnings growth. However, over the long-term revenue growth has roughly kept pace with inflation—about 3%. Intel has steadily grown book value, and tangible book value, from $5.34/share in 2001 to $9.01/share in 2010—just under 5% CAGR.
REVENUE AND COST STRUCTURE
As can be seen from the table below, Intel generates the majority of their revenue, approximately 93%,from microprocessors supporting the PC Client Group and Data Center Group—basically the PC and server markets. The company has a growing revenue volume in the Asia-Pacific region (57%). Revenue in the Americas and Japan over the last few years has been steady at 20% and 10%, respectively. The European market has decline from 19% in 2008 to 13% currently.
The table below summarizes Intel’s recent gross and net margins. Considering the cyclical nature of the industry, Intel has been able to increase gross margins from 49% a decade ago to the current 65% in 2010. Net margin 10 years ago was 4.9% and has additionally trended upward over the last decade. Intel’s average capex over the last 10 years is $5.1B, or about 15% of revenue.
The table below summarizes Intel’s recent history in operating expenses. In order for the company to maintain its technological superiority, they’ll need to continue to invest 15%, or more, of revenue in R&D.
MANAGEMENT EFFECTIVENESS
The table below summarizes a variety of metrics to illustrate management performance and competence in making money for their shareholders. It helps us answer the question: are they efficient allocators of capital?
Cash Return on Invested Capital (CROIC) tells us how efficiently a business’ operations and management can allocate capital into the business to create even more cash for other re-investment projects. In the short-term growth can happen at any rate. But over the long-term, a business will generally grow at the rate it can generate owner earnings. This growth depends on how much cash it generates based on total cash invested by shareholders and bondholders. In short, it can be viewed as the sustainable upper limit to a company’s growth.
Intel generates copious cash and has the management acumen to generate double-digit, albeit cyclical, ROE & CROIC. For every $100 invested, Intel generated $15 in owner earnings, on average.
BASIS OF COMPETITIVE ADVANTAGE
Intel benefits from a competitive advantage, or moat, working in its favor. Their extensive patent portfolio and brand recognition help; however, the bedrock of this advantage is based on their R&D budget which creates their ability to innovate continuously, push the envelope of performance, and keep manufacturing costs low with a vertically-integrated foundry process in their business model.
Additionally, through the creation of chipsets, they achieve product differentiation as their chips are optimized to work with their processors.
VALUATION
This section will value Intel based on two methods: a discounted cash flow (DCF) model, and an Equity-Bond approach. No valuation method is exact since the future is unknowable, we only have history and our sense of the future to make a judgment. The more consistent a company is, the closer the estimate will be. It’s better to be approximately right than precisely wrong and overpay.
To begin, Warren Buffett described the concept of owner earnings as the cash that a business can generate for its owners. He additionally considered owner earnings, not accounting earnings, to be the relevant item for valuation purposes. If one owned the business completely, the amount of cash, or owner earnings, that could conceivably be pulled out of the business without hurting its competitive position or ability to maintain operations is what’s of value.
DCF Model
To value Intel with DCF, we’re interested in our expectation of future owner earnings through a period of time, an appropriate growth rate, and a discount rate (or our required rate of return based on those expected owner earnings cash flows). The basic DCF equation is:
Fair Value = Growth Value (Years 1-10) + Terminal Value (Years 11-20) + Tangible Book Value
Due to the uncertainty presented by Moore’s Law beyond the next decade, it’s reasonable to assume Intel will perform close to their historic performance for the next 10 years and grow at their book value rate. The model slows growth by 10% every few years in the first decade. However, in the subsequent decade, since there’s a change in business model in addition to processor technology uncertainty, this model will assume no growth during that period. Lastly, since 2010 was a very good year for revenue and owner earnings growth, to account for cyclicality this model will start with 2010 owner earnings at the mid-point between 2009 and TTM owner earnings.
The following assumptions are used:
· Growth-Years 1-3 = 5% (in-line with 10-yrear book value growth)
· Growth-Years 4-7 = 4.5% (10% decay rate)
· Growth-Years 8-10 = 4.1% (10% decay rate)
· Terminal-Years 11-20 = 0% (see paragraph above)
· Discount rate = 9% (required rate of return)
· Margin of Safety = 25%
· Initial Owner Earnings (2010) = $7900M (see paragraph above)
· Tangible BV (2010) = $44899M
Projection of future Owner Earnings
[Note: all figures in $Millions]
DCF Fair Value = $25 (Current price = $22.80)
Equity-Bond Model
To value Intel based on an Equity-Bond model, the company’s owner earnings, compared to the price paid for those earnings (aka Owner Earnings Yield), are evaluated against the risk-free investment—the 10 year T-bill. In this view, a stock can be compared against a bond. The stock would have a growing “coupon” (owner earnings) through time roughly at the book value growth rate (5%), whereas a bond would have a static coupon. The Owner Earnings Yield is calculated as:
Owner Earnings Yield (OEY) = Owner Earnings (TTM) / Market Cap
OEY = 8.7% (by comparison, 10-year T-bill = 3.4%)
The DCF analysis indicates Intel is fairly valued. The Equity-Bond analysis indicates it is undervalued relative to the 10-year T-bill—by almost 2.5:1. This appears to be a fair price to pay for Intel’s growth going forward.
CATALYSTS
Intel has many opportunities ahead of it to capitalize on and provide opportunities for growth. They have a culture of innovation and have been able to fuel each technological change in microprocessors, pushing the envelope of Moore’s Law and IC density every two years. Catalysts for continue growth include:
· Higher enterprise information technology spending
· Improved consumer spending
· Increasing Internet use
· Proliferation of wireless communications devices
· Increasing semiconductor penetration in consumer electronics
DISCLOSURE
The author has no position in INTC.
SOURCES:
SEC Form 10-K, Intel (INTC), Fiscal Year ended December 25, 2010
Intel CFO Commentary on First-Quarter 2011 Results
Standard & Poor’s Report, Intel Corp (INTC), March 26, 2011
Morningstar Report, Intel Corporation (INTC), April 20, 2011
Morningstar Financial Data, Fiscal Years 2001-2011
Intel Corporation (INTC, Financial) is the world’s largest chipmaker by revenue ($43B in 2010) and unit shipments, eclipsing their peers in the semiconductor industry. They design, develop, and manufacture integrated digital technology, primarily integrated circuits (ICs, or aka chips) for original equipment manufacturers, original design manufacturers, personal computer (PC) and network communications products users, and other manufacturers of industrial and communications equipment. Additionally, they develop computing platforms—a combination of hardware and software designed to provide a solution.
Although known for their dominant market share in microprocessors, Intel is transforming from this focus to become a company that can deliver complete solutions consisting of hardware/software platforms and supporting services. As stated in their recent 10-K, their goal is to be “the preeminent computing solutions company that powers the worldwide digital economy.”
Intel was originally founded and incorporated in California in 1968 and later reincorporated in 1989 in Delaware. They are based in Santa Clara, California. Their company stock trades on the NASDAQ under the symbol INTC with a market capitalization of $111B. As of the end of 2010, Intel had 82,500 employees worldwide, of which 55% are employed in the US.
COMPANY HISTORY
Intel was founded in 1968 by three men: chemist and physicist, Gordon Moore (of Moore’s Law fame); physicist and co-inventor of the IC, Robert Noyce; and chemical engineer, Andy Grove. Their company’s original name was NM Electronics, before changing it a year later to Integrated Electronics, or “Intel” for short.
At its start, Intel set itself apart by their ability to produce semiconductors. Their primary products were static random access memory (SRAM) chips. During the 1970s their business grew and they expanded and improved their production processes to allow a wider range of products to be made.
Intel created the first commercially available 4-bit microprocessor, the Intel 4004, in 1971. In 1972, they introduced the first 8-bit processor, the 8008. In 1974, the 8080 came out which was wildly popular in the growing personal computer movement and spawned a variant from Zilog, the Z-80, popularized in Radio Shack’s offering, the TRS-80. Motorola got into the microprocessor game as well with its 6500-series processors being used extensively by Apple in its early models of the Apple, Apple II, and the more powerful 68000-series in the line of Macs.
Intel’s big break came with the introduction of the x86 architecture. The 16-bit 8086, and the 8-bit 8088, became the brains of the IBM PC in the early 1980s. Until then, Intel’s business was dominated by dynamic random access memory (DRAM) production. However, they faced increased competition by Japanese manufacturers reducing profitability of this commodity. Fortunately for Intel, the success of the IBM PC convinced then-CEO Andy Grove to shift their focus and concentrate on microprocessors.
Intel followed this success up with the 80186 and 80286 (1982), the 80386 (1985), and the 80486 (1989). By the end of the decade, the rapidly growing PC market buoyed Intel’s position and they became the primary and most profitable hardware supplier to the PC industry. In 1991, they introduced the “Intel Inside” campaign which linked brand loyalty to consumer selection. By the end of the 1990s, Intel’s Pentium line of processors was a household name.
In the 2000s, microprocessor demand slowed and competitors, mainly AMD, gained significant market share. Intel unsuccessfully attempted to diversify their business beyond chips. During 2005, CEO Paul Otellini, reorganized Intel to refocus the core processor and chipset business on platforms—enterprise, digital home, digital health, and mobility. To regain momentum from lost market share, Intel introduced the “tick-tock” product development model so they could regain their technological lead. This program alternated microarchitecture innovation with process innovation in yearly cycles to push a processor out approximately every two years.
Recently, Intel acquired McAfee, a manufacturer of computer security technology, and Infineon Technologies’ Wireless Solutions business. With the Infineon deal, Intel plans to use this technology in laptops, smart phones, netbooks, tablets, and embedded computers in consumer products.
INDUSTRY INFORMATION
In order to understand Intel’s business model, it’s necessary to spend a few moments understanding the basic nature of the technology.
Transistors, a form of semiconductor, are integral to all modern active electronics and ubiquitous in our everyday lives. The first patent for a transistor was in 1925; however, it wasn’t until Bell Labs’ work in 1947 where the potential of the transistor became apparent. Texas Instruments produced the first silicon transistor in 1954 and ushered in the age of miniaturization. Up until this point, active electronics used vacuum tubes and computers were massive along with their equally huge magnetic-core memory.
Transistors can be packaged as separate, discrete components, or integrated onto integrated circuits (ICs), aka chips, to produce complete electronic circuits. This process is accomplished with semiconductor device fabrication using wafers of silicon in a foundry, or fabrication plant, aka “fab.” An astonishingly low price per transistor can be achieved as transistor densities increase, and microprocessors use a lot of transistors. More transistors generally translates to more powerful processors.
Moore’s Law, made famous by co-founder of Intel, Gordon Moore, predicts that transistor density on ICs would double every two years. With each new generation of microprocessor, the envelope has been pushed further and further outward, generally keeping pace with Moore’s Law. However, physics imposes a lower limit on the size of silicon-chip transistors—below a certain size, they’ll produce excessive heat and have their functioning impaired by quantum effects. Experts predict the transistor density wall will come in about 10-20 years. When this happens, we will have reached the limits of silicon-based processor technology—processor speed and performance will plateau at that point, unless a replacement is developed.
Carbon nanotube technology has the promise to replace silicon chips and allow computer speeds to continue increasing. Nanotube-based transistors, known as carbon nanotube field-effect transistors (CNFETs), have been made that work at room temperature and are capable of digital switching. However, a major obstacle exists—the technology to mass produce them. Additionally, incorporating nanotubes into logic-gate circuits with the required densities to match transistor performance of the 1970s, let alone modern performance standards, hasn’t been demonstrated yet. The future is unclear for the next dominant microprocessor technology.
Additionally, there’s a growing trend to keep in mind—the “cloud.” The cloud is nothing more than the Internet; however, how we use and access it in the future will be changing. With an ever-increasingly connected world with computers and mobile devices, more people and more devices will be accessing the cloud. There’s a growing trend where services and file storage will be offered by the cloud and the consumer will access it, where ever he/she is at, with a cloud-connected device, ie. mobile phone, PC, tablet, etc. In effect, this trend means a transfer of raw computing power to the servers making up the backbone infrastructure of the Internet, or cloud, from the currently powerful desktop computers, as the servers would be the devices managing the requests, executing applications, and crunching data. There’s also a trend for faster and faster data transfer, driven in part by video streaming across the cloud. Not only will end devices need to be fast, but the interconnecting devices and infrastructure will need a build-out to accommodate ever-increasing speeds.
BUSINESS MODEL
Intel designs and manufactures computing and communications components, which include microprocessors, motherboards, chipsets, and wired and wireless connectivity devices. Microprocessors account for 15% of the total semiconductor industry’s revenues, and this field is dominated by Intel and AMD. Intel ships over 80% of the world’s microprocessors.
The company’s vision is to create a seamless continuum of personal computing experiences based on their proprietary architecture. Achieving this vision, in this increasingly interconnected world with mobile devices, will give consumers a portfolio of secure, consistent, and personalized computing experiences across a range of Internet-connected devices. These devices would then be consistent and interoperable amongst themselves for seamless connections and computing capability both locally and in the cloud.
Intel’s goal is to be the preeminent computing solutions company that powers the worldwide digital economy. To achieve this goal they’re transforming from a company with a main focus on the design and production of PC/server semiconductor chips to a computing company that can deliver complete solutions involving hardware, software, and supporting services.
Intel’s operations are broken into three main business segments: the PC Client Group (PCCG), the Data Center Group (DCG), and Other Intel Architecture (IA) that produce 74%, 20%, and 4% of revenue, respectively.
· PCCG: makes microprocessors and related chipsets for the desktop, notebook, and netbook segments. This group additionally makes desktop motherboards and wireless connectivity products.
· DCG: this group’s products includes microprocessors, chipsets, motherboards, and wired connectivity devices used in servers, workstations, storage, and other applications supporting the infrastructure for data centers and cloud computing.
· Other IA: this group is made up Intel’s smaller businesses which include the Embedded and Communications Group (ECG), the Digital Home Group (DHG), the Ultra-Mobility Group (UMG).
o ECG: produces scalable microprocessors and chipsets for various embedded applications in various market segments, such as industrial, medical, and in-vehicle infotainment.
o DHG: makes products used in various consumer electronics that are designed to access and share Internet, broadcast, optical media, and personal content through linked digital devices within a home.
o UMG: offers microprocessors and chipsets for mobile Internet platforms, such as the handheld market segment.
· Other Operating Segments
o NAND Solutions Group: this segment offers NAND flash memory products mainly used in portable memory storage devices, digital cameras, solid-state drives (SSDs), etc. These memory products are made by IM Flash Technologies, LLC (IMFT, Financial).
o Wind River Software Group: develops and licenses software optimization products for devices, including operating systems, to support the needs of consumers in the embedded and mobile market segments.
Intel’s manufacturing operations are mostly integrated. About 61% of their wafer fabrication, including microprocessors and chipsets, was performed in the US at their facilities in Arizona, New Mexico, Oregon, and Massachusetts. The remaining 39% was made outside of the US in plants located in Israel, Ireland, and China. Intel uses third-party foundries to make wafers for certain components, which include networking and communications products. They also use subcontractors to produce board-level products and systems, and buy some communications networking products from vendors in the Asia-Pacific region.
To continue to push the density and performance envelopes, Intel incurs significant start-up costs to ready each factory for each succeeding generation of process technology. Although costly, benefits of migrating to each generation includes higher transistor densities, reduced heat output per transistor, and/or increased quantity of integrated features per chip. The net result is higher performing processors which consume less power and cost less to make.
Intel products flow from manufacturing to assembly and test. These processes are performed at facilities in Malaysia, China, Costa Rica, and Vietnam. Intel augments capacity using subcontractors to assemble certain products, namely chipsets and networking and communications gear.
Intel focuses their R&D on advanced computing technologies, developing new microarchitectures, advancing the silicon manufacturing process technology, delivering the next generation of processors and chipsets, improving platform initiatives, and developing software and tools to support these technologies. They continue to make significant investments to develop systems on chips (SoCs) to enhance growth in areas such as mobile devices, embedded applications, and consumer electronics. They also invest in wireless technologies, graphics, and high-performance computing.
As mentioned previously, Intel operates their technology development on a “tick-tock” schedule which enables them to produce a new processor roughly every two years. Intel has provided leadership in silicon technology over the years, and will be manufacturing products using their new 22nm technology later in 2011. Their leadership in this area has enabled Moore’s Law to become a reality.
Their R&D model is built on a global network of organizations that emphasizes collaboration in identifying and developing new technologies, leading standards setting initiatives, and influencing regulatory policies to speed up the adoption of new technologies. Additionally, their R&D is centrally managed, but de-centrally executed--the different internal business groups perform the R&D, but corporate manages cross-business group initiatives to align and prioritize resources and activities across the groups. Lastly, Intel augments their R&D by investing in companies or entering into agreements with companies that have similar focus areas, such as their joint development agreement with Micron for development of NAND flash memory technologies.
INTELLECTUAL PROPERTY PORTFOLIO
Intel’s intellectual property portfolio includes patents, copyrights, trade secrets, trademarks, and maskwork rights. Interestingly, up until 1984, US law didn’t recognize intellectual property rights related to microprocessor topology (aka microarchitecture), the layout of transistors in a silicon wafer to achieve a design. The Semiconductor Industry Association (SIA) and Intel pursued these rights eventually resulting in the Semiconductor Chip Protection Act of 1984.
Even though patents are integral to Intel’s success over the years, their business isn’t dependent on any single patent. Due to their “tick-tock” development process, their products are typically obsolete before they patents expire. In some cases, products are obsolete before the patent is even granted.
Additionally, the bulk of the software Intel generates and distributes is copyrighted, which includes software (aka firmware) embedded in component-level and system-level products. They also protect details relating to their processes, products, and strategies as trade secrets.
COMPETITION
The semiconductor industry is cyclical, fluid and marked by rapid technological advances and frequent product introductions. As mentioned in the previous section, the life span of their products is typically short. Intel’s ability to compete rests in their ability to out-innovate and improve products and processes faster than competitors. To accomplish this feat requires a massive R&D budget, which Intel has and they’ve successfully out-performed the competition through the years in what would otherwise be a commoditized product. However, Intel is largely dependent on the success of their microprocessor business, which makes up the bulk of their revenue.
Intel’s competitors by product line are:
| Product Line | Market Segment | Competitors |
| Microprocessors | PC Client | AMD, QUALCOMM, VIA |
| Server | AMD, IBM, Oracle | |
| Application Processors | AMD, Broadcom, Freescale Semiconductor, MediaTek, NVIDIA, QUALCOMM, Samsung Electronics, STMicroelectronics, Texas Instruments | |
| Mobile Processors | ARM, MediaTek, QUALCOMM, ST-Ericsson, Texas Instruments | |
| Chipsets | AMD, Broadcom, NVIDIA, Silicon Integrated Systems, VIA | |
| Flash Memory | Hynix Semiconductor, Micron, Samsung, SanDisk, Toshiba | |
| Connectivity | Atheros Communications, Broadcom, QUALCOMM |
To expand on the microprocessor segment from above, Intel faces competition with rival architecture designs. IBM (IBM), Sony (SNE), and Toshiba are jointly developing the Cell Broadband Engine Architecture; IBM offers the Power Architecture; Oracle provides the Sun Scalable Processor Architecture (SPARC); and ARM Limited developed the ARM Architecture, which is a 32-bit reduced instruction set computer (RISC) design. ARMS stands for Advanced RISC Machine. Additionally, NVIDIA began developing ARM-based processors to merge with their graphics processors, which offloads some of the traditional workload from a microprocessor to a graphics processor.
ARM processors were originally intended to be used in desktop personal computers; however, this market is now dominated by Intel’s x86 family of processors used extensively in IBM PC-compatible and Apple Macintosh computers. ARM processors are inherently simple and are viable in low-power applications, a feature which has made them dominant in the mobile phone and embedded electronics markets. In 2005, roughly 98% of the mobile phones used at least one ARM processor. As of 2009, about 90% of all embedded 32-bit RISC processors were ARM designs, which are found predominantly in PDAs, mobile phones, digital media/music players, hand-held games, calculators, hard drives, routers, etc. ARM doesn’t manufacture processors. They design, patent, and license the technology intellectual property and offer supporting software and services—a high gross margin enterprise.
Intel’s ability to compete with ARM lies in their ability to design and manufacture high-performance, low-power, processors at competitive prices. The Intel Atom processor is their foray into this market.
Additionally, Intel’s ability to compete and maintain leadership in the PC/server microprocessor space, depends on their talent to continue to out-innovate, and out-market AMD. This’ll require them to continue to push the boundaries of Moore’s Law, of which the ceiling is approaching in the next 10-20 years. Although AMD has shown up Intel occasionally in a single-generation of processor design, they’ve demonstrated a continuous inability over history to upset Intel.
RISKS TO BUSINESS MODEL
Besides the usual risks listed in a 10-K with respect to competition, litigation, margin compression, tax rates, currency exchange, etc., there are a few additional risks from my perspective going forward.
Moore’s Law: as mentioned previously in the Industry Information section, the upper limit for silicon transistor density in an integrated circuit will be coming in the next 10-20 years. This means we can expect another 5 to 10 generations of silicon-based microprocessors. However, unless a new technology appears, microprocessor power and speed will plateau at that point. To date, carbon nano-tube technology may hold promise; however, it’s not viable as a commercially available, mass-produced technology.
Converting semiconductor foundries over to carbon-based methods would obviously be quite expensive for Intel in the beginning. The change would incur a learning curve, as there was for silicon, in order to evolve the technology, increase densities and performance, and lower manufacturing costs. If Intel went to this expense, it would create cannibalization of sales, at least initially, as consumers convert from the old technology to the new.
If Intel doesn’t make the switch, either another firm will create it or the microprocessor industry will stagnate and Intel’s designs will eventually become commodity products. Due to the time required to design and build a foundry, Intel must decide soon when it intends to switch.
Change in business model: Intel’s goal is to be the preeminent computing solutions company that powers the worldwide digital economy. This fact represents a shift in business model from a pure-play microprocessor producer (which represents 75% of their revenue) to a model that’s akin to IBM’s. When companies diversify beyond their area of expertise, the results are usually dismal and they later divest themselves of the non-core business area(s). Additionally, it’s unclear how Intel’s competitive positioning, revenue growth, and earnings growth will fare under this new model.
Although not mentioned in the annual report, I believe this shift is related to the above issue with Moore’s Law.
Mobile Phone Processors: ARM Holdings has a virtual monopoly in the mobile phone processor space—the majority of these devices use processors built and licensed on ARM’s designs. Intel is dipping their toes in the water with their Atom chip, but I believe it’ll be a few years before Intel competes in any significant way with ARM.
However, I believe this aspect is not as dire as it seems—Intel will still benefit even if they never make it to market in this segment. As more and more mobile devices appear and connect to the cloud and demand more data and more services, the cloud build-out will need to occur. This activity would then have beneficial side-effects for Intel as more powerful servers and a higher quantity of servers gets used which will obviously need their microprocessors.
THE $1 PREMISE
In Warren Buffet’s 1983 Owner’s Manual for Berkshire shareholders he stated that earnings retention must, in the long run, deliver at least $1 of market value for every $1 retained in the business. To the extent a business is capable of meeting this test, they’re creating value. Otherwise, earnings retention isn’t being invested wisely and the business, or the management, is value-destructive. It’s a simple and easy acid test to apply.
| $1 Test | 2010 | 2001 | Change |
| Retained Earnings | $ 32,919 | $ 27,150 | $ 5,769 |
| Market Value | $ 128,855 | $ 211,092 | $ (82,237) |
Upon initial review, it would appear over the last decade Intel hasn’t created shareholder value according to Buffett’s test. The data indicates they’ve destroyed $14 in market value for every $1 retained. Re-examining this metric on the assumption the market cap was still immensely overvalued from the tech bubble of the late 1990s, the table below starts in 2006 which was a cyclical low point in their earnings and covers a sufficiently minimum amount of time:
| $1 Test | 2010 | 2006 | Change |
| Retained Earnings | $ 32,919 | $ 28,984 | $ 3,935 |
| Market Value | $ 128,855 | $ 116,762 | $ 12,093 |
On this basis, Intel has created shareholder value on the order of $3 in market value for every $1 retained in the business.
CAPITAL STRUCTURE
Intel is capitalized with 5.7B shares of common stock with a current market cap of approximately $120B. They also issued in 2005/2009 junior subordinated convertible debentures of $1.9B, in addition to $128M of bonds from Arizona due in 2035/2037. Their total long-term debt is approximately $2B.
Intel’s liquidity predominantly comes from their cash flow generating abilities. As of the end of 2010, they held $21.5B on the books in cash, cash equivalents, marketable debt, and short-term investments. The Board has an approved, ongoing authorization allowing Intel to borrow up to $3B. Maximum borrowings under this program in 2010 were $150M.
A review of their investments doesn’t reveal anything odd. Their portfolio includes a variety, including government bonds, commercial paper, corporate bonds, bank deposits, marketable equities, asset-backed securities, and money markets. Their marketable equities include: Micron (MU, Financial), Imagination Technologies Group PLC (IMG, Financial), VMWare (VMW, Financial), and Clearwire (CLWR, Financial).
The company has additional non-marketable equity investments which include: a 49% interest in IM Flash Technologies (IMFT) and a 22% interest in IM Flash Singapore (IMFS, Financial); a 45% stake in Numonyx B.V.; and a $1B investment in Clearwire LLC, a wholly owned subsidiary of CLWR. IMFT and IMFS are a joint venture formed with Micron to produce NAND flash memory. In 2008, Intel divested their NOR flash memory business in exchange for their interest in Numonyx. To a smaller extent, the company invested in the SMART Technology (SMT, Financial) IPO later selling one-third of the position for a gain. Lastly, they created an equal-ownership, joint venture with GE in the healthcare industry called Intel-GE Care Innovations, LLC, and was formed by combining GE’s Home Health division and Intel’s Digital Health Group.
Intel’s use of derivatives appears run-of-the-mill and not exotic. They maintain derivatives for hedging against risk is currency exchange, interest rates, commodity prices, and equity markets.
Lastly, Intel has a share buyback program. They’ve reduced the share count from 6.88B shares in 2001 to the current 5.7B shares.
REVENUE, PRE-TAX PROFIT, AND OWNER EARNINGS
A visual review of Intel’s sales, pre-tax profit, net income, and owner earnings curves on a logarithmic graph will quickly illustrate the stability of its operations. Owner earnings are charted along with GAAP accounting earnings (net income), as it is the cash that a business can actually generate for its owners. If there are any accounting shenanigans going on, there’d be a visual divergence in the two curves. Owner earnings is calculated as:
Owner Earnings = Net Income + Depreciation & Amortization + Non-cash charges – Capex
Since owner earnings and net income track each other fairly well, one can conclude Intel is honest in their accounting.
A “wonderful” company with excellent business economics will have straight curves on this chart that parallel each other and rise to the right—a visual indication of excellent management, cost control, and the presence of a competitive advantage. A not-so-wonderful company (ie. those built on commoditized products) will have choppiness in these financial figures, and the curves will be erratic, bouncing up and down.
The semiconductor industry is notoriously cyclical, and many of the chips are essentially commodity products, including microprocessors. As can be seen from the chart above, even though the industry is cyclical, Intel has been able to achieve fairly consistent, “secular-style” revenue growth, in spite of it. However, the cyclicality does affect its underlying cost structure which can be visualized by the pre-tax profit curve. The company’s “tick-tock” processor development model is a major determinant as it causes capital expenditures to increase every couple of years as they ramp up new fabrication processes for their new technology.
Turning to growth rates, the company has enjoyed strong cash flow generation and a massive R&D budget. It’s this budget that allows them to carve a competitive advantage and push processor performance and lower manufacturing costs at a quicker pace than AMD and others that use third-party foundries instead of being vertically-integrated as Intel is. In the last year, owing to a rebounding economy, Intel enjoyed massive net income and owner earnings growth. However, over the long-term revenue growth has roughly kept pace with inflation—about 3%. Intel has steadily grown book value, and tangible book value, from $5.34/share in 2001 to $9.01/share in 2010—just under 5% CAGR.
| Growth Rates | 1 Yr | 3 Yr | 5 Yr | 10 Yr |
| Revenue | 24.2% | 4.4% | 2.4% | 2.6% |
| Operating income | 172.9% | 23.8% | 5.2% | 4.1% |
| Net Income | 161.0% | 19.4% | 7.5% | 2.9% |
| Owner Earnings | 122.1% | 17.2% | 7.9% | 36.9% |
| Book Value | 18.8% | 7.0% | 8.4% | 4.9% |
| Tangible Book Value | 20.4% | 4.9% | 6.8% | 4.3% |
REVENUE AND COST STRUCTURE
As can be seen from the table below, Intel generates the majority of their revenue, approximately 93%,from microprocessors supporting the PC Client Group and Data Center Group—basically the PC and server markets. The company has a growing revenue volume in the Asia-Pacific region (57%). Revenue in the Americas and Japan over the last few years has been steady at 20% and 10%, respectively. The European market has decline from 19% in 2008 to 13% currently.
| Revenue ($B) | 2008 | 2009 | 2010 | MRQ | Avg % of Sales |
| PCCG | $27.966 | $26.175 | $31.598 | $8.6 | 74% |
| DCG | $6.590 | $6.450 | $8.693 | $2.5 | 19% |
| Other IA | $1.763 | $1.402 | $1.784 | $1.1 | 4% |
| All Other | $1.267 | $1.1 | $1.548 | $0.6 | 3% |
| Total | $37.586 | $35.127 | $43.623 | $12.8 |
The table below summarizes Intel’s recent gross and net margins. Considering the cyclical nature of the industry, Intel has been able to increase gross margins from 49% a decade ago to the current 65% in 2010. Net margin 10 years ago was 4.9% and has additionally trended upward over the last decade. Intel’s average capex over the last 10 years is $5.1B, or about 15% of revenue.
| Margins | 2008 | 2009 | 2010 | MRQ |
| Gross | 55.5% | 55.7% | 65.3% | 62.3% |
| Net | 14.1% | 12.4% | 26.3% | 24.6% |
The table below summarizes Intel’s recent history in operating expenses. In order for the company to maintain its technological superiority, they’ll need to continue to invest 15%, or more, of revenue in R&D.
| Operating Expenses | 2008 | 2009 | 2010 |
| R&D | 15.2% | 16.1% | 15.1% |
| Sales | 44.5% | 44.3% | 34.7% |
| MG&A | 14.6% | 22.6% | 14.5% |
MANAGEMENT EFFECTIVENESS
The table below summarizes a variety of metrics to illustrate management performance and competence in making money for their shareholders. It helps us answer the question: are they efficient allocators of capital?
Cash Return on Invested Capital (CROIC) tells us how efficiently a business’ operations and management can allocate capital into the business to create even more cash for other re-investment projects. In the short-term growth can happen at any rate. But over the long-term, a business will generally grow at the rate it can generate owner earnings. This growth depends on how much cash it generates based on total cash invested by shareholders and bondholders. In short, it can be viewed as the sustainable upper limit to a company’s growth.
| Efficiency & Profitability | 2001 | 2002 | 2003 | 2004 | 2005 | 2006 | 2007 | 2008 | 2009 | 2010 | Avg |
| CROIC | 2.0% | 11.3% | 21.0% | 23.5% | 21.1% | 11.6% | 15.7% | 12.3% | 11.9% | 22.1% | 15.3% |
| FCF/Sales | 2.4% | 14.0% | 25.5% | 25.0% | 19.2% | 11.8% | 17.7% | 12.5% | 14.0% | 25.0% | 16.7% |
| ROA | 2.8% | 7.0% | 12.4% | 15.8% | 18.0% | 10.4% | 13.4% | 9.9% | 8.4% | 19.7% | 11.8% |
| ROE | 3.5% | 8.7% | 15.4% | 19.7% | 23.2% | 13.8% | 17.6% | 12.9% | 10.8% | 25.2% | 15.0% |
Intel generates copious cash and has the management acumen to generate double-digit, albeit cyclical, ROE & CROIC. For every $100 invested, Intel generated $15 in owner earnings, on average.
BASIS OF COMPETITIVE ADVANTAGE
Intel benefits from a competitive advantage, or moat, working in its favor. Their extensive patent portfolio and brand recognition help; however, the bedrock of this advantage is based on their R&D budget which creates their ability to innovate continuously, push the envelope of performance, and keep manufacturing costs low with a vertically-integrated foundry process in their business model.
Additionally, through the creation of chipsets, they achieve product differentiation as their chips are optimized to work with their processors.
VALUATION
This section will value Intel based on two methods: a discounted cash flow (DCF) model, and an Equity-Bond approach. No valuation method is exact since the future is unknowable, we only have history and our sense of the future to make a judgment. The more consistent a company is, the closer the estimate will be. It’s better to be approximately right than precisely wrong and overpay.
To begin, Warren Buffett described the concept of owner earnings as the cash that a business can generate for its owners. He additionally considered owner earnings, not accounting earnings, to be the relevant item for valuation purposes. If one owned the business completely, the amount of cash, or owner earnings, that could conceivably be pulled out of the business without hurting its competitive position or ability to maintain operations is what’s of value.
DCF Model
To value Intel with DCF, we’re interested in our expectation of future owner earnings through a period of time, an appropriate growth rate, and a discount rate (or our required rate of return based on those expected owner earnings cash flows). The basic DCF equation is:
Fair Value = Growth Value (Years 1-10) + Terminal Value (Years 11-20) + Tangible Book Value
Due to the uncertainty presented by Moore’s Law beyond the next decade, it’s reasonable to assume Intel will perform close to their historic performance for the next 10 years and grow at their book value rate. The model slows growth by 10% every few years in the first decade. However, in the subsequent decade, since there’s a change in business model in addition to processor technology uncertainty, this model will assume no growth during that period. Lastly, since 2010 was a very good year for revenue and owner earnings growth, to account for cyclicality this model will start with 2010 owner earnings at the mid-point between 2009 and TTM owner earnings.
The following assumptions are used:
· Growth-Years 1-3 = 5% (in-line with 10-yrear book value growth)
· Growth-Years 4-7 = 4.5% (10% decay rate)
· Growth-Years 8-10 = 4.1% (10% decay rate)
· Terminal-Years 11-20 = 0% (see paragraph above)
· Discount rate = 9% (required rate of return)
· Margin of Safety = 25%
· Initial Owner Earnings (2010) = $7900M (see paragraph above)
· Tangible BV (2010) = $44899M
Projection of future Owner Earnings
| Year | 2011 | 2012 | 2013 | 2014 | 2015 | 2016 | 2017 | 2018 | 2019 | 2020 |
| Earnings | 8,295 | 8,710 | 9,145 | 9,421 | 9,845 | 10,288 | 10,751 | 10,853 | 11,293 | 11,750 |
| Year | 2021 | 2022 | 2023 | 2024 | 2025 | 2026 | 2027 | 2028 | 2029 | 2030 |
| Earnings | 11,750 | 11,750 | 11,750 | 11,750 | 11,750 | 11,750 | 11,750 | 11,750 | 11,750 | 11,750 |
[Note: all figures in $Millions]
DCF Fair Value = $25 (Current price = $22.80)
Equity-Bond Model
To value Intel based on an Equity-Bond model, the company’s owner earnings, compared to the price paid for those earnings (aka Owner Earnings Yield), are evaluated against the risk-free investment—the 10 year T-bill. In this view, a stock can be compared against a bond. The stock would have a growing “coupon” (owner earnings) through time roughly at the book value growth rate (5%), whereas a bond would have a static coupon. The Owner Earnings Yield is calculated as:
Owner Earnings Yield (OEY) = Owner Earnings (TTM) / Market Cap
OEY = 8.7% (by comparison, 10-year T-bill = 3.4%)
The DCF analysis indicates Intel is fairly valued. The Equity-Bond analysis indicates it is undervalued relative to the 10-year T-bill—by almost 2.5:1. This appears to be a fair price to pay for Intel’s growth going forward.
CATALYSTS
Intel has many opportunities ahead of it to capitalize on and provide opportunities for growth. They have a culture of innovation and have been able to fuel each technological change in microprocessors, pushing the envelope of Moore’s Law and IC density every two years. Catalysts for continue growth include:
· Higher enterprise information technology spending
· Improved consumer spending
· Increasing Internet use
· Proliferation of wireless communications devices
· Increasing semiconductor penetration in consumer electronics
DISCLOSURE
The author has no position in INTC.
SOURCES:
SEC Form 10-K, Intel (INTC), Fiscal Year ended December 25, 2010
Intel CFO Commentary on First-Quarter 2011 Results
Standard & Poor’s Report, Intel Corp (INTC), March 26, 2011
Morningstar Report, Intel Corporation (INTC), April 20, 2011
Morningstar Financial Data, Fiscal Years 2001-2011
