ab
UBS Investment Research NetLogic Microsystems Inc
Excellence in Search: Solid Fundamentals With Further Upside; Initiate With Buy
Strong position drives a sustainable model with valuation upside NetLogic has built a strong competitive position in its core search processing chip business, with capabilities that make the entry barrier very high. Further enhancing its model have been acquisitions for high-end embedded processors (RMI) and high-speed interface chips. From this position and a design win trajectory that we believe drives sustained revenue growth, a primary valuation determinant, we derive a $51 price target, and given the upside, we rate the stock Buy. Positives – Solid sales drivers, expense control to expand op margin, FCF We model a 13% sales CAGR (2010-13) for NetLogic based on its search processor design win pipeline and expectations for RMI sales to accelerate on 40nm processor growth. We expect operating margin expansion on steady 65-67% gross margin and opex to grow at 60% the rate of sales (exceeding 25% nonGAAP in 2012). We expect FCF to remain positive and 20%+ FCF margin in ‘11. Risks – End-customer and end-market dependence We view NetLogic’s dependence on Cisco and ultimately the communications infrastructure market as the weakest links in the NetLogic thesis. However, key to mitigating these risks is the increasing market need for NetLogic technology along with the design win diversification, outside of Cisco and increasingly within Cisco. Valuation: Initiate coverage with a $51 price target, Buy rating Our 12-month price target of $51 (30x our $1.70 2010 non-GAAP EPS estimate) is based on our DCF model assumptions: 11.0% 10-year revenue CAGR, 25% longterm operating margin, 10.1% WACC, and 2% terminal growth.
Global Equity Research
Americas Semiconductors 12-month rating 12m price target Price
RIC: NETL.O BBG: NETL US
Buy
Prior: Not Rated
US$51.00 US$43.76
22 September 2009
Trading data 52-wk range Market cap. Shares o/s Free float Avg. daily volume ('000) Avg. daily value (US$m) Balance sheet data 12/09E Shareholders' equity P/BV (UBS) Net Cash (debt) Forecast returns Forecast price appreciation Forecast dividend yield Forecast stock return Market return assumption Forecast excess return EPS (UBS, US$) From 12/09E To (0.18) (0.10) (0.12) (0.72) (1.25) (1.84) Cons. 0.29 0.35 0.32 0.35 1.33 1.65 12/08 Actual 0.05 0.10 0.06 (0.05) +16.5% 0.0% +16.5% 6.0% +10.5% US$0.40bn 2.6x US$0.01bn US$44.69-15.19 US$0.96bn 22.0m (ORD) 82% 378 14.7
Highlights (US$m) Revenues EBIT (UBS) Net Income (UBS) EPS (UBS, US$) Net DPS (UBS, US$) Profitability & Valuation EBIT margin % ROIC (EBIT) % EV/EBITDA (core) x PE (UBS) x Net dividend yield %
12/07 109 (1) 3 0.14 0.00 5-yr hist av. -38.9 (6.2) -
12/08 140 0 4 0.16 0.00 12/08 0.1 0.1 29.8 NM 0.0
12/09E 168 (27) (30) (1.25) 0.00 12/09E -16.0 (10.9) >100 NM 0.0
12/10E 291 (53) (56) (1.84) 0.00 12/10E -18.3 (14.7) 77.7 NM 0.0
12/11E 338 (27) (26) (0.82) 0.00 12/11E -8.0 (8.6) 27.7 NM 0.0
Q1 Q2 Q3E Q4E 12/09E 12/10E
Performance (US$)
60.0 50.0 40.0 30.0 20.0 10.0 07/06 10/06 01/07 04/07 07/07 10/07 01/08 04/08 07/08 10/08 01/09 04/09 07/09 0.0 50 0 Stock Price (US$) Rel. S & P 500 200 150 100
Source: Company accounts, Thomson Financial, UBS estimates. (UBS) valuations are stated before goodwill-related charges and other adjustments for abnormal and economic items at the analysts' judgement. Valuations: based on an average share price that year, (E): based on a share price of US$43.76 on 21 Sep 2009 19:39 EDT
Price Target (US$) (LHS) Rel. S & P 500 (RHS)
Stock Price (US$) (LHS)
Source: UBS
Steven Eliscu
Analyst steven.eliscu@ubs.com +1-415-352 5674
Uche Orji
Analyst uche.orji@ubs.com +1 212 713 4015
www.ubs.com/investmentresearch
This report has been prepared by UBS Securities LLC ANALYST CERTIFICATION AND REQUIRED DISCLOSURES BEGIN ON PAGE 45. UBS does and seeks to do business with companies covered in its research reports. As a result, investors should be aware that the firm may have a conflict of interest that could affect the objectivity of this report. Investors should consider this report as only a single factor in making their investment decision.
NetLogic Microsystems Inc 22 September 2009
Contents
Investment Thesis
—
page 3 4
Steven Eliscu
Analyst steven.eliscu@ubs.com +1-415-352 5674
Key Position in Growth Market .............................................................................3 1) Strong Competitive Position in Search Processing ...........................................4 2) Market Trends Driving Need for Search and High-End Embedded Processing.5 3) Well-managed Expense Control.......................................................................6 4) Strong Cash Flow ............................................................................................7 5) Reduced Volatility Model..................................................................................7
Positives
— — — — —
Uche Orji
Analyst uche.orji@ubs.com +1 212 713 4015
Sensitivity Analysis
—
8 9
DCF Valuation Drives $51 Price Target................................................................8 1) Still Highly Dependent on Cisco .......................................................................9 2) Cisco Is NetLogic’s Largest Search Processing “Competitor”.........................10 3) Dependent on Comm Infrastructure Market....................................................10 4) RMI Integration ..............................................................................................10 5) Difficult to Maintain Leadership in High-End Embedded Processing...............10 6) Long New Product Gestation Period ..............................................................11
UBS vs Consensus Expectations
— — — — — —
Forthcoming Catalysts Valuation Stake Ownership Company Overview
—
11 14 15 16 20 21 23 23 26 27
Products.............................................................................................................17
NetLogic Management Earnings and Estimate Summary
—
Revenue—EPS Sensitivity Analysis ...................................................................22
Share Price Drivers Revenue Drivers Gross/Operating Margin Drivers Cash Flow Drivers
— —
Inventory Analysis ..............................................................................................27 Consideration of Stock-based Comp Expenses..................................................28
Appendix NSE/KBP Overview
— — — —
29 30
TCAM Is Underlying Technology........................................................................30 Cisco Enabled TCAM Market .............................................................................30 Cisco’s Gen 3 TCAM Enabled NetLogic’s Growth ..............................................31 Algorithmic Approach to Search Processing.......................................................31
Processor Overview PLP Overview 7-Layer OSI Model
— — — — — — — —
33 35 37
Various Layers Result in Specialized Communications Infrastructure.................37 Layer 1—Physical Layer; Muxing Equipment .....................................................37 Layer 2—Datalink Layer; Switching Equipment ..................................................38 Layer 3—Network Layer; Routing Equipment.....................................................38 Layer 4—Transport Layer ..................................................................................39 Layer 5—Session Layer; VoIP Equipment..........................................................39 Layer 6/7—Presentation/Application Layer.........................................................39 NetLogic Products Encompass all 7 Layers........................................................39
UBS 2
NetLogic Microsystems Inc 22 September 2009
Investment Thesis
Key Position in Growth Market
The bane of tech stocks is competition and the strength of a company’s competitive position is often the litmus test of our conviction level. NetLogic Microsystems has an enviable position in hardware-based search processing, as its technology depends on an amalgam of expertise that is rare in the semiconductor industry – high-speed memory, high-speed I/O and intimate knowledge of Internet Protocol (IP) based network architectures. Search processing is the necessary function that enables IP network traffic to flow from its source to its destination – NetLogic search processors are targeted into high-end, performance-sensitive networking equipment. The company brands its search processors as Knowledge-based Processors (KBPs), which utilize the knowledge in the network to perform search and more. We model a 13% CAGR (2010-13) for NetLogic’s search processor revenue. NetLogic’s recently announced acquisition of RMI (for $183m + a maximum of $68m potential earn-outs), a high-end embedded processor company, adds a new dimension to its focus on the networking equipment market. While suppliers for high-end embedded processing have had a history of passing the baton to successive incumbents (e.g., Broadcom, PMC-Sierra, Freescale), we believe a dearth of new entrants will likely limit NetLogic’s competition to Cavium and Freescale. However, the RMI acquisition makes NetLogic’s model more uncertain: 1) potentially higher opex intensity, as it battles a highly competent competitor, Cavium, which is currently the incumbent for recent high-end designs at Cisco; and 2) it adds a new sales channel through which its sells RMI’s low-end Alchemy processors. We expect a 15% CAGR (2010-13) for NetLogic’s RMI processor revenue. Along with limited competition, NetLogic has a history of sound execution while controlling expenses. Key to NetLogic’s ability to control R&D is a high degree of intellectual property (IP) reuse for new product development as well as a limited customer base that can be largely serviced on a direct basis (excluding the newly adopted Alchemy processor customer base). No less important has the been the ability for the company to build deep OEM relationships, led by Cisco Systems, which have allowed it to gain intimate knowledge of customers’ future requirements, enhancing the sustainability of its model, which we believe translates to an enhanced valuation. We expect operating expenses to grow at 60% the rate of sales growth, supporting our long-term 25% GAAP operating margin expectation. On a non-GAAP basis, we expect operating margin to reach 26% in 2012. Our 12-month price target valuation based on our DCF model is $51, which is supported by our analysis. Given the upside to our price target, we initiate coverage with a Buy rating.
NETL – Buy - $51 price target
UBS 3
NetLogic Microsystems Inc 22 September 2009
Positives
NetLogic’s strong competitive position, increasing end-market requirement for its technology and solid expense control provide strength and resilience to its model that we believe can result in solid free cash flow margins:
1) Strong Competitive Position in Search Processing
NetLogic currently has no direct merchant semiconductor competitors in search processors, as Cypress Semiconductor and Integrated Device Technology have exited the market, with the assets having been purchased by NetLogic. Additionally, Renesas is essentially a captive supplier of Cisco. (In separate transactions, NetLogic purchased much of Cypress’s search processing portfolio in February 2006 for $58.5m in NetLogic shares and the remainder in August 2007 for $14.4m in cash, and it purchased IDT’s portfolio in May 2009 for $90m in cash.) We do not believe NetLogic faces any new merchant semiconductor competition near-term given the significant barriers to entry: Required skill set combination – very few companies possess the capabilities of being able to design chips with large, advanced memories, high speed logic and high-speed I/O in the most advanced process nodes – this exclusive club includes Intel and AMD (advanced server chips) as well as Altera and Xilinx (high-end FPGAs); Broadcom and Marvell also have some of these capabilities including high-speed I/Os but do not design large, handcrafted high-speed memories and have been increasingly focusing their R&D efforts on the wireless device market; where they have needed NetLogic’s technology, they have chosen to partner rather than compete. Traditional high-speed memory companies such as Cypress Semiconductor and Integrated Device Technology struggled to execute increasingly advanced search processors on constrained budgets and exited the market. Ironically, while memory as a standalone product is a commodity, in the context of being integrated into advanced processors, it becomes a barrier to entry, especially given the limited population of skilled memory designers. Furthermore, as NetLogic combines two types of memory technology in its “hybrid” search processors, its primary implementation, it has further raised the bar. Of equal importance, we view NetLogic’s high-speed I/O technology (see Appendix) from its Aeluros acquisition as another important barrier to new entrants given the difficulty of executing this technology. Thus, we believe it is unlikely a competitor will emerge from the mainstream merchant semiconductor supplier base. Concerning M&A, we believe it is possible that NetLogic could become a target for Broadcom or Marvell to increase penetration into Cisco (as Cisco has increasingly relied on captive silicon to maintain differentiation). We think an acquisition from Intel is unlikely, as Intel is focused on penetrating its x86 architecture into communications, while the FPGA names, Altera and Xilinx, historically have made only small acquisitions and would likely seek much smaller targets to bolster their positions for the next generation of 28nm FPGAs.
Ironically, while memory as a standalone product is typically as a commodity, in the context of advanced processors, it becomes a barrier to entry, especially given the limited number of skilled memory designers
We believe it is unlikely a search processor competitor will emerge from the mainstream merchant semiconductor supplier base
UBS 4
NetLogic Microsystems Inc 22 September 2009
Time frame and cost to develop a competing mature operating model – The ability to successfully execute large complex chips with a lot of highspeed memory in advanced process nodes takes years to refine. If an engineering team started from scratch, even if it had all of the necessary intellectual property at hand, we estimate it would still need to go through a 5-year learning process (i.e., ~3 design generations) to execute such highly complex chips. If anything, we are concerned that the lack of direct competition could result in slower market growth, as OEMs may be hesitant to use NetLogic in all but the most performance sensitive applications. However, we believe no commercially viable company would put itself at a competitive disadvantage for its leading edge products simply to ensure multiple sourcing of components. Somewhat more concerning as a limiter of growth is Cisco’s in-house search processing team based on its acquisition of Spans Logic, which we discuss below in the Risks section.
The ability to successfully execute large complex chips with a lot of highspeed memory in advanced process nodes takes years to refine
Even as NetLogic has no merchant search processor competition, we believe no commercially viable OEM would put itself at a competitive disadvantage and avoid using NetLogic products simply to ensure multiple sourcing of components
2) Market Trends Driving Need for Search and High-End Embedded Processing
We believe NetLogic can sustain growth based on OEMs requiring the combination of increasing performance and more complex functionality for their high-end IP networking equipment. Performance is typically measured by the amount of data per second communicating between network nodes (e.g., 10gigabits per second). Over the next several years, high-end network interfaces will be migrating to 40 Gbps and 100 Gbps interfaces, up to a 10x increase over today’s fastest data rates. In addition, as video and voice traffic become more common, it will be required for telecom carriers to differentiate among services to ensure quality of service (QoS) at the lowest cost. Despite political winds that may require so-called Net Neutrality, where carriers cannot selectively discriminate among traffic types, differentiated treatment of traffic will likely exist in some form, especially on wireless networks, in order to ensure bandwidth availability. Wireless carriers are currently using the crude method for limiting network bottlenecks by placing monthly caps on the maximum amount of data (e.g., 5 gigabytes) that can be downloaded without the user incurring surcharges. This is a less than ideal solution, as it defeats the purpose of multimedia smartphones and notebooks/netbooks for usage other than downloading e-mail, which can be accomplished on a standard business Blackberry. NetLogic search processors can be used to determine data types and feed into so-called traffic managers, which are the functions within switches/routers that can “shape” traffic patterns to enable networks to most optimally deliver combined voice, video and data. In addition to being able to differentiate among various services, we believe network security will remain an important issue as the number and sophistication of users grow, including those who use networking technology for nefarious purposes. NetLogic’s NETL7 products paired with RMI processors can be used by equipment makers to enable carriers to stay ahead of these evolving threats by being able to search through the details of the data traffic in real-time to determine its content (e.g., if it is harboring a virus).
NetLogic’s NETL7 products paired with RMI processors can be used to stay ahead of evolving security threats NetLogic search processors enable telecom carriers to differentiate among services to ensure quality of service (QoS) at the lowest cost
UBS 5
NetLogic Microsystems Inc 22 September 2009
Because performance and improved traffic analysis are simultaneously increasing requirements, we believe NetLogic’s model is highly sustainable as long as it can continue to execute new products and win new designs. As Cisco is likely to retain a dominant position in many of its markets, this sustainability is still contingent upon maintaining a strong search processing position at Cisco and winning new designs on RMI processors. Complicating the RMI design win opportunity at Cisco is the potential mismatch between the latest XLP RMI processor, which is a highly threaded architecture (8-core, 4 threads/core), while Cisco’s Internet Operating System (IOS), its cross-platform software environment, contains a lot of code that was written before advanced multicore/multithreaded processors became available. To better address’s Cisco’s needs, RMI’s XL-series processor can operate in a nonthreaded mode that is better suited to run this legacy code. Thus, assuming RMI delivers XLP samples to Cisco by 1Q10, we believe it is possible to secure a Cisco design win before the end of 2010. Additionally, as RMI expands its XLP family to include lower core-count versions (including single-core), RMI has the opportunity to proliferate its technology beyond just very high-end applications.
3) Well-managed Expense Control
NetLogic has a culture for tightly managing operating expenses. From an R&D point of view, this is reflected in: 1) a high degree of intellectual property reuse that maximizes R&D productivity; 2) “right the first time” design methodology that avoids costly retooling charges; and 3) managing tradeoffs between features and customer requirements to minimize risk, time-to-market and product costs. From an SG&A perspective, the company has been focused on servicing its key accounts with primarily a direct sale force and operates its headquarters with a single administrative assistant. This expense control is key to our expectation for the company to reach a longterm GAAP operating margin of 25% based on our expectation for opex to grow at about 60% the rate of sales growth, especially as we expect stock-based compensation as a percentage of sales to decline as the company focuses on its goal to lower dilution to 3-3.5% of shares outstanding in new grants.
Chart 1: NetLogic Operating Margin, Y/Y Opex/Y/Y Revenues
70% 60% Y/Y Opex/Y/Y Revenues 50% 40% 30% 20% 10% 0% 2010E 2011E 2012E 2013E 2014E 2015E 2016E 2017E 2018E 2019E -15% -25% 25% 15% 5% -5% Operating Margin, GAAP
Y/Y Opex /Y/Y Rev enues
Source: UBS estimates
Operating Margin, GAAP
UBS 6
NetLogic Microsystems Inc 22 September 2009
With the acquisition of RMI, we expect it will take some time for NetLogic to optimally rationalize RMI opex as it integrates a new design team with its own and adopts a new sales force. At some time in the future, we expect NetLogic and RMI to relocate into a single building to best capitalize on potential operational synergies.
4) Strong Cash Flow
While GAAP earnings losses could extend until 1Q12, we expect NetLogic’s free cash flow to remain positive and sustain a 20% free cash flow margin from 2011, a level it has achieved historically. We expect annual free cash flow to exceed net income by as much as $95m even through 2013. In our valuation analysis, we show that companies with free cash flow margins of about 20% trade at nearly 4x sales ($338m in 2011), further validating our $51 12-month price target.
5) Reduced Volatility Model
We believe NetLogic’s valuation is supported from what we believe is inherently a reduced volatility model. One way to show this is to use annual price volatility data, which is a measure of a stock’s average annual price movement to a high and low from a mean price for the past year. NetLogic’s price volatility of 42% indicates the stock’s annual high and low price has shown a variation of +42% to -42% from its annual average price. This is in line with our peer group average and below larger cap names including SanDisk, AMD, ON Semiconductor and NVIDIA. While seemingly susceptible to lumpy sales patterns given the nature of carrier and data center spending, NetLogic’s model benefits from the broad customer and end-market diversity that comes with the near universal IP networking OEM requirement for its technology. This diversity exists not just outside of Cisco, but even increasingly within Cisco, where the number of design wins has doubled over the past two years. Also supporting lower volatility is a model where the company’s quarterly sales guidance is fully covered by orders shipped-to-date and backlog coverage. In addition, by the time the earnings conference call occurs, which is about one month into the current quarter, NetLogic typically has shipped about half of its sales guidance while its 12-week lead-time fully covers the balance of the quarter. NetLogic has the capability to accommodate some “turns” (i.e., book and ship in the same quarter) business, but this number is typically small, thus resulting in few surprises. Given this high certainty of sales and very limited COGS leverage, the potential for gross margin surprises is also limited.
NetLogic’s model benefits from the broad diversity and thus lower volatility that comes with the near universal OEM requirement for its technology Table 1: Annual Price Volatility (last 12 months)
Price Volatility SIGM SNDK AMD CY ONNN NVDA DIOD MRVL ATML BRCM SWKS NETL MU IRF ISIL FCS TXN ADI MXIM ALTR INTC Average Source: FactSet 59.9% 53.3% 52.6% 52.4% 52.3% 52.2% 46.6% 45.4% 43.6% 43.6% 43.5% 42.5% 41.2% 40.7% 40.2% 39.9% 32.6% 30.9% 30.3% 30.2% 29.8% 43.0%
UBS 7
NetLogic Microsystems Inc 22 September 2009
Sensitivity Analysis
DCF Valuation Drives $51 Price Target
Our DCF analysis leads to a 12-month price target of $51, equivalent to 30x our 2010 $1.70 non-GAAP EPS estimate. We assume an 11.0% 10-year revenue CAGR, 25% long-term operating margin, 10.1% WACC and 2% terminal growth rate. Table 3 provides a sensitivity overview, showing potential upside and downside to our base-case scenario.
Table 3: NetLogic DCF Sensitivity Analysis Table 2: NetLogic DCF
Key Assumptions 10-yr CAGR LT EBIT LT Capex/Sales WACC Terminal Growth
Source: UBS estimates
11.0% 25.0% 0.8% 10.1% 2.0%
Sensitivity Analysis Terminal Growth\WACC
51.00 0.0% 1.0% 1.5% 2.0% 2.5% 3.0% 3.5% 4.0% 4.5% 5.0% 5.5% 51.00 20% 21% 22% 23% 24% 25% 26% 27% 28% 29% 30% 51.00 20% 21% 22% 23% 24% 25% 26% 27% 28% 29% 30% 9.1% $50.78 $54.62 $56.92 $59.55 $62.57 $66.08 $70.21 $75.15 $81.17 $88.63 $98.16 9.6% $47.53 $50.82 $52.77 $54.98 $57.50 $60.39 $63.76 $67.73 $72.47 $78.24 $85.41
10.1% $44.62 $47.46 $49.12 $51.00 $53.12 $55.53 $58.31 $61.55 $65.36 $69.91 $75.45 10.1% $44.35 $45.68 $47.01 $48.34 $49.67 $51.00 $52.33 $53.65 $54.98 $56.31 $57.64 2.0% $44.35 $45.68 $47.01 $48.34 $49.67 $51.00 $52.33 $53.65 $54.98 $56.31 $57.64
10.6% $42.00 $44.46 $45.90 $47.50 $49.30 $51.33 $53.65 $56.32 $59.43 $63.08 $67.45 10.6% $41.47 $42.67 $43.88 $45.09 $46.29 $47.50 $48.71 $49.91 $51.12 $52.33 $53.53 2.5% $46.07 $47.48 $48.89 $50.30 $51.71 $53.12 $54.53 $55.94 $57.35 $58.75 $60.16
11.1% $39.64 $41.79 $43.03 $44.40 $45.94 $47.67 $49.62 $51.85 $54.41 $57.39 $60.91 11.1% $38.90 $40.00 $41.10 $42.20 $43.30 $44.40 $45.50 $46.60 $47.70 $48.80 $49.90 3.0% $48.02 $49.52 $51.03 $52.53 $54.03 $55.53 $57.03 $58.54 $60.04 $61.54 $63.04
Implied P/E F2010, Non-GAAP
9.1% 29.9x 32.2x 33.5x 35.1x 36.8x 38.9x 41.4x 44.3x 47.8x 52.2x 57.8x 9.6% 28.0x 29.9x 31.1x 32.4x 33.9x 35.6x 37.6x 39.9x 42.7x 46.1x 50.3x 10.1% 26.3x 28.0x 28.9x 30.0x 31.3x 32.7x 34.3x 36.2x 38.5x 41.2x 44.4x 10.1% 26.1x 26.9x 27.7x 28.5x 29.3x 30.0x 30.8x 31.6x 32.4x 33.2x 33.9x 2.0% 26.1x 26.9x 27.7x 28.5x 29.3x 30.0x 30.8x 31.6x 32.4x 33.2x 33.9x 10.6% 24.7x 26.2x 27.0x 28.0x 29.0x 30.2x 31.6x 33.2x 35.0x 37.2x 39.7x 10.6% 24.4x 25.1x 25.8x 26.6x 27.3x 28.0x 28.7x 29.4x 30.1x 30.8x 31.5x 2.5% 27.1x 28.0x 28.8x 29.6x 30.5x 31.3x 32.1x 32.9x 33.8x 34.6x 35.4x
11.1% 23.3x 24.6x 25.3x 26.2x 27.1x 28.1x 29.2x 30.5x 32.0x 33.8x 35.9x 11.1% 22.9x 23.6x 24.2x 24.9x 25.5x 26.2x 26.8x 27.4x 28.1x 28.7x 29.4x 3.0% 28.3x 29.2x 30.1x 30.9x 31.8x 32.7x 33.6x 34.5x 35.4x 36.2x 37.1x
Terminal EBIT %\WACC
9.1% $51.39 $53.03 $54.66 $56.29 $57.92 $59.55 $61.18 $62.81 $64.44 $66.07 $67.70 1.0% $41.49 $42.68 $43.88 $45.07 $46.26 $47.46 $48.65 $49.84 $51.04 $52.23 $53.42 9.6% $47.64 $49.10 $50.57 $52.04 $53.51 $54.98 $56.45 $57.92 $59.39 $60.85 $62.32 1.5% $42.84 $44.10 $45.35 $46.61 $47.87 $49.12 $50.38 $51.64 $52.90 $54.15 $55.41
Implied P/E F2010, Non-GAAP
9.1% 30.3x 31.2x 32.2x 33.2x 34.1x 35.1x 36.0x 37.0x 38.0x 38.9x 39.9x 1.0% 24.4x 25.1x 25.8x 26.5x 27.2x 28.0x 28.7x 29.4x 30.1x 30.8x 31.5x 9.6% 28.1x 28.9x 29.8x 30.7x 31.5x 32.4x 33.2x 34.1x 35.0x 35.8x 36.7x 1.5% 25.2x 26.0x 26.7x 27.5x 28.2x 28.9x 29.7x 30.4x 31.2x 31.9x 32.6x
Terminal EBIT %\Terminal Growth
Implied P/E F2010, Non-GAAP
Source: Company reports and UBS estimates
UBS 8
NetLogic Microsystems Inc 22 September 2009
UBS vs Consensus Expectations
Our 2010 sales/EPS estimates of $291m/$1.70 non-GAAP exceed company guidance of $275m/$1.60 non-GAAP and consensus at $278m/$1.65 on expectations for more meaningful search processor revenue growth, given NetLogic’s strong design win pipeline as well as a successful integration of RMI.
Table 4: NetLogic Microsystems – Estimate Summary vs Consensus
Revenue UBS F3Q09E F4Q09E FY2009E FY2010E FY2011E $39.3m $65.6m $167.8m $291.1m $338.4m Consensus $45.0m $64.0m $171.8m $277.6m $319.2m Difference -13% +3% -2% +5% +6% UBS -$0.12 -$0.72 -$1.25 -$1.84 -$0.82 UBS nonGAAP $0.32 $0.30 $1.25 $1.70 $2.37 EPS Consensus $0.32 $0.35 $1.33 $1.65 $2.07 Difference +0% -14% -6% +3% +14% Gross Margin 65.0% 64.8% 65.6% 65.2% 66.0% Op Margin -5.0% -27.9% -16.0% -18.3% -8.0% Op Profit -$2.0m -$18.3m -$26.8m -$53.3m -$27.1m
Source: FactSet and UBS estimates
Risk Analysis
While we believe NetLogic has a compelling story around search processing, there are still material risks:
1) Still Highly Dependent on Cisco
Cisco gave birth to the “network layer” (i.e., Layer 3, see Appendix for more details) search processor market, and given its dominant 63% market share in Ethernet Switching and 54% share in Routing (2Q09, Dell’Oro), it is no surprise that Cisco drives this market. In fact, it was Integrated Device Technology’s decision to pass on the opportunity to build the Gen 3 (also called TCAM3) search processor for Cisco that gave NetLogic the opening to establish a toehold in the market. Thus, viability as a search processing vendor has been mainly predicated on success at Cisco. Even as NetLogic diversifies its customer base, its proliferation of design wins at Cisco makes it likely that Cisco will remain the largest customer for the foreseeable future. However, Cisco will likely continue to utilize Renesas as an effective captive source for much of its volume requirements in the mid-range and its own internal capability at the low-end, both of which could limit growth and result in Cisco continuing to represent a smaller percentage of NetLogic’s search processing revenue relative to Cisco’s switch/router market share. As an indicator that this trend could continue, we note that even as the number of cumulative Cisco design wins has doubled over the past two years, the number is nearly 2.5x for all other search processor customers (although those many of those other design wins are likely to be smaller in size vs Cisco).
UBS 9
NetLogic Microsystems Inc 22 September 2009
2) Cisco Is NetLogic’s Largest Search Processing “Competitor”
Cisco is NetLogic’s largest customer. Given the exit of NetLogic’s major merchant competitors, Cisco through its Spans Logic acquisition and Renesas, which is building TCAM products for only Cisco, are NetLogic’s primary competitors. However, we do not believe Cisco’s in-house capability represents a substantive longer-term threat because of its inability to support 10-gigabit per second data rates (let alone the emerging 40/100 Gbps requirement). We expect Cisco to use both Spans Logic and Renesas in more cost-sensitive enterprise switch/router applications. We are not aware of any merchant solutions on the horizon that could challenge NetLogic.
3) Dependent on Comm Infrastructure Market
Even as NetLogic diversifies from Cisco while expanding the number of designs it has at Cisco, it is ultimately beholden to the slow growth of the Communications Infrastructure market. For 2005-10, UBS forecasts just a 3.2% CAGR for global carrier capital expenditures. Thus, with the trend towards higher speeds (10-gigabit per second interfaces moving to 40 Gbps and 100 Gbps over the next several years) and more intelligence in the network (which is used to maximize network efficiency and security), growth of the overall market is likely to be limited, especially given the depth of the global downturn.
4) RMI Integration
While NetLogic’s other acquisitions have been small teams or largely for gaining legacy product revenue and intellectual property, its acquisition of RMI is significant, as RMI’s revenue base represents nearly half of the former NetLogic. As positives, we expect RMI to remain fairly autonomous, and the two companies have a history of working together on higher end products, particularly the pairing of NetLogic’s NETL7 with RMI’s XL-series processor, which logically could be combined onto a single piece of silicon in the future. However, the integration is not without risk. It is possible that as the company focuses its resources on execution of its upcoming flagship XLP processor, the effort on its lower-end Alchemy family could be diluted. At least initially, we do not expect that to happen, especially as Alchemy opens the door for NetLogic to enter more consumer-oriented markets. We estimate Alchemy processor revenue comprises a little less than half of RMI’s total revenue, so it remains important.
5) Difficult to Maintain Leadership in High-End Embedded Processing
The high-end embedded processing market has had a history of successive incumbents (e.g., Broadcom, PMC-Sierra, Freescale) and unsuccessful challengers (e.g., AMD, PA Semiconductor, IDT, Exponential Technology/ Intrinsity). Because of the rising chip development costs, limited market size ($0.3bn in 2008) and execution track record of the remaining vendors, we believe the high-end embedded processor market is likely to evolve into a two or three horse race, where NetLogic would face-off against Cavium while Freescale could compete on lower-performance applications. With NetLogic’s advanced chip design capabilities and position as a strategic supplier to Cisco, we believe NetLogic brings likely revenue synergies with RMI, which should provide for more sustainability as a high-end embedded processing player.
Process technology and advanced circuit design are two primary requirements for high-end embedded processor leadership, which with the addition of NetLogic and Aeluros, RMI is more likely to improve on and sustain
UBS 10
NetLogic Microsystems Inc 22 September 2009
6) Long New Product Gestation Period
While the company has the benefit of long production life cycles that create a high degree sustainability of revenues, it also takes about 5 years between new product definition and resulting high-volume production. Thus, the attractiveness of NetLogic as an investment idea rests largely on the products it has already developed or has in the pipeline. On balance, while the die is cast for NetLogic’s near-to-mid-term revenue streams, long gestation periods reflect the very long total product life cycle that can easily span more than a decade. We believe this gives greater predictability of future product revenue streams based on its design win trajectory and lowers the volatility of NetLogic’s model.
Forthcoming Catalysts
We believe two primary near-to-mid-term catalysts will be key drivers for the stock and could drive higher long-term valuation: (1) Return of double digit revenue growth and resulting operating margin leverage – In 4Q09, we expect NetLogic to return to organic double-digit y/y revenue growth, given an expected economic recovery on weak comps. We believe this growth is sustainable based on expectations for: 1) growth of its Cisco business, which has been a flatliner over the past 3 years as the company grows revenue in new designs with limited or no competition and takes previously unserved TCAM3 business from Renesas; 2) growth of Huawei 3G wireless and IP video applications – NetLogic should continue to be a direct beneficiary of China’s wireless and IPTV buildouts; 3) growth of Alcatel-Lucent on the proliferation of NetLogic products, including its successful TiMetra 7750 Carrier Services Router used for both core wired and wireless infrastructure applications; and 4) growth of wireless applications, including new product ramps at Ericsson/Redback, Nokia-Siemens Networks (NSN) and ZTE. (a) Cisco growing again – Having reached a trough in 1Q09, we expect sales into Cisco to rise to near 2008’s yearly run-rate ($53m, excluding the recently acquired search processing business from IDT) in 2H09 on the ramp of new designs, especially the “Gen 4 TCAM.” or NL8000-family, which is used in high-end switching and routing platforms. (TCAM refers to a Ternary Content Addressable Memory, the technical term for the primary memory technology used in a search processing device – please see the Appendix for more details.) NetLogic faces competition from Renesas, which serves about half of Cisco’s merchant search processing requirements. In the future, because of the stringent technical requirements, NetLogic could be the sole-source for Cisco’s “Gen 5” (i.e., fifth generation) search processor and an important supplier of Physical Layer products (PLPs) with the transition to SFP+ optical interfaces (please see the Appendix for more details). Cisco also has its own lower performance search processing technology gained from its 2007 Spans Logic acquisition used in more cost-sensitive applications. Given an estimated order of magnitude lower performance of Spans Logic vs NetLogic’s current generation
UBS 11
NetLogic Microsystems Inc 22 September 2009
products, we would expect Cisco, at best, to compete with what are now low-end multi-sourced products that are not expected to drive much of NetLogic’s growth. (b) Design wins moving to production – Because of the long life cycle of communications infrastructure products, we believe cumulative design wins are a good proxy for total active design wins. As the portfolio of search processor cumulative design wins grows further (surpassed 275 in 2Q09), we expect NetLogic to grow revenue consistently despite the lumpy nature of its end markets. For example, as growth of China’s 3G wireless infrastructure products could slow in 2H09, this revenue should be augmented by new revenue from Cisco, Brocade/Foundry Networks, lower-end NETLite-based products mainly used in enterprise networking equipment and IPTV-related infrastructure. Additionally, we expect more recent wins to result in production revenue including Ericsson/Redback in late 2009 as well as NSN and ZTE in 2010.
Chart 2: Dependence of Sales From Cisco Has Been Declining as Cumulative Search Processor Design Wins Have Been Rising
300 Cumulative SPU Design Wins 250 200 150 100 50 0 Sep-04 Sep-05 Sep-06 Sep-07 Sep-08 Mar-04 Mar-05 Mar-06 Mar-07 Mar-08 Mar-09 100% 80% 70% 60% 50% 40% 30% 20% % of Revenue from CSCO 90%
Total cumulativ e design w ins
Source: Company reports, UBS estimates
NETL CSCO Rev enue Percentage
(c) Expansion into adjacent market areas – For lower end search processing requirements, NetLogic has developed a product family aimed at the lower-end called NETLite. Products in this family have been designed for more cost sensitive applications that augment merchant switch/router chipsets from Broadcom and Marvell as well as network processors from EZchip Technologies and Xelerated. Although NETLite average selling prices are lower (typically $75-125 vs $150-350 its higher-end search processors), NETLite is margin neutral relative to NetLogic’s overall mix, as NetLogic has costreduced the family using leading-edge 55nm technology (from 130nm prior). For the high-end, NetLogic has developed a product family it calls NETL7 products used for deep packet inspection (DPI), a function required for content aware networking products such as security appliances. The NETL7 family supports a range of performance needs including systems with 10-gigabit per second interfaces, the fastest currently in volume production.
UBS 12
NetLogic Microsystems Inc 22 September 2009
(2) Potential for RMI processor wins at Cisco – we believe long-run success for RMI products will likely be judged based on NetLogic’s ability to meaningfully penetrate Cisco and dislodge incumbent Cavium in at least some new designs. If indeed RMI’s performance features prove advantageous in Cisco’s applications and result in material design wins, RMI revenues should accelerate over the next 2-3 years. Raising the likelihood this will occur is RMI’s inclusion of a specific processor operating mode that better matches Cisco’s legacy software requirements with RMI’s leading-edge architecture. NetLogic’s track record for execution and strong relationship with Cisco suggest this is possible and could position it for at least several years of accelerated growth as RMI design wins would move to production. The flip side is the lack of RMI wins at Cisco by the end of 2010 in our opinion would weigh on the stock, as the overhead associated with RMI is likely to slow NetLogic’s ability to reach GAAP profitability. For modeling purposes, we use a 50% probability for RMI design wins at Cisco in our analysis, with a breakout of our RMI revenue forecast and RMI ex-Cisco as follows; note that even without the contribution of Cisco revenues, our RMI revenue CAGR forecast is 10%. In Chart 3, the right hand axis shows Cisco revenues, of which 50% are added to our RMI exCisco revenue contribution, based on our 50% probability assignment.
Chart 3: NetLogic RMI Revenue Contribution
$35m $30m RMI Revenues $25m $20m $15m Dec-10 Dec-11 Dec-12 Dec-13 Mar-10 Mar-11 Mar-12 Mar-13 Sep-10 Sep-11 Sep-12 Sep-13 Jun-10 Jun-11 Jun-12 Jun-13 $20m $15m $10m $5m $0m RMI Revenues from Cisco
Raising the likelihood of RMI processor design wins at Cisco is RMI’s inclusion of a specific processor operating mode that better matches Cisco’s legacy software requirements with RMI’s leading-edge architecture
RMI Rev enues RMI Cisco Rev enues
Source: UBS estimates
RMI Rev enues ex -Cisco
While the Cisco revenue contribution looks small, on an EPS basis, it is material, especially as much of the gross profit from Cisco design wins should fall straight to operating income. Even using our revenue sensitivity model that we show below, which does not account for potentially better than expected operating leverage, each 1% rise in revenue raises EPS by ~1.5c. Thus, raising the probability to 100% from 50% for RMI processor Cisco design wins would raise 2012 GAAP EPS by 3c (+13%) and 2013 by 5c (+5%) – clearly a material catalyst.
UBS 13
NetLogic Microsystems Inc 22 September 2009
Valuation
Our 12-month price target of $51 (30x our $1.70 2010 non-GAAP EPS estimate) is based on our DCF model assumptions: 11.0% 10-year revenue CAGR, 25% long-term operating margin, 10.1% WACC, and 2% terminal growth. Because of the large difference between GAAP and non-GAAP earnings for NetLogic, we believe it is useful to examine free cash flow metrics for evaluating NetLogic’s valuation. While we primarily utilize a DCF model, which is based on our GAAP estimates where we expect operating margins over time to rise to 25%, we also examine at price-to-sales vs free cash flow margin (as a proxy for net margin).
Chart 4: Price-to-2010 Sales vs 2010 Free Cash Flow Margin
5.0x Price-to-Sales 4.0x 3.0x 2.0x 1.0x 0.0x 0% 5% 10% 15% FCF Margin
Source: Company reports, UBS estimates
NVDA SNDK AMD
NETL ADI TXN XLNX NSM ISIL MRVL BRCM INTC MXIM SWKS SIGM ONNN DIOD CY ATML MU RFMD FCS 20% 25%
ALTR
30%
While NetLogic valuation appears rich on 2010 estimates, with free cash flow margin expected to grow to 23% in 2011 from 16% in 2010, NetLogic valuation looks much more attractive. Among its fabless peers in our universe, NetLogic is priced above the group average on 2011 EV/EBITDA and EV/sales but below the group average on P/B and 2011 P/FCF.
Table 5: Peer Group Valuation Overview
EPS - Cal. Ticker NETL ALTR BRCM NVDA MRVL SIGM XLNX Average Source: Company reports, UBS estimates (Prices as of Sep 21st). Rating Buy Buy Neutral Neutral (CBE) Buy (CBE) Neutral Buy Price $43.52 $20.46 $29.84 $14.99 $15.74 $15.14 $22.82 Price Target $51.00 $23.00 $27.00 $14.00 $18.00 $15.00 $25.00 2010E -$1.84 $0.97 $0.49 $0.28 $0.81 $0.55 $1.15 2011E -$0.82 $1.11 $0.91 $0.36 $0.96 $0.61 $1.39 EV/Sales 2010E 2011E 3.2x 3.7x 2.6x 2.0x 2.8x 0.6x 3.3x 2.5x 2.6x 3.4x 2.2x 1.9x 2.3x 0.5x 3.1x 2.3x EV/EBITDA 2010E 2011E 12.9x 40.0x 18.4x 11.1x 5.4x 11.7x 16.6x 27.6x 11.4x 20.3x 14.5x 9.1x 4.3x 9.9x 13.9x 2010E 21.1x 19.5x 27.7x 19.8x 22.1x P/E 2011E 18.4x 32.8x 16.4x 24.7x 16.5x 21.7x P/B 2010E 3.5x 8.6x 3.3x 3.2x 2.1x 1.3x 2.9x 3.6x 2010E 20.9x 16.1x 22.0x 19.9x 17.4x 18.6x 19.1x P/FCF 2011E 12.5x 15.1x 17.1x 22.6x 17.0x 28.7x 15.2x 18.3x
UBS 14
NetLogic Microsystems Inc 22 September 2009
Stake Ownership
In 2Q09, Next Century Growth Investors, Turner Investment Partners and Franklin Advisers held the largest institutional positions, while Wasatch Advisors, historically a top holder, is still among the top 10. Founder Norm Godinho has several trust positions, together which total 2.5m shares, a position that has declined slowly over the past three years. Next Century Growth Investors, Goodman & Co. Investment Partners and Turner Investment Partners had the largest position increases in 2Q while Fidelity, Zweig-DiMenna Associates and Wasatch Advisors had the largest decreases.
Table 6: NetLogic Top 20 Holder History
Holder Name Next Century Growth Investors LLC Turner Investment Partners, Inc. Franklin Advisers, Inc. Daruma Asset Management, Inc. Barclays Global Investors NA Fred Alger Management, Inc. Vanguard Group, Inc. Wasatch Advisors, Inc. Goodman & Co. Investment Counsel Ltd. Columbus Circle Investors Dimensional Fund Advisors, Inc. Ranger Investment Management LP Arbor Capital Management LLC (Minnesota) Deutsche Investment Management Americas, Inc. BlackRock Advisors, Inc. MFS Investment Management Fuller & Thaler Asset Management, Inc. Evergreen Investment Management Co., Inc. Cadence Capital Management LLC Chartwell Investment Partners LP Source: FactSet 06-30-09 Pos 979,934 928,812 859,870 857,300 842,530 829,338 780,837 759,276 729,100 701,052 444,700 415,110 409,100 402,550 393,981 393,130 374,900 338,810 323,609 317,187 03-31-09 Pos 0 418,530 716,070 842,000 767,893 457,330 676,398 1,031,806 0 593,393 486,316 560,854 453,900 437,700 401,109 0 377,800 426,140 455,219 197,071 12-31-08 Pos 0 0 893,500 539,700 806,704 0 564,802 1,114,886 0 482,660 487,616 518,317 321,200 600,000 318,069 0 382,500 442,250 0 553,846 09-30-08 Pos 0 10,210 711,100 0 793,842 0 531,150 1,181,026 0 460,095 542,038 669,926 556,400 341,958 248,369 0 381,900 415,211 0 640,126 06-30-08 Pos 0 0 630,200 0 625,212 0 508,588 1,412,960 0 332,953 541,638 335,350 559,300 382,864 270,969 0 353,300 480,925 0 560,236 03-31-08 Pos 0 0 774,000 0 582,649 0 478,104 2,168,281 0 0 535,446 419,706 600,600 386,000 80,400 0 0 437,076 0 496,356 12-31-07 Pos 0 0 929,400 0 572,691 0 463,409 1,534,648 0 0 483,846 414,420 687,400 256,900 64,326 0 0 352,691 0 489,765 09-30-07 Pos 0 0 1,039,034 0 548,270 0 426,426 1,248,101 0 284,590 510,942 97,220 682,800 0 72,442 0 0 405,858 0 493,430
UBS 15
NetLogic Microsystems Inc 22 September 2009
Company Overview
NetLogic Microsystems is a fabless semiconductor company that designs, develops, and markets search processors, high-end embedded processors and high-speed physical layer products for wireline and wireless networking systems. NetLogic’s products enable network and content awareness for Internet Protocol (IP) infrastructure, such as routers, switches, network access equipment, wireless infrastructure equipment, network security appliances, and networked storage devices. It provides search processing devices (knowledgebased processors or KBPs), NETLite processors (low-end KBPs), legacy network search engines (NSEs), high-end embedded microprocessors (through its RMI acquisition) and optical interface devices. NetLogic sells its products directly to networking OEMs in the US, Asia and Europe, as well as through distributors and international stocking reps. NetLogic was founded in 1995 and is headquartered in Mountain View, California. About 90% of NetLogic’s revenue pre-RMI and IDT acquisitions have come from search processing, with about one-third of revenues from Cisco, down from more than 70% as recently as 2005. Much of NetLogic’s product is shipped to Asia Pacific, based on multinationals manufacturing product in Asia as well as indigenous manufacturers in the region.
Chart 5: 2Q09 Revenue by Search Customer
Phy sical Lay er Products, $3m Cisco, $11m
Chart 6: 2Q09 Revenue by Geography
EU/ROW, $5m US, $9m
Other NonCisco, $9m AlcatelHuaw ei, $3m Lucent, $5m
Asia Pacific, $18m
Source: Company reports, UBS estimates
Source: Company reports, UBS estimates
Post its recent acquisitions (IDT search processing, RMI), the company’s revenue base will be more diversified, as search processing will represent less than two-thirds of revenue, while Cisco should represent about a quarter of revenues (note about three-fourths of the IDT revenue will be part of Cisco revenue).
Chart 7: 4Q09E Revenue by Product Category
Phy sical Lay er Products, $5m Search Processing, $40m
Chart 8: 4Q09E Revenue Category, by Search Customer
Phy sical Lay er Products, $5m AlcatelRMI, $21m Lucent, $6m Huaw ei, $3m Other NonCisco, $11m IDT, $8m Cisco, $13m
RMI, $21m
Source: Company reports, UBS estimates
Source: Company reports, UBS estimates UBS 16
NetLogic Microsystems Inc 22 September 2009
Products
Search Processing Products (KBPs, NETLite)
Search processing technology is detailed in the appendix, but the key with these products is they are used to accelerate the intelligent routing of Internet Protocol (IP) packets, which is required for high-end enterprise, data center and carrier networks. The advancing capability of NetLogic’s search processors allows for increasingly more intelligent decision making as to how packets move from one router to the next in an IP network at increasingly faster interface speeds – up to 10-gigabit per second, currently increasing to 40 Gbps and 100 Gbps over the next several years. The search processor market has largely stagnated since 2006, mainly as Cisco has been able to utilize new suppliers to lower the pricing for its high-volume second and third generation TCAM products (in fact, a second version of the TCAM2 was called TCAM2-CR, i.e., cost reduced). Ex-Cisco, the search processor market is growing rapidly with an estimated 25% revenue CAGR (2005-10). In addition, we expect search processors at Cisco to grow more sustainably from 2009, as NetLogic’s technology has become more widely adopted in Cisco’s product line.
Chart 9: Search Processor Market by Vendor
$250m $200m $150m $100m $50m $0m 2009E 2010E 2003 2004 2005 2006 2007 2008
Chart 10: Search Processor Market – Cisco vs non-Cisco Split
$250m $200m $150m $100m $50m $0m 2009E 2010E
UBS 17
2003
2004
2005
2006
2007
NETL
IDTI
CY
Renesas
Non-Cisco
Source: Company reports, UBS estimates
Cisco
Source: Company reports, UBS estimates
Further supporting that growth at Cisco has been NetLogic’s cost reduction of its NL5000, with a revision that allows it to drop into a Renesas TCAM3 design for Cisco’s XR 12000-series router, which could provide some incremental upside to our estimates.
2008
NetLogic Microsystems Inc 22 September 2009
Search Processing Customer Base
NetLogic has organically won more than 275 search processing designs since inception, mainly among the major communications equipment manufacturers:
Table 7: Key Customer Design Win Examples
Customer Cisco Systems Alcatel-Lucent Huawei Juniper Brocade/Foundry Arris ALAXALA Networks Ericsson Nokia Siemens Networks Tellabs ZTE Platform Multiple Multiple Switches/Routers/IPTV E,M,T-series Routers Switches CMTS Router Redback 3G wireless equipment Access equipment 3G wireless Recent Product Family Wins NL6000/7000/8000/NETLite/TCAM3 NL7000/NETLite NL7000 NL9000 NL7000 NL7000 NL7000 NL7000/9000 NL7000/9000 NL7000/9000 NL7000/9000
Source: Company reports, UBS estimates
Embedded Processing Products (RMI)
Through its acquisition of RMI in 2009 (expected October close), NetLogic gains two primary product lines – the high-end XL-series processors, with the latest 40nm generation XLP claimed to be the industry flagship when it samples in early 2010 and the lower-end Alchemy processor family.
Chart 11: RMI Revenues
$140m $120m $100m $80m $60m $40m $20m $0m 2009E 2010E 2011E 2012E 2013E 2004 2005 2006 2007 2008
Source: Company reports, UBS estimates
RMI’s revenue growth trajectory prior to the downturn had been strong (boosted in part by the 2006 Alchemy acquisition), and we expect revenue growth to accelerate from 2009 on the revenue synergies from the NetLogic acquisition. We estimate this acquisition likely triples NetLogic’s total addressable market, and RMI shares many of NetLogic’s current customers including AlcatelLucent, Juniper Networks, Motorola, Huawei, ZTE, Fujitsu and NEC. RMI focuses on high-performance and low-power multicore, multi-threaded
UBS 18
NetLogic Microsystems Inc 22 September 2009
processors for networking/communications, security, storage, industrial and consumer applications. Average selling prices for its current products are the 10s of dollars, with the new XLP family products expected to drive ASPs of several hundreds of dollars across the family. RMI is based in Silicon Valley near NetLogic headquarters and has offices globally. Formerly known as Raza Microelectronics, Inc., RMI was incorporated in Delaware in December 2001 and changed its name to RMI Corporation in December 2007. Like standard x86 microprocessors, RMI’s 40nm XL-series processors are highly complex 64-bit machines, but are more focused on achieving low power consumption through the use of a simpler, more-efficient reduced instruction set computer (RISC) architecture (licensed from MIPS) and multithreading. We expect the XLP family to be initially offered as an 8-core product in early 2010 and later in 2010 to be offered in 1, 2 and 4-core configurations. If RMI’s performance claims are borne out and it can execute on its XLP and follow-on products and translate that execution to design wins at Cisco, we believe it could provide a significant catalyst for the stock. In 2008, NEC, Broadcom and PMC-Sierra led the $0.3bn 64-bit embedded processor market with combine 60% market share, but most of this revenue came from legacy products and revenue from those vendors was either flat or in decline. Cavium and RMI, the growth leaders, each had about 15% share. RMI also acquired in 2006, AMD’s Alchemy processor business (founded in 1999 and acquired by AMD in 2002), which is a line of 32-bit MIPS processors focused on more consumer-oriented, low-power applications. The Alchemy products do not have a clear strategic fit within NetLogic, and it is arguable if Alchemy’s low-power expertise in the long-run can enable material power savings for high-end RMI products that the core team could not develop anyway. However, the products do provide the company options to branch beyond its core IP communications infrastructure business. With some added complementary skill sets and a business that is unlikely to be much of a drag to either NetLogic’s financials or management focus, we believe the Alchemy business is unlikely to provide a catalyst, either positive or negative in the nearto-mid-term. In 2008, the 32-bit embedded processor market was a $2.5bn led by Freescale with more than $1bn of revenue, while RMI had 1% market share. For the combined 32/64-bit embedded processor market, Gartner forecasts a 5% revenue CAGR (2009-13).
Physical Layer Products (PLPs)
NetLogic acquired Aeluros in October 2007 to gain the expertise and intellectual property to integrate high-speed interfaces for its search processors. However, Aeluros products have also provided an attractive adjacent set of products that expand NetLogic’s reach into high-end communications infrastructure where search processors may not be used. Aeluros products are collectively referred to as serializer-deserializer or more commonly, SerDes chips, which are specialized interface devices that support very high-speed chip-to-chip communications. Serial interfaces have the significant advantage of reducing system complexity and power consumption by collapsing the number of parallel signal interfaces among the various chips needed to operate a high-speed
UBS 19
NetLogic Microsystems Inc 22 September 2009
system. The trade-off is cost, as the implementation of a SerDes is non-trivial (since a serialized signal must now operate at much higher speeds to support the data transmission rate of an equivalent set of parallel signals). The downside with SerDes devices, is that even though they are hard to implement, especially at low power consumption levels, they are a pure hardware play and given competition, there is much less product stickiness. Thus, even if NetLogic is the incumbent supplier, if a competitor beats it to market on the next generation of products, NetLogic could be totally displaced. Therefore, while this business may show success over time, we do not view it as a primary catalyst to the stock, as the category is always likely to retain its commodity-like status.
NetLogic Management
CEO, Ron Jankov has been with NetLogic since 2000 in the CEO, having been VP/General Manager at NeoMagic and Cyrix prior. We believe his prior expertise with microprocessors (Texas Instruments, Cyrix) is important towards building NetLogic into a leading specialized processor company. VP/CFO, Mike Tate joined NetLogic in 2007 after having held various financial roles at Galileo Technology/Marvell since 1997. We believe his extensive experience in fast growing environments in the fabless semiconductor business should support both revenue growth and tight expense control. CTO/VP of Product Development, Dr. Varad Srinivasan co-founded NetLogic (along with Norm Godinho) in 1995 and has a background in highspeed memory design. With advanced memory as a core strength, NetLogic was able to execute early generation search devices to establish its position in the market. While functional enhancements have become increasingly important over time, most of the hundreds of millions or billions of transistors in a search processor are still memory, reinforcing a key NetLogic competitive advantage. Senior VP of Worldwide Sales, Marcia Zander joined NetLogic in 1999 after having been with a manufacturer’s representative (Hitachi was a primary line) for 12 years prior. We believe her sales skills have been a very important factor for the company’s success, especially in the early years when the product portfolio and customer penetration outside of Cisco was limited.
UBS 20
NetLogic Microsystems Inc 22 September 2009
Earnings and Estimate Summary
We are above consensus for our out year estimates based on expectations for higher revenue growth. We believe NetLogic’s revenue growth is the primary driver for the stock, given expectations it can maintain gross margin and control opex to grow slower than sales. We do not expect NetLogic to make any material acquisitions in the near term, as we believe it will need to show it can generate design wins at Cisco for its RMI processors as proof it can effectively execute a complex acquisition. Revenues – we expect NetLogic revenues to grow 25% in 2009 to $175m, supported by its IDT network search engine and RMI acquisitions, which support expected 67% growth in 2010. Assuming no additional acquisitions, we expect NetLogic revenue to grow 16% in 2011, supported by in-line growth from its search processing business, and slightly slower RMI processor growth, which we expect to accelerate in 2012/13 as design wins for its 40nm XLP processor ramp into volume production. Gross margin – we expect gross margin to hover in NetLogic’s recent 65-67% range (includes stock-based compensation but excludes amortization), as its proprietary product position gives it pricing power while its fabless model offers very limited COGS leverage. Operating margin and EPS – while we expect operating margin and EPS to remain negative on a GAAP basis for several years as NetLogic digests its recent acquisitions, we expect non-GAAP earnings, a reasonable proxy for its free cash flow, to remain positive and grow.
Table 8: NetLogic Microsystems – Estimate Summary
F3Q09E Total Sales Sales Growth Operating margin Non-GAAP EPS EPS Source: UBS estimates $39m +21.1% q/q -5.0% q/q $0.32 -$0.12 F4Q09E $66m +66.9% q/q -27.9% q/q $0.30 -$0.72 F2009E $168m +19.9% y/y -16.0% $1.25 -$1.25 F2010E $291m +73.5% y/y -18.3% $1.70 -$1.84 F2011E $338m +16.2% y/y -8.0% $2.37 -$0.82
For 3Q09, we expect sales/EPS of $39.3m/-$0.12 ($0.32 non-GAAP) with 65.0% gross margin. For 4Q09, we expect sales/EPS of $65.6m/-$0.72 ($0.30 non-GAAP) vs consensus of $64.0m/$0.35 non-GAAP. For the full year 2009, we expect sales/EPS of $168m/-$1.25 ($1.25 non-GAAP). For 2010, we expect sales/EPS of $291m/-$1.84 ($1.70 non-GAAP), above consensus at $278m/$1.65 non-GAAP and guidance of $275m/$1.60 nonGAAP, and for 2011, we expect $338m/-$0.82 ($2.37 non-GAAP), above consensus of $319m/$2.07 non-GAAP.
UBS 21
NetLogic Microsystems Inc 22 September 2009
Table 9: NetLogic Microsystems – Estimate Summary vs. Consensus
Revenue UBS F3Q09E F4Q09E FY2009E FY2010E FY2011E $39.3m $65.6m $167.8m $291.1m $338.4m Consensus $45.0m $64.0m $171.8m $277.6m $319.2m Difference -13% +3% -2% +5% +6% UBS -$0.12 -$0.72 -$1.25 -$1.84 -$0.82 UBS nonGAAP $0.32 $0.30 $1.25 $1.70 $2.37 EPS Consensus $0.32 $0.35 $1.33 $1.65 $2.07 Difference +0% -14% -6% +3% +14% Gross Margin 65.0% 64.8% 65.6% 65.2% 66.0% Op Margin -5.0% -27.9% -16.0% -18.3% -8.0% Op Profit -$2.0m -$18.3m -$26.8m -$53.3m -$27.1m
Source: FactSet and UBS estimates
Revenue—EPS Sensitivity Analysis
Versus our baseline 2010 model, we estimate that for each percentage point change in revenue, EPS changes by ~1.5c. Excluding mix, gross margin is largely insensitive to revenue changes given its variable cost fabless model as shown by the historical relationship:
Chart 12: NetLogic Y/Y Gross Profit vs Y/Y Revenue
60% Y/Y NETL Revenues 40% 20% 0% -20% Dec-06 Dec-07 Dec-08 Mar-07 Mar-08 Mar-09 Sep-06 Sep-07 Sep-08 Jun-07 Jun-08 Jun-09 60% 40% 20% 0% -20% Y/Y NETL Gross Profit
-2% $285m -$1.87
NETL Y/Y Rev enues NETL Y/Y Gross Profit (ex -SBC, amort)
Source: Company reports, UBS estimates
Our underlying assumptions are: Y/Y gross profit = 1.0x Y/Y revenue; Y/Y, operating expenses = 0.6x Y/Y revenue:
Table 10: NetLogic 2010 Sensitivity to Revenue Changes
-8% Revenue EPS $268m -$1.96 Change in EPS = Source: UBS estimates -6% $274m -$1.93 -4% $279m -$1.90 0% $291m -$1.84 2% $297m -$1.81 4% $303m -$1.78 6% $309m -$1.75 8% $314m -$1.72
1.5c for each percentage change in revenue
UBS 22
NetLogic Microsystems Inc 22 September 2009
Share Price Drivers
We believe NetLogic’s stock price is a largely a function of its ability to drive top-line growth and fall through to operating profit.
Chart 13: NetLogic Revenue vs. Share Price
$40m $35m NETL Revenues $30m $25m $20m $15m $10m $5m Sep-04 Sep-05 Sep-06 Sep-07 Sep-08 Mar-05 Mar-06 Mar-07 Mar-08 Mar-09 $40 $35 $30 $25 $20 $15 $10 $5 $0
Chart 14: NetLogic Op Profit (ex-SBC, amort) vs. Share Price
$10m $44 $36 $32 $28 $24 $20 $2m Sep-05 Sep-06 Sep-07 Sep-08 Mar-06 Mar-07 Mar-08 Mar-09 $16 Avg Q NETL Share Price
UBS 23
Avg Q NETL Share Price
$40 NETL Op Profit $8m $6m $4m
NETL Rev enues Av erage/current stock price for period (EPS currency , US$)
Source: Company reports and UBS estimates
NETL Operating Profit (ex -SBC, amort) Av erage/current stock price for period (EPS currency , US$)
Source: Company reports and UBS estimates
We believe the company will be able to generate a return on capital employed greater than its cost of capital over time, as it realizes the revenue synergies from its Aeluros and RMI acquisitions.
Revenue Drivers
We believe NetLogic’s revenue growth is the primary driver for the stock, especially given expectations for opex growth relative to sales to decline. Thus, we focus primarily on the components of its top-line drivers:
Chart 15: NetLogic Revenue Forecast
$100m $80m $60m $40m $20m $0m Dec-09 Dec-10 Dec-11 Sep-09 Sep-10 Sep-11 Sep-12 Dec-12 Jun-10 Jun-11 Mar-10 Mar-11 Mar-12 Jun-12
Cisco-NL5000 Revenues Cisco-Other Revenues RMI Revenues
Source: UBS estimates
IDT Revenues Search Processor Revenues ex-Cisco, IDT Physical Layer Products Revenues
To build our NetLogic revenue model, we focus on the key driver – design wins. For search processors, we aggregated NetLogic’s business excluding the IDTI acquisition and compared this with cumulative design wins 2 years prior, which is a likely gestation period from declaration of a design win to volume production. We have estimated the design win count based on the number (or
NetLogic Microsystems Inc 22 September 2009
range) the company has stated on its earnings conference calls over its history and used the cumulative number to date, as this is likely to be a reasonable proxy for active designs given the typically long life cycles of communication equipment. We used this trend to calibrate our trajectory for search processing revenue over the next two years, after which we assume a slowing rate of growth as the market matures:
Chart 16: NetLogic Search Processing Revenue ex-IDT vs Cumulative Design Wins
$50m $45m NETL Revenues $40m $35m $30m $25m $20m Sep-05 Sep-06 Sep-07 Sep-08 Sep-09 Sep-10 Sep-11 Mar-06 Mar-07 Mar-08 Mar-09 Mar-10 Mar-11 300 Cumulative Design Wins 250 200 150 100 50 0
NETL Revenues less IDTI, RMI, PLP Total cumulative design wins - 2 years prior
Source: Company reports, UBS estimates
Given the potential for an increasing number of RMI processor design wins post the acquisition, especially at Cisco, enabled by NetLogic’s reputation and balance sheet, we expect RMI y/y revenues to grow from the low-teens to the mid-teens by 2012. We expect physical layer product revenue growth to decline over time as competition increases (mainly from Broadcom) and the category becomes more commoditized. As NetLogic’s largest customer, Cisco is clearly a driver. However, the correlation between Cisco’s switching and routing revenues and NetLogic revenues is not as close as may be expected.
Chart 17: NetLogic Revenues vs Cisco Switch/Router Revenues
$50m NETL Revenues $40m $30m $20m $10m $0m Mar-04 Jun-04 Sep-04 Dec-04 Mar-05 Jun-05 Sep-05 Dec-05 Mar-06 Jun-06 Sep-06 Dec-06 Mar-07 Jun-07 Sep-07 Dec-07 Mar-08 Jun-08 Sep-08 Dec-08 Mar-09 Jun-09 $6.0bn $5.5bn $5.0bn $4.5bn $4.0bn $3.5bn $3.0bn $2.5bn
Chart 18: NetLogic @ Cisco vs Cisco Switch/Router Revenues
$20m $6.0bn $5.5bn $5.0bn $4.5bn $4.0bn $3.5bn $3.0bn $2.5bn Mar-04 Jun-04 Sep-04 Dec-04 Mar-05 Jun-05 Sep-05 Dec-05 Mar-06 Jun-06 Sep-06 Dec-06 Mar-07 Jun-07 Sep-07 Dec-07 Mar-08 Jun-08 Sep-08 Dec-08 Mar-09 Jun-09
CSCO Switch/Route Revs
NETL CSCO Revenues
$15m $10m $5m $0m
NETL Rev enues CSCO Sw itching + Routing Rev enues
Source: Company reports
NETL CSCO Rev enues CSCO Sw itching + Routing Rev enues
Source: Company reports
UBS 24
CSCO Switch/Route Revs
NetLogic Microsystems Inc 22 September 2009
In addition, we believe there are other growth companies that are positioned similarly to NetLogic that make good comparisons. In semiconductors, Cavium, especially now with NetLogic’s acquisition of RMI, is likely to be considered the closest comparable:
Chart 19: NetLogic vs Cavium Revenues
$50m $45m NETL Revenues $40m $35m $30m $25m $20m $15m Sep-05 Sep-06 Sep-07 Sep-08 Sep-09 Sep-10 Mar-05 Mar-06 Mar-07 Mar-08 Mar-09 Mar-10 $40m $35m $25m $20m $15m $10m $5m $0m Cavium Revenues $30m
NETL Rev enues ex -IDTI, RMI
Source: FactSet (consensus estimates), company reports, UBS estimates
CAVM Rev enues
An important trend in networking equipment is the growth of Layer 4-7 Application Delivery Controllers (ADCs), where F5 and Citrix are among the market leaders. Driving this growth is IT network, server and storage resource consolidation, with virtualization as an enabling technology. While Intel-based servers remain an important silicon solution for many network appliances like ADCs, specialized processors from Cavium and NetLogic/RMI are used in higher end products. Notably, revenues for F5 and Citrix have tracked with NetLogic, and given a focus on addressing similar market problems, we believe they are also good proxies for growth:
Chart 20: NetLogic vs F5 Revenues
$50m NETL Revenues $40m $30m $20m $10m $0m Sep-03 Mar-04 Sep-04 Mar-05 Sep-05 Mar-06 Sep-06 Mar-07 Sep-07 Mar-08 Sep-08 Mar-09 Sep-09 Mar-10 Sep-10 $100m $50m $0m $200m NETL Revenues FFIV Revenues $150m
Chart 21: NetLogic vs Citrix Revenues
$50m $40m $30m $20m $10m $0m Mar-04 Sep-04 Mar-05 Sep-05 Mar-06 Sep-06 Mar-07 Sep-07 Mar-08 Sep-08 Mar-09 Sep-09 Mar-10 Sep-10 $500m $450m $400m $350m $300m $250m $200m $150m $100m
NETL Rev enues ex -IDTI, RMI
FFIV Rev enues
NETL Rev enues ex -IDTI, RMI
CTXS Rev enues
Source: FactSet (consensus estimates), company reports, UBS estimates
Source: FactSet (consensus estimates), company reports, UBS estimates
UBS 25
Citrix Revenues
NetLogic Microsystems Inc 22 September 2009
Gross/Operating Margin Drivers
As a fabless semiconductor maker with proprietary, high-end products, NetLogic’s gross profit sensitivity to changes in revenue has been very limited. While gross margin is being negatively affected by its recent acquisitions, we expect the company’s ability to manage costs and gain greater purchasing power as a result of limited competition to help it to return to the high-end of its historical 65-67% gross margin level over time. We believe NetLogic will be able to expand operating margin as it expands revenue and grows opex slower than revenue; in recent history Y/Y opex has been about 80% of Y/Y revenues. We expect opex growth to decelerate to about 60% of Y/Y revenues, as NetLogic focuses on bolstering its leading position in search processing and incrementally expands on the RMI XL-series processors and SerDes products.
Chart 22: NETL Y/Y Gross Profit vs Y/Y Revenue
60% Y/Y NETL Revenues 40% 20% 0% -20% Dec-06 Dec-07 Dec-08 Sep-06 Sep-07 Sep-08 Mar-07 Mar-08 Mar-09 Jun-07 Jun-08 Jun-09 60% 40% 20% 0% -20% Y/Y NETL Gross Profit Y/Y NETL Revenues
Chart 23: Y/Y Operating Profit vs Y/Y Revenue
60% 40% 20% 0% -20% Dec-06 Dec-07 Dec-08 Sep-06 Sep-07 Sep-08 Mar-07 Mar-08 Mar-09 Jun-07 Jun-08 Jun-09 48% 40% 32% 24% 16% 8% 0% -8% -16%
NETL Y/Y Rev enues NETL Y/Y Gross Profit (ex -SBC, amort)
Source: Company reports and UBS estimates
NETL Y/Y Rev enues NETL Y/Y Opex (ex -SBC, amort)
Source: Company reports and UBS estimates
On a non-GAAP model basis, NetLogic gross margin/operating margin has been around 65%/20%, and we believe over time, the company can expand operating margin largely on operating leverage expectations.
Chart 24: NetLogic Gross Margin, Operating Margin
70% 60% 50% 40% 30% 20% 10% 0% Dec-05 Dec-06 Dec-07 Dec-08 Dec-09 Dec-10 Dec-11 Dec-12 Jun-05 Jun-06 Jun-07 Jun-08 Jun-09 Jun-10 Jun-11 Jun-12
NETL Gross Margin (ex -SBC, amort) NETL Operating Margin (ex -SBC, amort)
Source: Company reports and UBS estimates
UBS 26
Y/Y NETL Opex
NetLogic Microsystems Inc 22 September 2009
Cash Flow Drivers
NetLogic’s Free Cash Flow is directly driven by Net Income growth, as we expect stock-based compensation and amortization of acquisition related intangibles to be about $25m/quarter for the next several years. We expect NetLogic to regain and sustain Free Cash Flow Margin of 20% or greater from 2011.
Chart 25: NetLogic Net Income vs Free Cash Flow
$10M $5M NETL Net Income $0M -$5M -$10M -$15M -$20M -$25M Mar-04 Sep-04 Mar-05 Sep-05 Mar-06 Sep-06 Mar-07 Sep-07 Mar-08 Sep-08 Mar-09 Sep-09 Mar-10 Sep-10 Mar-11 Sep-11 Mar-12 Sep-12 $30m $25m $20m $15m $10m $5m $0m -$5m -$10m
Chart 26: NetLogic Free Cash Flow vs Free Cash Flow Margin
$30m $25m $20m $15m $10m $5m $0m -$5m -$10m Mar-04 Sep-04 Mar-05 Sep-05 Mar-06 Sep-06 Mar-07 Sep-07 Mar-08 Sep-08 Mar-09 Sep-09 Mar-10 Sep-10 Mar-11 Sep-11 Mar-12 Sep-12 60% 40% 20% 0% -20% -40% -60% -80% NETL FCF Margin
UBS 27
NETL FCF
NETL Net Income
Source: Company reports and UBS estimates
NETL FCF
NETL FCF
NETL FCF
Source: Company reports and UBS estimates
NETL FCF Margin
Inventory Analysis
NetLogic’s inventory days have oscillated around 110 days with inventory more recently below 100 days. We believe NetLogic’s inventory risk is low, as it operates essentially under a “build-to-order” model and its products have a life cycle that can span up to a decade or more.
Chart 27: NetLogic Inventory and Inventory Days
$18m $15m Inventory Value $12m $9m $6m $3m $0m Sep-04 Sep-05 Sep-06 Sep-07 Sep-08 Mar-04 Mar-05 Mar-06 Mar-07 Mar-08 Mar-09 140 120 Inventory Days 100 80 60 40 20 0
Chart 28: NetLogic Inventory Days and Cash Conversion
140 120 100 80 Days 60 40 20 0 Sep-04 Sep-05 Sep-06 Sep-07 Sep-08 Mar-04 Mar-05 Mar-06 Mar-07 Mar-08 Mar-09
NETL Inv entory
Source: Company reports
NETL Inv entory Day s
Source: Company reports
NETL Inv entory Day s NETL Cash Conv ersion Cy cle
NetLogic Microsystems Inc 22 September 2009
Consideration of Stock-based Comp Expenses
In 2008, NetLogic’s stock-based compensation expenses were $16m (11% of sales). We expect 2010 stock options expenses to grow to $49m (17% of sales) based on the acquisition of RMI. In 2011, we expect stock-based compensation to be about the same level (declining to 15% of sales).
Chart 29: NetLogic Non-GAAP Operating Income, Stock-based Compensation as a Percentage of Non-GAAP Operating Income
$30m NETL non-GAAP Op Inc $25m $20m $15m $10m $5m $0m Sep-05 Sep-06 Sep-07 Sep-08 Sep-09 Sep-10 Sep-11 Sep-12 Mar-05 Mar-06 Mar-07 Mar-08 Mar-09 Mar-10 Mar-11 Mar-12 140% 100% 80% 60% 40% 20% 0% SBC % non-GAAP Op Inc 120%
NETL Operating Profit (ex -SBC, amort) NETL Stock Option Ex penses as Percentage of Op Income ex -SBC, amort
Source: Company reports and UBS estimates
UBS 28
NetLogic Microsystems Inc 22 September 2009
Appendix
UBS 29
NetLogic Microsystems Inc 22 September 2009
NSE/KBP Overview
TCAM Is Underlying Technology
Underlying the product categories of Network Search Engines (NSEs) and Knowledge-based Processors (KBPs) are TCAMs or ternary content addressable memories. NetLogic has branded its products as Knowledge-based Processors (KBPs), all of which are built on TCAM technology. NetLogic invented the term Knowledge-Based Processor (KBP), in which it adds advanced decision-making capability around the TCAM, utilizing the knowledge base in the network. TCAMs are different in both form and function vs standard memories – while standard memories are used to store a particular piece of data at a particular address (location), TCAMs perform a high-speed pattern match of the input data vs the data in all of the storage locations in the TCAM, which in turn provides the location(s) of the matching result. Hence, when reading from a standard memory, the input is an address location and the output is the data stored at that address. A TCAM in essence performs the opposite function. The input is a piece of data (e.g., the destination for an Internet Protocol packet) and the output is the address location(s) at which that data is stored (which is used to tell the switch/router onto which port interface it should transmit the packet).
“T” in TCAM Is for Ternary—Needed for IP Networks
The “T” in TCAM stands for ternary, as the memory cell in a TCAM is physically a 3-state memory, which adds an additional state called “don’t care” (denoted as “X”), in addition to the standard memory states, 0 or 1. This is useful for masking off a suffix of the stored data, which is most applicable Internet Protocol (IP) routing. The mask can vary in length, as it can include locations just within the local “sub-network” (usually called a subnet) where the final destination resides or some larger network encompassing many subnets. Thus, when a new piece of data is presented to the TCAM, it is actually matching the input against stored prefixes and looks for the matching result with the longest prefix, i.e., “longest-prefix match,” with the resident address location in the TCAM as the output result. The longest-prefix match is most desired because it provides the closest location in the network to the packet’s final destination. (In the case of no matches, a more time consuming process involving the switch/router’s packet processor can be used to determine how to get the IP packet to its destination.) For cost and power considerations, the result from a TCAM usually is fed to a static RAM (SRAM), a conventional type of high-speed memory. The output from the SRAM will then tell the packet processor in the switch/router which port interface the packet should be transmitted in order to reach the next switch/router (often referred to as the next “hop”) towards its final destination.
Cisco Enabled TCAM Market
Cisco Systems enabled the TCAM market in the late 1990s when technology industry pioneer, Andreas Bechtolsheim (founded Sun Microsystems) led a decision-making process within Cisco that concluded next generation switching/routing systems would need TCAMs to achieve desired levels of performance and functionality. Integrated Device Technology (IDT), Cypress Semiconductor and NetLogic were all early pioneers in the TCAM market. IDT
UBS 30
NetLogic Microsystems Inc 22 September 2009
led the market after Cisco chose IDT over Micron and Freescale (then Motorola) to be the initial supplier for TCAM2 (also known as the Gen 2 TCAM). Later, Cisco pulled in Cypress Semiconductor and Renesas as alternate sources.
Cisco’s Gen 3 TCAM Enabled NetLogic’s Growth
For its third generation device, TCAM3, Cisco chose IBM as its lead vendor; incumbent IDT decided not to bid on the program, as initial production volume estimates were small, and IDT sought early to deploy its resources to diversify from Cisco. This turned out to be a strategic mistake, as NetLogic’s NL5000 was adopted for use by Cisco in certain TCAM3 applications, which became the inflection point for the company. Only later did IDT buy IBM’s TCAM3 business to become an alternate source to NetLogic but was never able to reestablish a sustained leading position versus NetLogic and later Renesas. Recently, NetLogic began shipping a cost-competitive TCAM3 compliant device, challenging Rensesas’s current leadership.
Expanded Uses for Search Processing
Most commonly, the hardware search function is used for network routing – i.e., moving an Internet Protocol (IP) packet from one router to the next towards its end destination. However, these routes are not simply dependent on the source and destination (and choosing the most optimal path in-between) but also the sender (and his network access privileges), the type of data (e.g., voice video, data) and the packet contents. This not only raises the utility of the network, but also it raises the complexity and accordingly the performance requirements for NetLogic search processing products, which is positive for supporting revenue growth expectations.
Algorithmic Approach to Search Processing
Algorithmic search processors are not TCAMs but rather use standard SRAM or DRAM technology along with some logic to emulate a TCAM. While algorithmic search processors use less power and can support larger densities than TCAMs, they have the drawback of longer latency (i.e., the time delay to generate a result, which can negatively affect overall performance). In all cases where OEMs are implementing hardware-based search and not utilizing TCAMbased search processors (from NetLogic), they are implementing their own proprietary search processing algorithm. TCAMs and algorithmic search engines are quite complementary. TCAMs provide for the lowest latency, which is required for the more time dependent operations, mainly the basic operation of determining the best way for packets to reach their destination, as this bottleneck effectively determines the throughput of the packet processor with which the search processor is working in tandem. However, in an example of inspecting packet contents for viruses, the algorithmic search processor needs to disassemble, inspect and reassemble a series of packets, making latency much less of factor for determining overall performance, as the search processor needs large chunks of information before it can start to do any useful work. The complementary nature of theses approaches is a key reason that NetLogic’s search processors will utilize a configurable hybrid TCAM/algorithmic approach
UBS 31
NetLogic Microsystems Inc 22 September 2009
subsequent to its current hybrid NL9000-family flagship. We expect future search processors to continue to evolve using its base of TCAM/algorithmic search technology.
Cisco’s Spans Logic Uses Algorithmic Approach
Cisco has publicly disclosed very little about its in-house search processing capability based on its 2007 Spans Logic acquisition. However, based on our industry sources, we have gleaned a few data points. Cisco acquired Spans Logic to gain access to lower cost DRAM-based search technology for lower performance routers, which could run on special purpose custom chips or even on general-purpose embedded processors.
Table 11: NetLogic vs Cisco/Spans Logic Search Processing Approach
NetLogic Search technology Typical throughput Latency Typical table size Source: Parimics TCAM/SRAM-based 100 Gbps (130m packets/s) > 1 packet time 1m entries
NetLogic believes it is gaining share vs Cisco’s in-house algorithmic search processing, as Cisco platforms including the ASR 9000 and XR 12000 have mission critical functions for which algorithmic search falls short. We see no evidence that Cisco’s internal search efforts are a serious competitive threat, although it could limit NetLogic’s ability to penetrate into lower end, higher volume applications.
UBS 32
NetLogic Microsystems Inc 22 September 2009
Processor Overview
NetLogic has acquired RMI Corporation mainly for its high-end processor capabilities, which complement NetLogic’s own search processing devices. What is the difference between search processing and general purpose processing? Search processing addresses specific problems related to determining how Internet Protocol packets move through a network where general purpose processors can be used for multiple high-end embedded processing tasks beyond communications equipment, such as high-end office automation equipment or medical equipment. However, the volume applications that drive much of the demand for high-end embedded processors are similar to what NetLogic already serves – communications infrastructure. As communications infrastructure adopts Internet Protocol with all of its feature rich potential at up to 100x the speeds of just a decade ago, both high-speed search processing and high-speed generalpurpose processors are required. Over time, we would expect NetLogic to integrate its layer 4-7 NETL7 search processors with its RMI processors for enabling high-performance content processing, such as what is needed in the fast growing category of application delivery controllers (ADCs), which help manage IT network, server and storage resource consolidation (F5, Cisco and Citrix are the market leaders). RMI processors have a similar to architecture to other multicore processors, such as Intel’s Nehalem family processors, where independent central processing units or CPUs have local high-speed memory known as (level 1, level 2) cache and share a last-level (level 3) cache. However, because high-end embedded processors are typically addressing more power-sensitive applications, they are more optimized for low power consumption (~50W for RMI’s 8-core XLP vs >100W for Intel’s 8-core Nehalem-EX).
Chart 30: RMI XLP Processor Block Diagram
L-3 Cache Memory Distributed Interconnect™
L2 Cache L1 Cache L2 Cache L1 Cache L2 Cache L1 Cache
CPU1
1 2 3 4
CPU2
5 6 7 8
CPUn
XLP™ XLP™ XLP
Advanced Fast Messaging Network™ I/O Distributed Interconnects I/O
Source: NetLogic Microsystems
Programmable Autonomous Accelerators
Versus its closest competitor, Cavium, NetLogic, claims a number of architectural and raw clock speed advantages, which coupled with its choice to use 40nm (vs Cavium’s latest generation Octeon II is at 65nm), could give it a performance edge, especially for applications that are highly parallel in nature:
UBS 33
NetLogic Microsystems Inc 22 September 2009
Table 12: RMI Processor Advantages
Key RMI Processor Features Core speed (2 GHz+) Multi-Threading Fast Messaging Network Memory Distributed Interconnect Autonomous Packet Accelerators Optimized memory architecture Inter-chip Coherency Interface Source: NetLogic Microsystems, UBS Advantage High raw performance Increased performance through parallelism Increased scalability of multiple processor cores on a chip Speeds up performance to main memory, which is often a bottleneck Off-load overhead tasks, for higher performance, lower power Dedicated resources, very high speed access to internal cache memory Increased scalability of multiple processors per system
The ability to realize these advantages highly depends on the application and the way in which the programmers take care to develop/adapt their software in order to take advantage of the multicore architecture. Ultimately, OEMs will typically benchmark competitors’ processors running their specific operating system and applications in a head-to-head competition in what is often referred to as a bakeoff. In addition to a vendor’s reputation and road map, the results of a bakeoff will be critical to which vendor the OEM chooses.
UBS 34
NetLogic Microsystems Inc 22 September 2009
PLP Overview
Through its acquisition of Aeluros, NetLogic gained a portfolio of physical layer products (PLPs or transceivers), which are specialized to cost-effectively transmit/receive the highest-speed electrical signals. Today such signals are in the range of 10-gigabits per second. Techniques to enable high-speed signaling include: 1) use of low voltages, as the time it takes to change the voltage from a logical zero state to a logical one state is less as the voltage swing is decreased; 2) differential pairs, where the same signal is transmitted in both its standard configuration and an inverted configuration so when the two signals are “subtracted” from each other by the receiver, the “noise” picked up from the line is largely cancelled out; 3) transmission of a clock signal with the data (source synchronous clocking), which ensures that the receiver can resync the data with the system clock to ensure that data integrity is maintained; 4) EDC or error detection and correction, where additional information is transmitted with the data so that errors can be at a minimum detected, and if the errors are small enough, can be corrected by the receiver; and 5) match printed circuit board (PCB) trace (i.e., wire) lengths to ensure multiple signals arrive at a receiver at the same time. In the world of nanotech, in one nanosecond or billionth of a second, an electrical signal travels about 10 inches, so if a receiver with multiple signal inputs receives them at slightly different times because of mismatched trace lengths, the likelihood the resulting data is “misinterpreted” can rise significantly. While electrical transmission is used for the vast majority of network interconnections, in the “backbone” of the carrier network and in the Internet/enterprise data center, signals are often transmitted optically (i.e., using laser light). When those signals need to be processed in almost any fashion (with the exceptions of being amplified to extend the distance or multiplexed without any payload processing occurring), they need to be converted to electrical signals. With the proliferation of equipment to handle 10 Gbps signals, physical layer products such as those from NetLogic are an important enabling technology. The Aeluros transceivers are sold to makers of optical “modules,” which include the optical transponder (into which an optical fiber cable is plugged in) and the electrical interface, which communicates via a standardized protocol, the most common being XAUI (10-gigabit attachment unit interface, pronounced ZOWee) – the “X” in this case is the Roman numeral 10, which is common among many electrical and optical interfaces for 10 Gbps communications. The 10 Gbps transceivers perform an additional function call Serializer/ Deserializer, better known as SerDes (pronounced SIR-deez). As the name indicates, it converts a high-speed serial signal, effectively the optical signal converted to an electrical signal which in turn is converted to a set of parallel (or deserialized) signals that operate at more conventional frequencies used across an entire circuit board. In the case of 10-gigabit Ethernet, a single 10 Gbps signal called SFI (SerDes Framer Interface) is converted to 4 lower-speed XAUI signals (2 in some of the newest generation systems – called reduced XAUI or RXAUI; accordingly RXAUI signals run at 2x the speed of XAUI).
UBS 35
NetLogic Microsystems Inc 22 September 2009
Historically, OEMs such as Cisco would purchase the optical “transponders,” which have included the SerDes chip from vendors such as NetLogic. Thus, even though Cisco could use a NetLogic SerDes chip embedded in an optical transponder, it would purchase the transponder from optical module makers such as Avago, Finisar and Opnext. The combined electrical/optical subsystem was originally defined using a standard called XENPAK in 2001. Standards subsequent to XENPAK have evolved to support higher physical port density and lower power consumption. NetLogic claims its has 10% of the XENPAK SerDes market, which it expects to rise to 50% for a follow-on generation of a smaller-sized standard called X2 in the 2010 time frame. Beyond XENPAK and X2 is a more recent standard called SFP+, in which the SerDes is on the OEM’s circuit board rather than in the optical transponder and communicates to the transponder via the SFI interface, which is now exposed. The ability to shrink the optical module while excluding the SerDes has allowed for further raising the port density. In turn, because there no longer needs to be a one-for-one relationship between the optics and the SerDes, a market for “dual” and “quad” SerDes has emerged so that a single chip integrating two or four SerDes can interface with two or four SFP+ transponders respectively, thus reducing system chip count and power consumption vs individual SerDes.
UBS 36
NetLogic Microsystems Inc 22 September 2009
7-Layer OSI Model
When NetLogic talks about the fact that it plays at all “layers” in the network, what does it mean? Like all complex systems, an IP network is divided into a hierarchy of complexity. In 1977, the International Organization for Standardization (ISO) initiated development of a framework, the Open Systems Interconnection (OSI) reference model, which partitioned a network into 7 layers, with each layer representing a way the data is processed at a node (i.e., within a piece of equipment) on the network.
OSI Model Is Typically Called a “Protocol Stack”
At the top of network protocol “stack” is the Application Layer (Layer 7), and at the bottom is the Physical Layer (Layer 1). “Applications” are not the typical software applications like Microsoft Word but rather network applications, the most common being HTTP (hypertext transfer protocol) for transmitting web pages. At each layer, the data is split into header information and payload, with each lower layer protocol “encapsulating” the successive higher layer protocol information.
Various Layers Result in Specialized Communications Infrastructure
To see why these layers matter, it is useful to understand how the equipment at each node is operating at each specific layer. Thus, while the OSI model seems very abstract, the implication for the equipment that operates at each layer is very real, and the requirements are highly distinct. While higher layer protocols involve more complexity, this is typically traded off for more throughput at the lower layers. Thus, the core of any carrier network typically comprises Layer 1 equipment (e.g., reconfigurable optical add-drop multiplexers or ROADMs) while closer to the user comprises equipment that operates at Layers 2/3 (switches/routers) along with specialized appliances that operate at layer 4-7 (e.g., Application Delivery Controllers).
Table 13: Open Systems Interconnection (OSI) Reference Model
Layer 7. Application 6. Presentation 5. Session 4. Transport 3. Network 2. Data Link 1. Physical Source: Wikipedia Segment Packet Frame Bit Data unit Data Function Network process to application Data representation and encryption Interhost communication End-to-end connections and reliability Path determination and logical addressing Physical addressing Media, signal and binary transmission Examples HTTP, FTP, SMTP SSL, MPEG SIP TCP IP Ethernet (MAC) Ethernet (PHY)
Layer 1—Physical Layer; Muxing Equipment
At Layer 1, signals are multiplexed or “muxed,” most often in the time domain. The most prominent example of a Layer 1 network is the traditional “TDM” or “time-division multiplexed” telephone network. Telephone networks are synchronized to a master clock so time slots do not get jumbled (even across thousands of miles), which would prevent the ability to maintain a telephone
UBS 37
NetLogic Microsystems Inc 22 September 2009
connection. Modern packet networks use high-speed optical signals at 10 or 40gigabits per second, where signals can be multiplexed among time slots or even different wavelengths of light (wave division multiplexing or WDM). Layer 1 processing also incorporates data integrity mechanisms to ensure that the data being transmitted from one end is the same as when it reaches its destination. Examples of Layer 1 equipment include the Tellabs 5500 Digital Cross Connect and the Fujitsu Flashwave 7500 Reconfigurable Add-Drop Multiplexer (ROADM). In addition, passive optical networks (PONs) utilize time division multiplexing to parse optical signals between the Optical Network Unit (ONU) and each Optical Network Terminal (ONT). Encapsulated within this Layer 1 payload is the end-user data as well as all of the necessary information for the higher layers. Higher layer processing is typically more compute intensive, so data is often transmitted at lower layers as much as possible for cost reasons, and when it gets closer to the user, it is terminated up to the application layer.
Layer 2—Datalink Layer; Switching Equipment
Layer 2 is known as the data link layer. Layer 2 signals are switched and are no longer bound by the rules of time slots or different wavelengths of light or radio frequency spectrum. In this way, a network can be provisioned for a greater amount of data than its theoretical maximum (overprovisioning) because most of the time, the network is not fully utilized. This property is known as statistical multiplexing and is a key reason that the telephone networks are a dying technology – while they are very reliable, they are also terribly inefficient. Operating at Layer 2 is the way most small-scale data networks operate, as switching offers the efficiency of statistical multiplexing while requiring very low overhead, as each switch on a Layer 2 network only needs to know which port to pass on a packet. The most common Layer 2 protocol is the ubiquitous Ethernet, which is not only a technology for homes and enterprises but increasingly carrier networks. Examples of Layer 2 equipment include the Netgear GS105 5-port gigabit switch (under $50) up through a Cisco Nexus 7000 Data Center switch. Examples of other Layer 2 technologies that have largely disappeared in new network deployments include Asynchronous Transfer Mode (ATM) and Token ring.
Layer 3—Network Layer; Routing Equipment
Layer 3 is known as the network layer. Layer 3 signals are moved through the network by routers, which can manage for much greater complexity and scale than at Layer 2. For example, if a packet has a choice of two different routes it can be transmitted across to reach its destination, rules can be defined for those networks. This could result in routers choosing different routes for the same packet depending on parameters such as who sent the packet, who will receive the packet, what kind of data is contained in the packet and the time the packet is being sent. Examples of Layer 3 equipment include the Cisco CRS-1 and Juniper T-series core routers.
UBS 38
NetLogic Microsystems Inc 22 September 2009
Layer 4—Transport Layer
Layers 4, 5 and 6 sit between IP packets and the Layer 7 Application Layer information encapsulated inside the IP packet header. Layer 4 or the transport layer provides an end-to-end view of the IP traffic flow by establishing a “connection” between the source and destination. Because statistical multiplexing bets that most but not necessarily all of the time the network will have enough bandwidth, packets are occasionally dropped. If data integrity is important, the network application will use TCP (transmission control protocol) to ensure dropped packets are retransmitted. If time sensitivity is more important, the network application should use UDP (user datagram protocol), which attempts to move data across the network with the lowest latency and does not attempt to retransmit dropped packets. When networks are referred to as “TCP/IP networks,” this is a generic term that encompasses UDP as well. Most network applications run over TCP, but real-time applications such as voice over IP use UDP. During a VoIP phone call, dropped packets result in degraded voice quality.
Layer 5—Session Layer; VoIP Equipment
Layer 5 is the Session Layer and is an extension to layer 4, increasingly associated with Voice/Video over IP (VoIP) using SIP (Session Initiation Protocol). A telephone conversation may be defined as a session, which is maintained for the duration of the call. An example of Layer 5 equipment is the Acme Packet Session Border Gateway Net-Net family.
Layer 6/7—Presentation/Application Layer
Layer 6 or the Presentation Layer is associated with Layer 7 and “presents” decompressed or decrypted information in a standard format to the Layer 7 application. As Layers 5 and 6 are extensions of Layers 4 and 7, respectively, the OSI model is sometimes simplified as a 5 layer model. Products that operate at Layer 7 (which encompass Layers 4-7), including Application Delivery Controllers (ADCs), are used for server load balancing and denial of service (DoS) attack prevention. These include the F5 BigIP appliance, Cisco Content Services Switches and Brocade/Foundry ServerIron family.
NetLogic Products Encompass all 7 Layers
As alluded to by this overview, understanding networking technology in depth requires gaining knowledge of a vast alphabet soup of acronyms. We have touched only on a small number. To remain relevant to the context of NetLogic products, we believe it is more useful to think about functionality at a high level: Layer 1 = Muxing, Layer 2 = Switching, Layer 3 = Routing and Layer 7 = Applications. NetLogic Physical Layer Products are Layer 1 interface devices, NetLogic Knowledge-based (Search) Processors address Layers 2-3 and NETL7 products address through Layers 4-7.
UBS 39
Table 14: NetLogic (NETL) Income Statement ($ millions, except per share data)
NetLogic Microsystems
Income Statement Fiscal year ends Dec 31 ($ in millions, except EPS) Total Revenues Cost of Revenues Gross Profit R&D Expenses SG&A Expenses Intangible Asset Amortization, Other Costs Total Operational Expenses Operating Income Interest Income and Other, net Pretax Income Income Tax Net Income Net Income ex-SBC, Charges Stock-based compensation (net of tax) Extraordinary items (net of tax) GAAP Net Income Diluted EPS GAAP EPS Diluted EPS ex-SBC, Charges Diluted Shares Outstanding (GAAP) Diluted Shares Treasury (PF) Percentage of Revenues Gross Profit R&D Expenses SG&A Expenses Operating Income Interest Income and Other, net Pretax Income Income Tax Tax Rate Net Income Q-Q Growth Total Revenues Net Income Y-Y Growth Total Revenues Net Income 2007 109.0 37.7 71.3 45.2 19.7 7.0 71.8 (0.5) 4.5 3.9 0.2 3.7 26.7 16.0 (1.1) 2.6 0.17 0.12 1.20 21.9 22.2 FY2008 Mar-08 Jun-08 Sep-08 Q1A Q2A Q3A 34.2 36.5 38.3 11.5 12.4 12.8 22.6 24.2 25.5 12.2 12.4 13.6 6.5 6.2 7.2 3.8 3.6 4.0 22.5 22.2 24.8 0.1 2.0 0.7 0.5 0.3 0.4 0.6 2.2 1.1 (0.5) (0.1) (0.2) 1.1 2.3 1.3 8.7 3.7 1.1 0.05 0.05 0.38 22.1 22.7 9.2 3.3 2.3 0.10 0.10 0.40 22.5 23.2 9.8 4.6 1.3 0.06 0.06 0.42 22.8 23.5 Dec-08 Q4A 30.9 10.1 20.8 13.4 6.7 3.4 23.5 (2.6) 0.4 (2.3) (1.1) (1.1) 6.8 4.5 (1.1) (0.05) (0.05) 0.29 21.7 23.1 2008 139.9 46.8 93.1 51.6 26.6 14.8 92.9 0.1 1.5 1.6 (1.9) 3.6 34.4 16.1 3.6 0.16 0.16 1.54 22.3 22.3 Mar-09 Q1A 30.4 10.2 20.1 12.2 6.8 3.3 22.3 (2.2) 0.2 (2.0) 1.9 (3.9) 6.7 4.3 (3.9) (0.18) (0.18) 0.29 21.8 23.4 FY2009E Jun-09 Sep-09 Q2A Q3E 32.5 39.3 10.7 13.8 21.8 25.6 14.1 15.0 7.4 8.1 4.7 4.4 26.2 27.5 (4.3) (2.0) 0.2 (0.6) (4.1) (2.6) (1.9) 0.2 (2.2) (2.8) 8.4 6.0 (2.2) (0.10) (0.10) 0.35 22.0 24.1 8.0 6.3 (2.8) (0.12) (0.12) 0.32 22.5 24.6 Dec-09 Q4E 65.6 23.1 42.5 29.0 16.5 15.3 60.8 (18.3) (0.6) (18.9) 2.0 (20.9) 9.4 15.0 (20.9) (0.72) (0.72) 0.30 29.0 31.5 2009E 167.8 57.8 110.0 70.4 38.7 27.8 136.9 (26.8) (0.8) (27.6) 2.2 (29.8) 32.5 31.6 (29.8) (1.25) (1.25) 1.25 23.8 25.9 Mar-10 Q1E 67.6 23.7 43.8 29.3 15.5 15.3 60.1 (16.3) (0.4) (16.7) 2.0 (18.7) 9.1 12.5 (18.7) (0.63) (0.63) 0.29 29.5 32.0 FY2010E Jun-10 Sep-10 Q2E Q3E 71.7 74.2 25.0 25.8 46.7 48.5 29.6 30.1 15.2 15.4 15.3 15.3 60.1 60.9 (13.4) (12.4) (0.2) (13.6) (12.4) (13.6) (12.4) 14.2 12.5 (13.6) (0.45) (0.45) 0.44 30.0 32.5 15.4 12.5 (12.4) (0.41) (0.41) 0.47 30.5 33.0 Dec-10 Q4E 77.6 26.8 50.8 30.9 15.8 15.3 62.0 (11.1) 0.2 (10.9) (10.9) 16.9 12.5 (10.9) (0.35) (0.35) 0.50 31.0 33.5 2010E 291.1 101.3 189.8 119.9 61.9 61.3 243.1 (53.3) (0.4) (53.7) 2.0 (55.7) 55.6 50.0 (55.7) (1.84) (1.84) 1.70 30.3 32.8 Mar-11 Q1E 80.7 27.6 53.1 32.1 16.1 13.5 61.8 (8.7) 0.4 (8.3) (8.3) 17.8 12.6 (8.3) (0.26) (0.26) 0.53 31.4 33.9 FY2011E Jun-11 Sep-11 Q2E Q3E 84.7 85.0 28.8 28.7 55.9 56.2 32.3 32.3 16.5 16.5 13.5 13.5 62.3 62.3 (6.4) (6.1) 0.4 0.4 (6.0) (5.7) (6.0) (5.7) 20.2 12.7 (6.0) (0.20) (0.20) 0.61 30.6 33.1 20.6 12.8 (5.7) (0.18) (0.18) 0.60 31.7 34.2 Dec-11 Q4E 87.9 29.7 58.2 33.7 16.8 13.5 64.1 (5.9) 0.4 (5.5) (5.5) 20.9 12.9 (5.5) (0.18) (0.18) 0.63 31.0 33.5 2011E 338.4 114.9 223.4 130.4 65.9 54.2 250.5 (27.1) 1.6 (25.5) (25.5) 79.7 51.0 (25.5) (0.82) (0.82) 2.37 31.2 33.7 2012E 382.5 127.4 255.0 137.6 70.3 42.8 250.7 4.3 3.2 7.5 0.8 6.7 96.8 47.3 6.7 0.21 0.21 2.85 32.6 34.0 2013E 425.0 141.6 283.4 146.7 74.6 34.0 255.3 28.1 5.2 33.3 3.3 30.0 112.8 48.8 30.0 0.89 0.89 3.25 33.8 34.8
NetLogic Microsystems Inc 22 September 2009
65.4% 41.4% 18.0% -0.5% 4.1% 3.6% 0.2% 6.3% 3.4%
66.2% 35.7% 18.9% 0.4% 1.4% 1.8% -1.5% -83.6% 3.3%
66.1% 66.5% 33.8% 35.6% 17.1% 18.8% 5.4% 1.8% 0.7% 1.1% 6.1% 2.8% -0.3% -0.4% -5.2% -15.4% 6.4% 3.3%
67.4% 43.4% 21.6% -8.5% 1.1% -7.4% -3.7% 50.1% -3.7%
66.5% 36.9% 19.0% 0.1% 1.1% 1.2% -1.4% -118.1% 2.6%
66.3% 40.2% 22.4% -7.2% 0.6% -6.6% 6.3% -94.2% -12.9%
67.2% 43.5% 22.7% -13.3% 0.8% -12.6% -5.9% 47.2% -6.6%
65.0% 38.2% 20.5% -5.0% -1.5% -6.5% 0.5% 0.0% -7.0%
64.8% 44.2% 25.1% -27.9% -0.9% -28.9% 3.0% 0.0% -31.9%
65.6% 41.9% 23.1% -16.0% -0.5% -16.5% 1.3% -7.9% -17.7%
64.9% 43.4% 22.9% -24.1% -0.6% -24.7% 3.0% 0.0% -27.7%
65.1% 41.3% 21.2% -18.8% -0.3% -19.0% 0.0% 0.0% -19.0%
65.3% 40.6% 20.8% -16.7% 0.0% -16.7% 0.0% 0.0% -16.7%
65.5% 39.8% 20.3% -14.4% 0.3% -14.1% 0.0% 0.0% -14.1%
65.2% 41.2% 21.3% -18.3% -0.1% -18.4% 0.7% -3.7% -19.1%
65.8% 39.8% 19.9% -10.8% 0.5% -10.3% 0.0% 0.0% -10.3%
66.0% 38.1% 19.5% -7.6% 0.5% -7.1% 0.0% 0.0% -7.1%
66.2% 38.0% 19.4% -7.2% 0.5% -6.7% 0.0% 0.0% -6.7%
66.2% 38.4% 19.1% -6.7% 0.5% -6.2% 0.0% 0.0% -6.2%
66.0% 38.6% 19.5% -8.0% 0.5% -7.5% 0.0% 0.0% -7.5%
66.7% 36.0% 18.4% 1.1% 0.8% 2.0% 0.2% 11.0% 1.7%
66.7% 34.5% 17.6% 6.6% 1.2% 7.8% 0.8% 10.0% 7.1%
-6.0% 6.9% 4.8% -19.4% -- -134.9% 106.9% -46.1% -190.6%
---
-1.7% 244%
7.0% -45.0%
21.1% 28%
66.9% 657%
---
3.0% -11%
6.1% -26.9%
3.5% -9.0%
4.6% -12%
---
4.0% -24.2%
5.0% -27.5%
0.3% -5%
3.5% -4%
---
---
---
12.6% -68.7%
46.0% -33.1%
41.4% 39.1% -3.6% -58.5%
-4.2% -65.8%
28.3% -4.2%
-11.2% -451%
-11.1% -195%
2.6% -323%
112.4% 1277%
19.9% -880.0%
122.5% 253%
120.8% 363%
88.8% 230%
18.3% -51%
73.5% 47.2%
19.4% -58%
18.2% -57%
14.5% -56%
13.3% -50%
16.2% -56%
13.0% -125%
11.1% 333%
Source: Company reports and UBS estimates
UBS 40
Table 15: NetLogic (NETL) Balance Sheet ($ millions, except per share data)
NetLogic Microsystems
Balance Sheet Statement Fiscal year ends Dec 31 ($ in millions, except EPS) Assets Cash and ST Investments Accounts Receivable, net Inventories Deferred income taxes Prepaid Expenses and Other Assets Total Current Assets Property and Equipment, net Goodwill Intangible Assets Other Assets Other LT assets Total Assets Liabilities Accounts Payable Accrued Salaries and Wages Deferred Revenue Capital Lease Obligations Total Current Liabilities LT Liabilities LT Debt Capital Lease Obligations, Other Total Liabilities Convertible Preferred stock Common stock APIC Accumulated Deficit/Retained Earnings Deferred compensation expense Notes Receivable from Stockholders Accumulated other comprehensive (loss)/income Shareholder Equity Total Liabilities & Shareholder Equity Operating Activity Day Sales Outstanding Inventory Turns Inventory Days Payable Period Cash Conversion Cycle 2007 50.7 14.8 12.9 12.7 91.2 5.7 55.4 52.8 0.1 108.4 205.3 FY2008 Mar-08 Jun-08 Sep-08 Q1A Q2A Q3A 55.8 19.8 16.7 12.8 105.1 5.2 54.1 49.5 0.1 103.7 214.0 73.7 14.7 16.1 14.3 118.7 5.7 53.8 46.2 0.1 100.1 224.5 85.9 14.7 14.4 14.5 129.6 6.3 53.8 42.9 0.2 96.8 232.7 Dec-08 Q4A 96.5 8.4 13.7 5.2 123.8 5.5 68.7 39.5 8.2 116.5 245.8 2008 96.5 8.4 13.7 5.2 123.8 5.5 68.7 39.5 8.2 116.5 245.8 Mar-09 Q1A 88.9 8.3 10.9 2.9 3.9 114.9 4.8 68.7 36.2 5.2 110.1 229.8 FY2009E Jun-09 Sep-09 Q2A Q3E 88.3 8.0 9.7 2.9 6.9 115.7 7.9 68.7 32.9 21.1 122.7 246.3 68.1 13.1 13.8 2.9 6.9 104.8 11.9 101.7 243.9 21.1 366.7 483.4 Dec-09 Q4E 56.4 21.9 23.1 2.9 6.9 111.2 12.0 101.7 228.5 21.1 351.4 474.6 2009E 56.4 21.9 23.1 2.9 6.9 111.2 12.0 101.7 228.5 21.1 351.4 474.6 FY2010E Mar-10 Jun-10 Sep-10 Dec-10 2010E Q1E Q2E Q3E Q4E 48.9 26.3 23.7 2.9 6.9 108.7 12.1 101.7 213.2 21.1 336.1 456.9 50.4 27.9 25.0 2.9 6.9 113.1 12.2 101.7 197.9 21.1 320.7 446.1 38.9 28.9 25.8 2.9 6.9 103.3 12.3 101.7 182.6 21.1 305.4 421.0 43.4 30.2 26.8 2.9 6.9 110.2 12.4 101.7 167.2 21.1 290.1 412.7 43.4 30.2 26.8 2.9 6.9 110.2 12.4 101.7 167.2 21.1 290.1 412.7 Mar-11 Q1E 61.1 31.4 27.6 2.9 6.9 129.8 12.5 101.7 153.7 21.1 276.6 418.9 FY2011E Jun-11 Sep-11 Q2E Q3E 79.6 33.0 28.8 2.9 6.9 151.2 12.6 101.7 140.2 21.1 263.0 426.8 85.4 33.1 28.7 2.9 6.9 157.0 12.7 101.7 126.6 21.1 249.5 419.2 Dec-11 2011E Q4E 105.2 34.2 29.7 2.9 6.9 178.9 12.8 101.7 113.1 21.1 235.9 427.7 105.2 34.2 29.7 2.9 6.9 178.9 12.8 101.7 113.1 21.1 235.9 427.7 2012E 202.0 38.6 33.1 2.9 6.9 283.6 13.2 101.7 70.3 21.1 193.2 489.9 2013E 317.3 42.4 36.3 2.9 6.9 405.9 13.6 101.7 36.3 21.1 159.2 578.6
NetLogic Microsystems Inc 22 September 2009
7.1 13.3 0.3 2.5 23.2 8.0 2.1 33.4 0.2 251.2 (79.6) (0.0) 171.9 205.3
12.3 11.2 0.5 2.1 26.0 10.8 36.8 255.6 (78.4) (0.0) 177.2 214.0
10.2 11.6 2.1 1.1 25.0 11.1 0.4 36.5 264.1 (76.1) (0.0) 188.0 224.5
10.8 11.5 1.8 0.9 25.1 11.6 0.5 37.2 270.4 (74.8) (0.0) 195.5 232.7
7.6 25.9 1.6 0.8 35.9 0.5 9.1 45.5 0.2 276.0 (76.0) (0.0) 200.3 245.8
7.6 25.9 1.6 0.8 35.9 0.5 9.1 45.5 0.2 276.0 (76.0) (0.0) 200.3 245.8
8.3 10.1 0.6 0.8 19.8 0.0 9.4 29.3 0.2 280.2 (79.9) 0.0 200.5 229.8
9.6 13.2 0.5 3.4 26.7 10.9 37.6 0.2 290.6 (82.1) 208.8 246.3
11.5 6.0 0.5 3.4 21.4 55.0 76.4 467.6 (60.6) 407.0 483.4
19.3 6.0 0.5 3.4 29.2 44.0 73.2 467.6 (66.2) 401.4 474.6
19.3 6.0 0.5 3.4 29.2 44.0 73.2 467.6 (66.2) 401.4 474.6
18.5 6.0 0.5 3.4 28.4 33.0 61.4 467.6 (72.1) 395.5 456.9
19.5 6.0 0.5 3.4 29.4 22.0 51.4 467.6 (73.0) 394.6 446.1
20.0 6.0 0.5 3.4 30.0 11.0 41.0 467.6 (87.6) 380.0 421.0
20.8 6.0 0.5 3.4 30.8 30.8
20.8 6.0 0.5 3.4 30.8 30.8
21.5 6.0 0.5 3.4 31.4 31.4 467.6 (80.1) 387.5 418.9
22.4 6.0 0.5 3.4 32.4 32.4 467.6 (73.1) 394.5 426.8
22.4 6.0 0.5 3.4 32.3 32.3 467.6 (80.7) 386.9 419.2
23.1 6.0 0.5 3.4 33.1 33.1 467.6 (73.0) 394.6 427.7
23.1 6.0 0.5 3.4 33.1 33.1 467.6 (73.0) 394.6 427.7
23.9 6.0 0.5 3.4 33.9 33.9 467.6 (11.5) 456.1 489.9
26.2 6.0 0.5 3.4 36.2 36.2 467.6 74.9 542.4 578.6
467.6 467.6 (85.7) (85.7) 381.9 381.9 412.7 412.7
49.7 2.9 123.4 68.6 104.5
52.0 2.8 131.9 96.1 87.8
36.1 3.1 118.5 74.0 80.6
34.6 3.6 102.4 76.0 60.9
24.4 2.9 124.1 68.0 80.5
21.9 3.4 105.3 59.4 67.8
24.5 3.8 97.3 73.1 48.8
22.2 4.4 82.6 80.6 24.2
30.0 4.0 91.3 75.0 46.3
30.0 4.0 91.3 75.0 46.3
47.6 2.5 144.0 121.7 69.9
35.0 4.0 91.3 70.0 56.3
35.0 4.0 91.3 70.0 56.3
35.0 4.0 91.3 70.0 56.3
35.0 4.0 91.3 70.0 56.3
37.8 3.8 95.2 75.1 58.0
35.0 4.0 91.3 70.0 56.3
35.0 4.0 91.3 70.0 56.3
35.0 4.0 91.3 70.0 56.3
35.0 4.0 91.3 70.0 56.3
36.9 3.9 93.1 73.5 56.6
36.9 3.8 93.5 68.5 61.9
36.4 3.9 92.4 67.6 61.2
Source: Company reports and UBS estimates
UBS 41
Table 16: NetLogic (NETL) Statement of Cash Flows ($ millions, except per share data)
NetLogic Microsystems
Statement of Cash Flows Fiscal year ends Dec 31 ($ in millions, except EPS) Net Income Depreciation & amortization Stock-based compensation Other Charges Changes in assets and liabilities Accounts receivable, net Inventories Prepaid expenses and other assets Accounts payable Deferred revenue Other accrued liabilities Net cash flow from operating activities Investments in property, plants, and equipme Investments, net Restricted cash Cash paid for acquisition Other Net cash flow from investing activities Capital lease obligations Repurchase of restricted and common stock Proceeds from issuance of common stock Proceeds from payment of notes receivable Proceeds from notes payable Repayment of notes payable Other Net cash flow from financing activities FX Effects Change in cash Beginning cash Ending cash 2007 2.6 9.1 16.0 (2.7) (0.1) (4.5) 1.5 (1.3) 1.0 0.3 2.9 24.9 (2.2) 39.8 (70.2) (32.6) 8.3 (0.7) 7.7 (0.0) (0.1) 50.8 50.7 FY2008 Mar-08 Jun-08 Sep-08 Q1A Q2A Q3A 1.1 2.3 1.3 4.3 4.2 4.4 3.7 3.3 4.6 0.5 0.3 1.1 (4.0) 4.5 0.8 (4.9) 5.2 (0.1) (2.4) 0.4 0.6 (0.2) (1.4) (0.4) 5.2 (2.1) 0.7 0.2 1.7 (0.3) (1.9) 0.8 0.4 5.6 14.6 12.1 (0.4) (0.2) (0.7) Dec-08 Q4A (1.1) 4.4 4.9 (3.9) 5.3 6.3 (0.1) 2.7 (3.2) (0.2) (0.3) 9.5 (0.1) (13.0) 0.2 (0.4) (0.2) (0.7) (12.9) 2008 3.6 17.2 16.5 (2.0) 6.6 6.6 (1.4) 0.6 0.5 1.3 (1.0) 41.9 (1.4) (13.0) 0.2 (14.3) 7.9 (2.7) 5.2 (0.0) 32.8 50.7 83.5 FY2009E Mar-09 Jun-09 Sep-09 Q1A Q2A Q3E (3.9) (2.2) (2.8) 4.3 4.3 5.4 4.3 6.0 6.3 3.6 0.9 0.0 0.4 1.9 (7.3) 0.1 0.3 (5.1) 2.5 0.4 (4.1) (2.0) (3.1) 0.7 1.3 1.9 (1.0) (0.1) 0.1 3.2 8.7 11.0 1.7 (0.3) (6.7) (15.5) (22.5) (0.2) 19.8 (15.0) 4.5 (0.8) (253.0) (253.8) (0.8) Dec-09 Q4E (20.9) 16.3 15.0 0.0 (10.3) (8.8) (9.3) 7.8 0.1 (0.8) 2009E (29.8) 30.4 31.6 4.5 (15.3) (13.5) (10.6) (5.1) 11.6 (1.1) 3.3 21.4 (2.1) 13.1 (283.5) (272.5) 181.9 54.2 (11.0) (1.1) 224.0 0.0 (27.1) 83.5 56.4 Mar-10 Q1E (18.7) 16.3 12.5 0.0 (5.8) (4.4) (0.6) (0.8) 4.3 (0.8) FY2010E Jun-10 Sep-10 Q2E Q3E (13.6) (12.4) 16.3 16.3 12.5 12.5 0.0 0.0 (1.9) (1.1) (1.6) (1.0) (1.3) (0.7) 1.0 0.6 13.3 (0.8) 15.3 (0.8) (15.0) (0.8) (0.8) (15.8) (0.8) Dec-10 Q4E (10.9) 16.3 12.5 0.0 (1.5) (1.3) (1.0) 0.8 16.3 (0.8) 2010E (55.7) 65.3 50.0 (10.4) (8.3) (3.7) 1.6 49.2 (3.2) (15.0) (18.2) (44.0) (44.0) 0.0 (13.0) 56.4 43.4 Mar-11 Q1E (8.3) 15.5 12.6 0.0 (1.4) (1.2) (0.8) 0.6 18.5 (0.8) FY2011E Jun-11 Sep-11 Q2E Q3E (6.0) (5.7) 14.5 14.5 12.7 12.8 0.0 0.0 (1.8) (0.1) (1.6) (0.1) (1.2) 0.1 0.9 (0.1) 19.4 (0.8) 21.6 (0.8) (15.0) (0.8) (0.8) (15.8) (0.8) Dec-11 Q4E (5.5) 14.5 12.9 0.0 (1.4) (1.1) (1.0) 0.8 20.6 (0.8) 2011E 2012E 2013E
NetLogic Microsystems Inc 22 September 2009
(25.5) 6.7 30.0 59.2 47.8 39.0 51.0 52.6 54.2 (4.7) (7.0) (4.7) (4.0) (4.5) (3.8) (2.9) (3.4) (3.2) 2.3 0.8 2.3 80.0 100.0 118.5 (3.2) (15.0) (18.2) 0.0 61.8 43.4 105.2 (3.2) (3.2) 0.0 96.8 105.2 202.0 (3.2) (3.2) 0.0 115.3 202.0 317.3
0.4
5.3
1.8
0.4
(0.1)
5.0 (0.8)
177.0 55.0 (11.0) 232.0 (20.1) 88.3 68.1 (11.0) (11.7) 68.1 56.4
(11.0) (11.0) (7.5) 56.4 48.9
(11.0) (11.0) 1.5 48.9 50.4
(11.0) (11.0) (11.5) 50.4 38.9
(11.0) (11.0) 4.5 38.9 43.4
(0.5) (0.1) (0.0) 5.1 50.7 55.8
(1.8) 3.5 (0.0) 17.9 55.8 73.7
(0.9) 0.9 (0.0) 12.3 73.7 85.9
0.5 1.0 (0.0) (2.5) 85.9 83.5
(0.4) (0.5) (14.3) 83.5 69.1
(0.7) 3.6 19.1 69.1 88.3
17.7 43.4 61.1
18.6 61.1 79.6
5.8 79.6 85.4
19.8 85.4 105.2
Source: Company reports and UBS estimates
UBS 42
NetLogic Microsystems Inc
Income statement (US$m)
Revenues Operating expenses (ex depn) EBITDA (UBS) Depreciation Operating income (EBIT, UBS) Other income & associates Net interest Abnormal items (pre-tax) Profit before tax Tax Profit after tax Abnormal items (post-tax) Minorities / pref dividends Net income (local GAAP) Net Income (UBS) Tax rate (%) Pre-abnormal tax rate (%)
12/04
48 (49) (1) (2) (3) 0 0 (4) (7) 0 (7) 0 0 (7) (3) 0 0
12/05
82 (62) 19 (2) 17 0 1 0 19 0 18 0 0 18 18 2 2
12/06
97 (84) 13 (5) 8 0 4 (11) 1 0 1 0 0 1 11 44 4
12/07
109 (100) 9 (9) (1) 0 4 (2) 2 0 2 0 0 2 3 11 20
12/08
140 (123) 17 (17) 0 0 2 0 2 2 4 0 0 4 4 0 0
12/09E
168 (164) 4 (30) (27) 0 (1) 0 (28) (2) (30) 0 0 (30) (30) 0 0
% ch
19.9 34.0 -79.6 76.4 -
12/10E
291 (279) 12 (65) (53) 0 0 0 (54) (2) (56) 0 0 (56) (56) 0 0
% ch
73.5 69.9 239.5 115.1 98.7 -48.8 94.5 -7.9 87.0 87.0 87.0 -
12/11E
338 (306) 32 (59) (27) 0 2 0 (26) 0 (26) 0 0 (26) (26) 0 0
% ch
16.2 9.7 167.0 -9.4 -49.1 -52.5 -54.2 -54.2 -54.2 -
Per share (US$)
EPS (local GAAP) EPS (UBS) Net DPS Cash EPS BVPS
12/04
(0.63) (0.29) 0.00 (0.14) 2.74
12/05
0.96 0.96 0.00 1.15 3.84
12/06
0.03 0.53 0.00 0.82 7.02
12/07
0.09 0.14 0.00 0.59 8.11
12/08
0.16 0.16 0.00 0.97 9.17
12/09E
(1.25) (1.25) 0.00 0.02 16.85
% ch
-97.5 83.7
12/10E
(1.84) (1.84) 0.00 0.32 12.62
% ch
47.2 47.2 1193.9 -25.1
12/11E
(0.82) (0.82) 0.00 1.08 12.67
% ch
-55.5 -55.5 240.2 0.3
Balance sheet (US$m)
Cash and equivalents Other current assets Total current assets Net tangible fixed assets Net intangible fixed assets Investments / other assets Total assets Trade payables & other ST liabilities Short term debt Total current liabilities Long term debt Other long term liabilities Total liabilities Equity & minority interests Total liabilities & equity
12/04
41 15 56 3 0 0 59 11 0 11 0 0 11 48 59
12/05
66 16 81 4 0 0 86 16 0 16 0 1 17 69 86
12/06
90 20 110 6 42 0 158 14 0 14 0 2 15 143 158
12/07
51 40 91 6 108 0 205 23 0 23 0 10 33 172 205
12/08
97 27 124 6 108 8 246 36 0 36 0 10 46 200 246
12/09E
56 55 111 12 330 21 475 29 0 29 44 0 73 401 475
% ch
-41.6 101.0 -10.2 118.0 205.1 157.1 93.1 -18.7 -18.7 60.9 100.4 93.1
12/10E
43 67 110 12 269 21 413 31 0 31 0 0 31 382 413
% ch
-23.1 21.9 -0.9 3.3 -18.6 0.0 -13.0 5.4 5.4 -57.9 -4.9 -13.0
12/11E
105 74 179 13 215 21 428 33 0 33 0 0 33 395 428
% ch
142.4 10.4 62.4 3.2 -20.1 0.0 3.6 7.4 7.4 7.4 3.3 3.6
Cash flow (US$m)
Net income Depreciation Net change in working capital Other (operating) Net cash from operations Capital expenditure Net (acquisitions) / disposals Other changes in investments Cash from investing activities Increase/(decrease) in debt Share issues / (repurchases) Dividends paid Other cash from financing Cash from financing activities Cash flow chge in cash & equivalents FX / non cash items Bal sheet chge in cash & equivalents Core EBITDA Maintenance capital expenditure Maintenance net working capital Operating free cash flow, pre-tax
12/04
(7) 2 (9) 10 (4) (1) 0 8 7 0 40 0 (7) 33 36 (16) 20 (1) (1) (9) (11)
12/05
18 2 2 5 28 (2) 0 0 (2) 0 2 0 (2) 0 26 (2) 24 19 (2) 2 20
12/06
1 5 (8) 25 22 (2) (1) (40) (42) 0 5 0 (1) 4 (16) 40 24 13 (2) (8) 4
12/07
2 9 0 13 24 (2) (70) (30) (103) 0 8 0 (1) 8 (71) 32 (39) 9 (2) 0 6
12/08
4 17 7 14 42 (1) 0 (13) (14) 0 8 0 (3) 5 33 13 46 17 (1) 7 23
12/09E
(30) 30 (15) 36 21 (2) (284) (270) (556) 0 182 0 (1) 181 (354) 314 (40) 4 (2) (15) (14)
% ch
76.4 149.5 -48.9 44.6 3856.8 3359.7 2003.2 -79.6 44.6 -
12/10E
(56) 65 (10) 50 49 (3) (15) (15) (33) 0 0 0 0 0 16 (29) (13) 12 (3) (10) (2)
% ch
87.0 115.1 -31.9 38.5 129.9 53.9 -94.7 -94.0 -94.4 239.5 53.9 -31.9 -88.4
12/11E
(26) 59 (5) 51 80 (3) (15) (15) (33) 0 0 0 0 0 47 15 62 32 (3) (5) 24
% ch
-54.2 -9.4 -55.2 2.0 62.6 0.0 0.0 0.0 0.0 167.0 0.0 -55.2 -
Source: Company accounts, UBS estimates. (UBS) valuations are stated before goodwill-related charges and other adjustments for abnormal and economic items at the analysts' judgement. Note: For some companies, the data represents an extract of the full company accounts.
UBS 43
Global Equity Research
Americas Semiconductors 12-month rating 12m price target
NetLogic Microsystems Inc
Buy
US$51.00
Valuation (x)
P/E (local GAAP) P/E (UBS) P/CEPS Net dividend yield (%) P/BV EV/revenue (core) EV/EBITDA (core) EV/EBIT (core) EV/OpFCF (core) EV/op. invested capital
5Yr Avg
-
12/07
NM NM 49.7 0.0 3.6 5.0 63.1 NM NM 6.2
12/08
NM NM 28.4 0.0 3.0 3.7 29.8 NM 22.9 4.6
12/09E
NM NM NM 0.0 2.6 5.4 >100 NM NM 3.7
12/10E
NM NM NM 0.0 3.5 3.2 77.7 NM NM 2.6
12/11E
NM NM 40.5 0.0 3.5 2.6 27.7 NM NM 2.8
Company profile
NetLogic Microsystems is a fabless semiconductor company that designs, develops, and markets search processors, high-end embedded processors and high-speed physical layer products for networking systems including routers, switches, network access equipment, wireless infrastructure equipment, network security appliances, and networked storage devices. It provides search processing devices (knowledge-based processors or KBPs), NETLite processors (low-end KBPs), legacy network search engines (NSEs), high-end embedded microprocessors (through its RMI acquisition) and optical interface devices.
Enterprise value (US$m)
Average market cap + minority interests + average net debt (cash) + pension obligations and other - non-core asset value Core enterprise value
12/07
612 0 (70) 0 0 542
12/08
590 0 (74) 0 0 516
12/09E
961 0 (54) 0 0 907
12/10E
961 0 (28) 0 0 933
12/11E
961 0 (74) 0 0 887
Value (EV/OpFCF & P/E)
25.0x 20.0x 15.0x 10.0x 5.0x 0.0x 12/07 12/08 12/09E EV/OpFCF (LHS) P/E (RHS) 12/10E 0.0x -10.0x -20.0x -30.0x -40.0x -50.0x -60.0x 12/11E
Growth (%)
Revenue EBITDA (UBS) EBIT (UBS) EPS (UBS) Cash EPS Net DPS BVPS
5Yr Avg
68.5 -
12/07
12.6 -33.7 -73.5 -28.0 15.5
12/08
28.3 102.0 10.2 63.7 13.1
12/09E
19.9 -79.6 -97.5 83.7
12/10E
73.5 NM 98.7 47.2 NM -25.1
12/11E
16.2 167.0 -49.1 -55.5 NM 0.3
Margins (%)
EBITDA / revenue EBIT / revenue Net profit (UBS) / revenue
5Yr Avg
NM -38.9 NM 5Yr Avg NM 20.9 5Yr Avg NM 5Yr Avg 2.9 3.1 29.0 5Yr Avg 2.6 2.3 0.4 5Yr Avg (49.8) NM -
12/07
7.9 -0.5 2.9 12/07 NM NM 2.0 12/07 NM 12/07 1.3 1.3 9.3 12/07 NM 2.0 0.2 12/07 (29.5) (41.8) (13.0)
12/08
12.4 0.1 2.6 12/08 0.1 0.1 1.9 12/08 NM 12/08 1.2 1.2 32.7 12/08 NM 1.0 0.1 12/08 (48.2) (93.1) (14.3)
12/09E
2.1 -16.0 NM 12/09E NM NM (9.9) 12/09E NM 12/09E 0.7 0.7 19.9 12/09E 0.5 1.2 0.1 12/09E (3.1) (3.2) (6.0)
12/10E
4.1 -18.3 NM 12/10E NM NM (14.2) 12/10E NM 12/10E 0.8 0.9 9.5 12/10E 0.0 1.1 0.0 12/10E (11.4) (12.8) (3.0)
12/11E
9.5 -8.0 NM 12/11E NM NM (6.6) 12/11E NM 12/11E 1.1 1.3 8.8 12/11E NM 0.9 0.1 12/11E (26.7) (36.3) (8.4)
Profitability
5.0% 0.0% -5.0% -5.0% -10.0% -15.0% -20.0% 12/07 12/08 12/09(E) 12/10(E) ROIC (RHS) 12/11(E) -10.0% -15.0% 5.0% 0.0%
Return on capital (%) EBIT ROIC (UBS) ROIC post tax Net ROE Coverage ratios (x) EBIT / net interest Dividend cover (UBS EPS) Div. payout ratio (%, UBS EPS) Net debt / EBITDA Efficiency ratios (x) Revenue / op. invested capital Revenue / fixed assets Revenue / net working capital Investment ratios (x) OpFCF / EBIT Capex / revenue (%) Capex / depreciation Capital structure (%) Net debt / total equity Net debt / (net debt + equity) Net debt (core) / EV
EBIT margin (LHS)
ROE v Price to book value
5.0% 0.0% -5.0% -10.0% -15.0% 4.0x 3.5x 3.0x 2.5x 2.0x
12/07
12/08 ROE (LHS)
12/09(E)
12/10(E) Price to book value (RHS)
12/11(E)
Growth (UBS EPS)
350 300 250 200 150 100 50 12/07 12/08 Revenue (LHS) 12/09(E) 12/10(E) UBS EPS Growth (RHS) 12/11(E) 60% 40% 20% 0% -20% -40% -60% -80%
Source: Company accounts, UBS estimates. (UBS) valuations are stated before goodwill-related charges and other adjustments for abnormal and economic items at the analysts' judgement. Valuations: based on an average share price that year, (E): based on a share price of US$43.76 on 21 Sep 2009 19:39 EDT Market cap(E) may include forecast share issues/buybacks.
Steven Eliscu
Analyst steven.eliscu@ubs.com +1-415-352 5674
Uche Orji
Analyst uche.orji@ubs.com +1 212 713 4015
UBS 44
NetLogic Microsystems Inc 22 September 2009
NetLogic Microsystems Inc NetLogic Microsystems is a fabless semiconductor company that designs, develops, and markets search processors, high-end embedded processors and high-speed physical layer products for networking systems including routers, switches, network access equipment, wireless infrastructure equipment, network security appliances, and networked storage devices. It provides search processing devices (knowledge-based processors or KBPs), NETLite processors (low-end KBPs), legacy network search engines (NSEs), high-end embedded microprocessors (through its RMI acquisition) and optical interface devices.
Statement of Risk NetLogic Microsystems company-specific risks include: 1) dependence on Cisco as its largest customer, which is likely to comprise at least one-third of sales, 2) Cisco's captive search processing efforts, which could weigh on growth, 3) dependence on the communications infrastructure market, which is likely to dominate its end-market sales, 4) potential risks related to the integration of recently acquired high-end embedded processor specialist, RMI, 5) addressing the historical difficulty of maintaining a leading position in highend embedded processors, 6) a multi-year new product gestation period, which limits its ability to make any mid-course changes to the company’s business model without seeing results for a least a couple of years. In addition, NetLogic faces the semiconductor sector risk where revenue trends have historically been linked to corporate profitability, which could be challenged for growth as UBS expects global GDP growth to remain below trend line for an extended period and end market growth, which is likely to remain depressed as vendors consolidate and the Chinese vendors continue to exert pressure on industry margins.
Analyst Certification Each research analyst primarily responsible for the content of this research report, in whole or in part, certifies that with respect to each security or issuer that the analyst covered in this report: (1) all of the views expressed accurately reflect his or her personal views about those securities or issuers; and (2) no part of his or her compensation was, is, or will be, directly or indirectly, related to the specific recommendations or views expressed by that research analyst in the research report.
UBS 45
NetLogic Microsystems Inc 22 September 2009
Required Disclosures
This report has been prepared by UBS Securities LLC, an affiliate of UBS AG. UBS AG, its subsidiaries, branches and affiliates are referred to herein as UBS. For information on the ways in which UBS manages conflicts and maintains independence of its research product; historical performance information; and certain additional disclosures concerning UBS research recommendations, please visit www.ubs.com/disclosures. The figures contained in performance charts refer to the past; past performance is not a reliable indicator of future results. Additional information will be made available upon request.
UBS Investment Research: Global Equity Rating Allocations
UBS 12-Month Rating Buy Neutral Sell UBS Short-Term Rating Buy Sell Rating Category Buy Hold/Neutral Sell Rating Category Buy Sell Coverage 44% 39% 17% 3 Coverage less than 1% less than 1%
1
IB Services 38% 36% 25% 4 IB Services 33% 33%
2
1:Percentage of companies under coverage globally within the 12-month rating category. 2:Percentage of companies within the 12-month rating category for which investment banking (IB) services were provided within the past 12 months. 3:Percentage of companies under coverage globally within the Short-Term rating category. 4:Percentage of companies within the Short-Term rating category for which investment banking (IB) services were provided within the past 12 months. Source: UBS. Rating allocations are as of 30 June 2009.
UBS Investment Research: Global Equity Rating Definitions
UBS 12-Month Rating Buy Neutral Sell UBS Short-Term Rating Buy Sell Definition FSR is > 6% above the MRA. FSR is between -6% and 6% of the MRA. FSR is > 6% below the MRA. Definition Buy: Stock price expected to rise within three months from the time the rating was assigned because of a specific catalyst or event. Sell: Stock price expected to fall within three months from the time the rating was assigned because of a specific catalyst or event.
UBS 46
NetLogic Microsystems Inc 22 September 2009
KEY DEFINITIONS Forecast Stock Return (FSR) is defined as expected percentage price appreciation plus gross dividend yield over the next 12 months. Market Return Assumption (MRA) is defined as the one-year local market interest rate plus 5% (a proxy for, and not a forecast of, the equity risk premium). Under Review (UR) Stocks may be flagged as UR by the analyst, indicating that the stock's price target and/or rating are subject to possible change in the near term, usually in response to an event that may affect the investment case or valuation. Short-Term Ratings reflect the expected near-term (up to three months) performance of the stock and do not reflect any change in the fundamental view or investment case. EXCEPTIONS AND SPECIAL CASES UK and European Investment Fund ratings and definitions are: Buy: Positive on factors such as structure, management, performance record, discount; Neutral: Neutral on factors such as structure, management, performance record, discount; Sell: Negative on factors such as structure, management, performance record, discount. Core Banding Exceptions (CBE): Exceptions to the standard +/-6% bands may be granted by the Investment Review Committee (IRC). Factors considered by the IRC include the stock's volatility and the credit spread of the respective company's debt. As a result, stocks deemed to be very high or low risk may be subject to higher or lower bands as they relate to the rating. When such exceptions apply, they will be identified in the Company Disclosures table in the relevant research piece.
Research analysts contributing to this report who are employed by any non-US affiliate of UBS Securities LLC are not registered/qualified as research analysts with the NASD and NYSE and therefore are not subject to the restrictions contained in the NASD and NYSE rules on communications with a subject company, public appearances, and trading securities held by a research analyst account. The name of each affiliate and analyst employed by that affiliate contributing to this report, if any, follows. UBS Securities LLC: Steven Eliscu; Uche Orji.
Company Disclosures
Company Name 16 NetLogic Microsystems Inc Reuters NETL.O 12-mo rating Short-term rating Not Rated N/A Price US$43.76 Price date 21 Sep 2009
Source: UBS. All prices as of local market close. Ratings in this table are the most current published ratings prior to this report. They may be more recent than the stock pricing date 16. UBS Securities LLC makes a market in the securities and/or ADRs of this company.
Unless otherwise indicated, please refer to the Valuation and Risk sections within the body of this report.
NetLogic Microsystems Inc (US$)
Price Target (US$) 60.0 50.0 40.0 30.0 20.0 10.0 01-Jan-05 01-Jan-06 01-Jan-07 01-Jan-08 01-Jan-09 01-Apr-05 01-Apr-06 01-Apr-07 01-Apr-08 01-Apr-09 01-Jul-04 01-Jul-05 01-Jul-06 01-Jul-07 01-Jul-08 01-Oct-04 01-Oct-05 01-Oct-06 01-Oct-07 01-Oct-08 01-Jul-09 0.0 Stock Price (US$)
Buy 2 Neutral 2 No Rating
Source: UBS; as of 21 Sep 2009
UBS 47
NetLogic Microsystems Inc 22 September 2009
Note: On August 4, 2007 UBS revised its rating system. (See 'UBS Investment Research: Global Equity Rating Definitions' table for details). From September 9, 2006 through August 3, 2007 the UBS ratings and their definitions were: Buy 1 = FSR is > 6% above the MRA, higher degree of predictability; Buy 2 = FSR is > 6% above the MRA, lower degree of predictability; Neutral 1 = FSR is between -6% and 6% of the MRA, higher degree of predictability; Neutral 2 = FSR is between -6% and 6% of the MRA, lower degree of predictability; Reduce 1 = FSR is > 6% below the MRA, higher degree of predictability; Reduce 2 = FSR is > 6% below the MRA, lower degree of predictability. The predictability level indicates an analyst's conviction in the FSR. A predictability level of '1' means that the analyst's estimate of FSR is in the middle of a narrower, or smaller, range of possibilities. A predictability level of '2' means that the analyst's estimate of FSR is in the middle of a broader, or larger, range of possibilities. From October 13, 2003 through September 8, 2006 the percentage band criteria used in the rating system was 10%.
UBS 48
NetLogic Microsystems Inc 22 September 2009
Global Disclaimer
This report has been prepared by UBS Securities LLC, an affiliate of UBS AG. UBS AG, its subsidiaries, branches and affiliates are referred to herein as UBS. In certain countries, UBS AG is referred to as UBS SA. This report is for distribution only under such circumstances as may be permitted by applicable law. Nothing in this report constitutes a representation that any investment strategy or recommendation contained herein is suitable or appropriate to a recipient’s individual circumstances or otherwise constitutes a personal recommendation. It is published solely for information purposes, it does not constitute an advertisement and is not to be construed as a solicitation or an offer to buy or sell any securities or related financial instruments in any jurisdiction. No representation or warranty, either express or implied, is provided in relation to the accuracy, completeness or reliability of the information contained herein, except with respect to information concerning UBS AG, its subsidiaries and affiliates, nor is it intended to be a complete statement or summary of the securities, markets or developments referred to in the report. UBS does not undertake that investors will obtain profits, nor will it share with investors any investment profits nor accept any liability for any investment losses. Investments involve risks and investors should exercise prudence in making their investment decisions. The report should not be regarded by recipients as a substitute for the exercise of their own judgement. Any opinions expressed in this report are subject to change without notice and may differ or be contrary to opinions expressed by other business areas or groups of UBS as a result of using different assumptions and criteria. Research will initiate, update and cease coverage solely at the discretion of UBS Investment Bank Research Management. The analysis contained herein is based on numerous assumptions. Different assumptions could result in materially different results. The analyst(s) responsible for the preparation of this report may interact with trading desk personnel, sales personnel and other constituencies for the purpose of gathering, synthesizing and interpreting market information. UBS is under no obligation to update or keep current the information contained herein. UBS relies on information barriers to control the flow of information contained in one or more areas within UBS, into other areas, units, groups or affiliates of UBS. The compensation of the analyst who prepared this report is determined exclusively by research management and senior management (not including investment banking). Analyst compensation is not based on investment banking revenues, however, compensation may relate to the revenues of UBS Investment Bank as a whole, of which investment banking, sales and trading are a part. The securities described herein may not be eligible for sale in all jurisdictions or to certain categories of investors. Options, derivative products and futures are not suitable for all investors, and trading in these instruments is considered risky. Mortgage and asset-backed securities may involve a high degree of risk and may be highly volatile in response to fluctuations in interest rates and other market conditions. Past performance is not necessarily indicative of future results. Foreign currency rates of exchange may adversely affect the value, price or income of any security or related instrument mentioned in this report. For investment advice, trade execution or other enquiries, clients should contact their local sales representative. Neither UBS nor any of its affiliates, nor any of UBS' or any of its affiliates, directors, employees or agents accepts any liability for any loss or damage arising out of the use of all or any part of this report. For financial instruments admitted to trading on an EU regulated market: UBS AG, its affiliates or subsidiaries (excluding UBS Securities LLC and/or UBS Capital Markets LP) acts as a market maker or liquidity provider (in accordance with the interpretation of these terms in the UK) in the financial instruments of the issuer save that where the activity of liquidity provider is carried out in accordance with the definition given to it by the laws and regulations of any other EU jurisdictions, such information is separately disclosed in this research report. UBS and its affiliates and employees may have long or short positions, trade as principal and buy and sell in instruments or derivatives identified herein. Any prices stated in this report are for information purposes only and do not represent valuations for individual securities or other instruments. There is no representation that any transaction can or could have been effected at those prices and any prices do not necessarily reflect UBS's internal books and records or theoretical model-based valuations and may be based on certain assumptions. Different assumptions, by UBS or any other source, may yield substantially different results. United Kingdom and the rest of Europe: Except as otherwise specified herein, this material is communicated by UBS Limited, a subsidiary of UBS AG, to persons who are eligible counterparties or professional clients and is only available to such persons. The information contained herein does not apply to, and should not be relied upon by, retail clients. UBS Limited is authorised and regulated by the Financial Services Authority (FSA). UBS research complies with all the FSA requirements and laws concerning disclosures and these are indicated on the research where applicable. France: Prepared by UBS Limited and distributed by UBS Limited and UBS Securities France SA. UBS Securities France S.A. is regulated by the Autorité des Marchés Financiers (AMF). Where an analyst of UBS Securities France S.A. has contributed to this report, the report is also deemed to have been prepared by UBS Securities France S.A. Germany: Prepared by UBS Limited and distributed by UBS Limited and UBS Deutschland AG. UBS Deutschland AG is regulated by the Bundesanstalt fur Finanzdienstleistungsaufsicht (BaFin). Spain: Prepared by UBS Limited and distributed by UBS Limited and UBS Securities España SV, SA. UBS Securities España SV, SA is regulated by the Comisión Nacional del Mercado de Valores (CNMV). Turkey: Prepared by UBS Menkul Degerler AS on behalf of and distributed by UBS Limited. Russia: Prepared and distributed by UBS Securities CJSC. Switzerland: Distributed by UBS AG to persons who are institutional investors only. Italy: Prepared by UBS Limited and distributed by UBS Limited and UBS Italia Sim S.p.A.. UBS Italia Sim S.p.A. is regulated by the Bank of Italy and by the Commissione Nazionale per le Società e la Borsa (CONSOB). Where an analyst of UBS Italia Sim S.p.A. has contributed to this report, the report is also deemed to have been prepared by UBS Italia Sim S.p.A.. South Africa: UBS South Africa (Pty) Limited (Registration No. 1995/011140/07) is a member of the JSE Limited, the South African Futures Exchange and the Bond Exchange of South Africa. UBS South Africa (Pty) Limited is an authorised Financial Services Provider. Details of its postal and physical address and a list of its directors are available on request or may be accessed at http:www.ubs.co.za. United States: Distributed to US persons by either UBS Securities LLC or by UBS Financial Services Inc., subsidiaries of UBS AG; or by a group, subsidiary or affiliate of UBS AG that is not registered as a US broker-dealer (a 'non-US affiliate'), to major US institutional investors only. UBS Securities LLC or UBS Financial Services Inc. accepts responsibility for the content of a report prepared by another non-US affiliate when distributed to US persons by UBS Securities LLC or UBS Financial Services Inc. All transactions by a US person in the securities mentioned in this report must be effected through UBS Securities LLC or UBS Financial Services Inc., and not through a non-US affiliate. Canada: Distributed by UBS Securities Canada Inc., a subsidiary of UBS AG and a member of the principal Canadian stock exchanges & CIPF. A statement of its financial condition and a list of its directors and senior officers will be provided upon request. Hong Kong: Distributed by UBS Securities Asia Limited. Singapore: Distributed by UBS Securities Pte. Ltd or UBS AG, Singapore Branch. Japan: Distributed by UBS Securities Japan Ltd to institutional investors only. Where this report has been prepared by UBS Securities Japan Ltd, UBS Securities Japan Ltd is the author, publisher and distributor of the report. Australia: Distributed by UBS AG (Holder of Australian Financial Services License No. 231087) and UBS Securities Australia Ltd (Holder of Australian Financial Services License No. 231098) only to 'Wholesale' clients as defined by s761G of the Corporations Act 2001. New Zealand: Distributed by UBS New Zealand Ltd. An investment adviser and investment broker disclosure statement is available on request and free of charge by writing to PO Box 45, Auckland, NZ. China: Distributed by UBS Securities Co. Limited. Dubai: The research prepared and distributed by UBS AG Dubai Branch, is intended for Professional Clients only and is not for further distribution within the United Arab Emirates. The disclosures contained in research reports produced by UBS Limited shall be governed by and construed in accordance with English law. UBS specifically prohibits the redistribution of this material in whole or in part without the written permission of UBS and UBS accepts no liability whatsoever for the actions of third parties in this respect. © UBS 2009. The key symbol and UBS are among the registered and unregistered trademarks of UBS. All rights reserved.
ab
UBS 49