Seeking Alpha
Profile| Send Message|
( followers)  

Intel (NASDAQ:INTC) is bombarded by negative headlines: lower revenue guidance, missing the smartphone and tablet train, ARM's (NASDAQ:ARMH) poised invasion into its home turf, etc. The share price has plummeted more than 20% this year. Is Intel really becoming irrelevant or even obsolete? Media hype aside, Intel still holds tremendous competitive advantages. The low valuation offers adequate return even though we only account for its PC and data center business. In addition, Intel possesses great potential to regain a foothold in the mobile processor market. At current share price, this upside is a free option for long-term value investors.

Is PC market really dying in "post-PC era"?

Let's start from Intel's PC processor business. Intel's dominance in this market is well known. The company has 75% market share in desktop and 85% in notebook. The only competitor, AMD, occupies mainly the low-end market and holds neither technology nor cost advantage over Intel. Of course, being the leader in a dying business may not be a positive thing, but the PC market is far from dying despite all the talk of the post-PC era. Smartphones and tablets offer an alternative to many tasks traditionally performed by PCs. But those devices are better suited to content consumption and light communication. For serious production work such as content creation, editing or analysis, a full-fledged desktop or laptop remains indispensable. You may opt for a Mac instead of a Windows PC, but Mac also uses an Intel processor.

We can break down PC demand into two categories: new installation and replacement. In the following table, we performed a simple analysis of worldwide PC demand. Start from the data on a PC-installed base and shipment (from Gartner and IDC). The annual net new PC units in the table are simply the difference between two years' installed base. The rest of the PC shipment will be for replacement purposes. To estimate the replacement cycle, we first discount the PC installed base in the year by a number of years' compound growth rate. This gives the approximate unit base to be replaced. We then estimate the average time a PC gets replaced by dividing this replacement base by the annual replacement units. (In the table, we set the number of years to discount the unit base to be the same as the replacement period and solve for the result.) The resulting replacement period of four to five years looks reasonable by real-world standards. Tablets and smartphones are picking up some demand from PCs. So the replacement period may indeed extend. To estimate the impact, we assume the replacement period to trend up and stabilize at seven years by 2014. The replacement period will not extend much beyond that. Since 2008, 60% of the PCs shipped are for commercial use. Business cannot live with too old machines. Also, nearly 60% of the new PCs are notebooks. Laptops tend to be upgraded more often than desktops.

We further assume the net newly installed PCs will grow at a 5% annual rate. At this growth rate, our estimated net increase of PC units between 2012 and 2015 is 17% lower than in Gartner's forecast. Put these things together, the annual PC shipment will still average 325 million units in the next four years. Notice that once the replacement period peaks, annual PC shipment will grow again because of the larger installed base. In this dire (but not insurmountable) scenario, global PC demand will decrease only 1 to 2% per year on average by 2015, though much of the drop may be front-loaded due to the extension of the replacement period.

Worldwide PC shipment analysis and forecast

Year

PC units installed

PC units shipment

New PC units

Replaced units

Replacement cycle

PC units shipment

unit

million

million

million

million

Years

yoy

A

B

Ct=At-At-1

D=B-C

E=A/(1.117E)/D

2005

798

208

2006

885

235

87

148

3.9

13.3%

2007

982

268

97

171

3.8

13.7%

2008

1,100

288

118

170

4.1

7.4%

2009

1,250

296

150

146

4.9

2.9%

2010

1,400

347

150

197

4.3

17.2%

2011

1,540

353

140

213

4.4

1.8%

2012E

1,694

347

147

200

5.0

-1.8%

2013E

1,780

312

154

157

6.0

-10.1%

2014E

2,240

315

162

153

7.0

1.0%

2015E

2,300

327

170

157

7.0

3.9%

Annual growth rate

2006-11

11.7%

8.5%

10.0%

7.5%

9.4%

2011-15

10.5%

5.0%

-1.8%

Numbers in bold font are from Gartner, IDC

Don't forget servers

In contrast to the PC market, the server market is still growing even in 2012. The shift toward cloud computing and online storage creates a tail wind for data center demand. Gartner's preliminary statistics show that worldwide server unit shipments in 3Q12 increased 5.1% year over year. 1Q and 2Q shipments were also up from a year before. Within the server market, the x86 based server has been gaining popularity at the expense of mainframe and RISC/UNIX based servers. According to IDC, more than 96% of 2009 server units shipped were x86 servers. The trend toward server virtualization generates demand for more powerful x86 servers to replace low-end UNIX servers. As a result, the average selling price in this segment has actually crept up in the past three years. Additionally, the recent shift toward blade servers further accelerates this transition, as 90% of blade servers are x86 based. In the first half of 2012, blade servers represent 16.9% of all server revenue and grow at a double-digit rate. Intel's position in this market is even more overwhelming than in the PC market. In 2011, Intel's share of x86 server revenue was 94.5%. Furthermore, IDC or Gartner may underestimate the actual server demand growth. Large web companies such as Google, Facebook and Amazon increasingly bypass traditional server makers and buy servers and chips directly from ODMs and Intel. In a recent article, Google was estimated to be the fifth-largest buyer of Intel server chips.

In 2011, Intel's revenue composite is roughly 66% from PC Client Group and 20% from Data Center Group. Using the aforementioned scenario, a flat ASP will translate into an annual growth rate of -2% for PC and 5% for server in the next three to four years. At the current sales mix, Intel's overall top line growth rate in this slump is close to 0.

A dreadful circumstance, hypothetically

For value investors with a longer time horizon, the main question is how much of this downside scenario is priced in, or in other words, what kind of long-term return investors can get in this circumstance.

To assess a stock's downside, I often find it useful to value it in a simple but extreme scenario. In Intel's case, let's assume sales stay flat forever. The top line pressure also prompts much talk on margin compression. Intel's operating margin was 32.4% in 2011. The past seven-year average was 25.2%. We take it further down to a new norm average of 24.5% to account for the long-term margin compression (this number is below the operating margin of other, better-run semiconductor companies and foundries such as Qualcomm and TSMC). At this margin, Intel can generate $13.23 billion operating income from $54 billion revenue. That is $2.66 per share based on the latest shares outstanding. Using 28% effective tax rate, the net income per share will be $1.92. (Intel's interest expense net interest income is less than $100 million.)

We can further translate these accrual-based earnings into cash flow space. Intel's seven-year average capital expenditure plus non-financial investment are roughly 1.6 times the depreciation and amortization. For a company with forever-flat revenue, we assume the management will be sensible and refrain from any excessive capital spending or acquisition. So Capex will eventually go down to the asset replacement level, which is close to depreciation and amortization. The net working capital will also stay at around the same level. (Intel's past seven-year total net working capital change is indeed close to 0.) Put these together, the free cash flow per share will be very close to the EPS of $1.92. Intel has a strong record of returning capital to shareholders. Its seven-year payout ratio (dividend + buyback as % of FCF) is 130%. The number is skewed by the large share buyback in 2011, which was partially funded by debt issuance. If we assume the entire FCF is paid out as dividends, at $20 share price, Intel's stock will become a 9.6% annuity-not bad in such a dire scenario (of course, management needs to be rational to cut Capex in this situation).

Falling empire, barbarians at the gate. Really?

Few companies can sustain zero growth forever. They either sink or reform and resurrect. For the investor, it is important to take a look at the forces that can potentially drive Intel into either circumstance.

Flat or even a slight decline of sales is often cited by media as the reason for Intel's share price drop in 2012. But for this revenue pressure to be permanent, you need one of the two things to happen:

  1. Structural decline of PCs in the current form and Intel and Microsoft do not reinvent themselves to adapt to the new personal computing demand.
  2. ARM-based CPU encroaches into Intel's home turf of the PC and server market.

We already discussed the first concern in the PC market analysis above. The shift of PC toward small and mobile form factor is nothing new. Worldwide laptop sales have exceeded those of desktops since 2008 when laptops became powerful enough to take on many commonly performed tasks originally done on an installed PC. The initial rise of tablets follows a similar path. But tablets are not designed for content creation or analysis purposes. The models currently on the market and in the pipeline are not nearly powerful enough to replace laptops. They can only partially substitute PCs, mainly on demand for content consumption. We are not saying that there will not be cannibalization on PC sales, but much of the impact has probably already taken place. When news came out that Microsoft was designing a hybrid tablet-laptop, Apple CEO Tim Cook's scoffed at the idea as converging a toaster and refrigerator and "no one wants it". But consumers disagree. In an online survey, more than 70% of the respondents say that they will buy such a device if it is available. The survey result reveals an underserved market as well as the limitation of tablets. The stage is set for both sides to bring their products to the next level. Intel (and Microsoft) faces a challenge as well as an opportunity to adapt and reinvent their products quickly to meet the new mobile computing demand. At the present time, it is unclear whether the tablet can permanently carve out a market as a dedicated device for content consumption or whether the notebook PC will become small enough and power efficient enough to reclaim this role. The war between tablet and PC is far from over.

If PCs are not going away, what if PC makers shift to ARM-based CPUs? Periodically we see news or rumors on ARM PCs, including the one from Microsoft: Surface RT. ARM processors dominate the mobile device market because they have lower power consumption. But the energy efficiency is at the expense of performance. Apparently, the Nvidia (NASDAQ:NVDA) Tegra 3-powered Surface RT is not powerful enough to run a full-scale Windows 8. Microsoft has to use a trimmed-down version with limited functions. ARM chip-based Surface RT is more a tablet than a laptop. Unlike the smartphone and tablet, a PC needs to support a multi-thread OS and other full-scale software rather than bite-size apps. The processing speed is at least as important as power consumption. Current ARM chips are barely up to these tasks. Samsung's Exynos 5 Dual, which is based on the latest ARM Cortex-A15, consumes a similar amount of power as Intel's soon-to-be-released Haswell processors but runs slower than two generations old Sandy Bridge processors. Even Apple still uses Intel's Ivy Bridge processors in its latest MacBook laptop despite the fact that all iPads and iPhones run on ARM-based processors. The barrier of entry in the server market is even higher. Nowadays, most new servers run on 64-bit x86-based CPUs. ARM just announced its 64-bit design in October 2011. There is no commercially available chip on the market yet. In addition, porting an operating system to a new set of core processors is no light task. For PC and server manufacturers and business users, the issues in compatibility and reliability are formidable barriers to switching. Plus, in the PC and server market ARM can no longer enjoy the ecosystem it built in the mobile device market, while Wintel can still count on the network effect advantage from its vast installed base.

Intel strikes back

So if Intel's home turf is not as threatened as the media proclaims, what is Intel's chance to regain a foothold in the mobile market? Intel faces an uphill battle to make inroads there. At present, Atom processors, Intel's answer to the mobile device market, hold no significant technology advance over ARM's chips. The latest Atom Clover Trail Z2760 may be slightly faster than Nvidia Tegra 3, which is based on Cortex-A9, the same ARM core (albeit a customized version) in iPad and iPhone. But the number of Clover Trail-based tablets is still small. In contrast, ARM has built a large installed base among popular tablet devices and their chip makers. They will be reluctant to switch. This established ARM processor ecosystem is a big obstacle for Intel to clear in the mobile device market.

But Intel is catching up. It has refocused its resource to cut power consumption in its processors. The company's Atom platform (~2w power level) targets the low-to-middle-range tablets and competes directly with ARM-based processors. Atom chips have not generated much buzz so far but the next generation Bay View-T may change the landscape in 2014 with better performance at a lower power level. Other Intel core processors also see improvement in energy efficiency. The 4th generation Haswell architecture, which covers both high-end laptops and servers, will consume as little as 8w power. In addition to processor architecture, Intel possesses a unique advantage in the fab process. Unlike ARM, Nvidia, Qualcomm and Apple, Intel has its own foundries. Its semiconductor processing technology is considered to be two generations ahead of competitors such as TSMC and Global Foundries. The vertical integrated model gives it an edge to improve its processors on both speed and power consumption. There are also rumors that Intel may soon manufacture chips for Apple iOS devices such as iPads and iPhones. Intel's CEO Otellini confirmed the negotiation with Apple in a recent buy-side meeting. If Intel pulls off the deal, it will help to maintain its gross margin as the idle capacity increases its cost of goods sold. The tablet market is still developing. Consumer behavior has yet to become deeply entrenched. Microsoft's Surface Pro, which will be released in January 2013 and runs on Intel mobile core i5, will be the first serious attempt by Wintel to regain a foothold in the tablet market. If it is well received, it may set a new category for the tablet-laptop hybrid and once again change the landscape of the mobile device market.

Back to stock

Now let's take a look at Intel's book. Intel's balance sheet is strong. At the end of 3Q12, the company has more than $10 billion cash and short-term investment against a $9 billion debt. Neither debt servicing nor maturity profile is a concern. Intel's capital allocation is also rational. Capex is adequate at around 15% of revenue and 1.2 times depreciation and amortization. Large acquisition is not frequent, though 2011's $6.7 billion acquisition of McAfee is controversial. Manager tends to return the excess capital to shareholders in the form of dividend and buyback. The quality of dividend is high, on average 1.8 times covered by free cash flows in the past five years. The rest of FCFs go to share buyback, which is also sensible currently given the low valuation of the stock.

So how much is Intel share worth? It depends on the assumptions used. We apply a two-stage DCF model in our equity valuation in two scenarios. In a bearish case, we assume 0 revenue growth from 2012 to 2019 to account for the slow sales when the PC market is in transition. Using historical average Capex intensity and 24.5% operating margin (to be conservative, we take a 1.4% haircut from a five-year average margin of 25.9%), Intel can generate an average $7.6 billion unlevered FCF per year in the next seven years. Using an 11% discount rate, the present value of this cash-flow stream is $37.6 billion. After 2019, we assume long-term revenue growth resumes to 3%. At an 11% discount rate, this translates to a terminal FCF multiple of 12.5 (12.875 in exact math) and a present value of $49 billion from the terminal value. Put these together and net out the net debt (factor in tax shield benefit at effective tax rate of 28%), Intel's market capitalization will be around $86.7 billion or $17.21 per share. We note that the assumed terminal growth rate of 3% is less than the trend line US nominal GDP growth rate. This scenario pretty much says that Intel will make little progress in mobile market while maintaining the same level of Capex as before. And after a long period of stagnation, the growth of PC and server market recovers to a rate below GDP growth. If such things happen, the downside risk of the stock is about 14% from the recent trade price of $20. In a more bullish scenario, we increase the seven years' sales growth to 2% and terminal growth rate to 5% but keep other assumptions intact. The share value rises to $21.81. This second scenario factors in some strength of the traditional PC market but barely prices in any upside from mobile processors, whether it is from tablet or hybrid tablet-laptop. Intel's average revenue growth in the past seven years was 6.7%. If the company successfully becomes one of the major players in the mobile chip market, the top line growth rate shall match that figure and share value will be much higher.

Conclusion

Intel owns a valuable technology platform in the PC and server market. This market is undergoing a secular transformation and so is Intel. The company still possesses a strong franchise and deep technology know-how to reinvent itself and adapt to the new demand. Intel's share price may continue to be under pressure for a while on media hypo or short-term earnings disappointment during its transition period. But at the current price, investors can get Intel's existing business with a reasonable margin of safety plus a free option if Intel successfully captures a piece of the mobile device market. The company's high-quality dividend stream also provides a cushion to enable long-term value investors to sustain the short term market fluctuation.

Bearish case

Equity valuation, $million

FYE2012

FYE2013

FYE2014

FYE2015

FYE2016

FYE2017

FYE2018

FYE2019

Revenue

53,999

53,999

53,999

53,999

53,999

53,999

53,999

53,999

yoy

0.0%

0.0%

0.0%

0.0%

0.0%

0.0%

0.0%

Gross Profit

30,779

30,779

30,779

30,779

30,779

30,779

30,779

30,779

gross margin

57%

57%

57%

57%

57%

57%

57%

57%

Operating Profit

13,230

13,230

13,230

13,230

13,230

13,230

13,230

13,230

operating margin

25%

25%

25%

25%

25%

25%

25%

25%

- Tax provision

3,704

3,704

3,704

3,704

3,704

3,704

3,704

3,704

tax rate

28%

28%

28%

28%

28%

28%

28%

28%

NOPAT

9,525

9,525

9,525

9,525

9,525

9,525

9,525

9,525

+ Depr & Amort

7,830

8,532

9,093

9,543

9,902

10,190

10,420

10,604

- Capex & Acq.

11,340

11,340

11,340

11,340

11,340

11,340

11,340

11,340

Capex % Sales

21%

21%

21%

21%

21%

21%

21%

21%

- Incr of net wc

1,048

0

0

0

0

0

0

0

Free Cash Flows

4,967

6,717

7,279

7,728

8,088

8,375

8,605

8,789

PV of FCFs

37,576

discount rate

11.0%

PV of terminal value

49,104

terminal FCF multiple

12.9

Enterprise value

86,680

- Net debt & liabilities

1,028

Market value

85,652

shares outstanding, million

4,976

Price per share

17.21

source: company financial statements and our forecast

Base case

Equity valuation, $million

FYE2012

FYE2013

FYE2014

FYE2015

FYE2016

FYE2017

FYE2018

FYE2019

Revenue

55,079

56,181

57,304

58,450

59,619

60,812

62,028

63,268

yoy

2.0%

2.0%

2.0%

2.0%

2.0%

2.0%

2.0%

Gross Profit

31,395

32,023

32,663

33,317

33,983

34,663

35,356

36,063

gross margin

57%

57%

57%

57%

57%

57%

57%

57%

Operating Profit

13,494

13,764

14,040

14,320

14,607

14,899

15,197

15,501

operating margin

25%

25%

25%

25%

25%

25%

25%

25%

- Tax provision

3,778

3,854

3,931

4,010

4,090

4,172

4,255

4,340

tax rate

28%

28%

28%

28%

28%

28%

28%

28%

NOPAT

9,716

9,910

10,108

10,311

10,517

10,727

10,942

11,161

+ Depr & Amort

7,830

8,577

9,221

9,784

10,282

10,730

11,138

11,515

- Capex & Acq.

11,567

11,798

12,034

12,275

12,520

12,770

13,026

13,286

Capex % Sales

21%

21%

21%

21%

21%

21%

21%

21%

- Incr of net wc

1,054

6

6

6

7

7

7

7

Free Cash Flows

4,925

6,683

7,289

7,813

8,272

8,680

9,047

9,383

PV of FCFs

38,317

discount rate

11.0%

PV of terminal value

71,249

terminal FCF multiple

17.5

Enterprise value

109,565

- Net debt & liabilities

1,025

Market value

108,540

shares outstanding, million

4,976

Price per share

21.81

source: company financial statements and our forecast

Source: Intel: Low Valuation And A Free Option