November 22, 2015 IT Hardware, HDD/SSD, & Mobile Weekly

IT Hardware, HDD/SSD, & Mobile Weekly

1

November 22, 2015

SC’15 Highlights Positive HPC / Supercomputing Trends; Strong Cray Positioning, China 4G / TD-LTE Subs Exits Oct. at 25.2%, Tsinghua (China) Chip Plans & More…

Aaron C. Rakers, CFA [email protected] (314) 342-8401 Joseph Quatrochi [email protected] (314) 342-2180 Andrew Shinn, CFA [email protected] (314) 342-2792

This Week’s Highlights:

SC’15 Conference - Positive Trends in HPC / Supercomputing: Last week we attended the SC’15 conference in Austin, TX – we are once again impressed by the level of attendance (+12,900 attendees) and the detailed content on next-gen. technologies (e.g., Intel Omni-Path, Xeon Phi Knights Landing, ARM CPUs in HPC, evolving strong hierarchy, etc.). The conference reinforced our positive views on growth trends in the HPC / supercomputing market, and Cray’s strong position. IDC estimated that the total HPC market would grow to $11.4B in 2015, +12% y/y. Cray’s flagship high-end supercomputing market is estimated to grow at a ~10% CAGR from 2014-2019, stronger than our prior +7%-8% estimate. The commercial adoption of HPC / supercomputers remains a key growth driver over the coming years (reminder: Cray is targeting a +15% commercial contribution in 2015, growing to 30%+ of total revenue over the coming years). A study found by IDC and DOE

estimated that the ROI on HPC deployments stands at ~$515 in revenue for every $1 spent on HPC deployments; +$40 in estimated profits per $1 spent. Intel noted that HPC is still utilized by less than 10% of U.S. manufacturing firms.

Cray Update – Positive: As highlighted in our reports last week and reiterated below, we believe last week’s SC’15 conference offered up several positive takeaways for Cray: 1.) investor meeting (not widely attended) highlighted the company’s confidence in its growth expectations – continued growth at 2x the overall market, or 15%-20% y/y; 2.) Cray’s increasing traction / contribution in the commercial market; 3.) Cray’s positioning to capitalize on the aging installed base of Blue Gene/Q systems – last week’s win announcement at the University of Warsaw represents the first announced replacement of a Blue Gene/Q system; 4.) our expectation of meaningful product cycles going into 2016 – particular focus on Intel’s Xeon Phi Knights Landing, as well as positioning for the ongoing evolution in storage; 5.) Cray’s strong showing in the updated (Nov ’15) Top500 list of global supercomputers – Cray is #1 in top 100, 50, and 10 global supercomputers, PGS deployment is the largest commercial supercomputer deployment, Cray leads in terms of share of supercomputer GPUs deployed (48.5% share), and Cray deployments are the only two new additions to the top 10 list, five of the only six new additions within the top 50, and six out of the 15 new additions to the top 100.

China 4G / TD-LTE Subscribers Base Grows by 26M in October (Apple): Late last week it was reported that China’s total 4G / TD-LTE subscriber base exited October totaling 328.38 million total subscribers, or reflecting the addition of 25,986 million subscribers when compared to September. This represents the fifth consecutive month in which new additions have trended between 25-27 million. China’s 4G / TD-LTE subscriber base now stands at ~25.2% of total mobile phone subscribers, or rather leaving ~418.9 million 3G subscribers and an implied ~554.5 million total 2G / other subscribers.

Other: 1.) Reuters published an interview of Tsinghua’s Chairman highlighting China’s aspirations in the chip industry, 2.) Nimble results highlight growth vs. profitability challenges in an increasingly disruptive / competitively aggressive enterprise storage market, 3.) given the focus on the WD / SanDisk combination, Intel’s comments last week on increased NAND price pressure in 2016 weighed on WD shares.

Past performance should not and cannot be used as an indicator of future results

Stifel 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. All relevant disclosures

and certifications appear on pages 28-29 of this report.

mailto:[email protected]



ENTERPRISE HARDWARE, HARD DISK DRIVE & MOBILE WEEKLY (NOVEMBER 22, 2015) STIFEL EQUITY RESEARCH

2

Highlighted Industry News/Thoughts:

SC’15 Conference – Stifel Quick Takeaways: Positive Market Growth Trends; Cray Remains Very Well Positioned Last week we attended the SC’15 conference in Austin, Texas – a conference that continues to impressive us in terms of the size and content (+12,900 total attendees). Below we provide some quick notes on our takeaways from the conference:

Attending the SC’15 conference left us with a positive view on the trends in the high-performance computing (HPC) market; focus remains on commercial adoption as a positive catalyst for Cray going forward. From a commercial perspective, IDC’s announcement that a study in conjunction with the DOE equated that the return on investment for HPC deployments could be considered at ~$515 for every $1 invested in HPC solutions; estimating more than $40 of profits per $1 HPC spend. IDC currently estimates the HPC market to grow a solid 12% y/y in 2015 to $11.4 billion – noting that there have been continued signs of broad strength in the HPC market. Consistent with our own view and Cray’s strategic vision, IDC believes there will continue to be accelerating trends around the convergence between high-performance computing and data analytics.

Cray hosted an investor event in conjunction with SC’15, which we believe (as summarized below – see Cray section) provided a positive review of the company’s strong / strengthening positioning in the supercomputing market – focus on expanding commercial presence as a key driver of the company’s expected growth rate at 2x the addressable market (i.e., 15%-20% y/y). Cray won six awards by HPCWire at this year’s conference – including top-5 vendor to watch, best HPC server product / technology (XC40 supercomputer), and best HPC software (OpenACC parallel programming for CPU/GPU systems).

IDC estimates that Cray had a ~14% share in the overall HPC market in 1H2015 – a market that totaled an estimated ~$1.5 billion. While this would look notably different if separating out the very high end of the market (i.e., non-clustered solutions), we would note that HP was estimated to be leading the market with a 38% share, followed by Lenovo at 12%. HP did not make meaningful announcements at SC’15, but rather emphasized the positioning of its Apollo clustered servers for the HPC market.

Cray’s CTO, Steve Scott, highlighted evolution architectures going forward, in particular: 1.) the evolution of on-chip memory package going forward, and 2.) the company’s opportunity to differentiate around the increasingly complex hierarchies of data storage. Cray believes Flash will continue to see increasing acceptance in primary storage environments, though HDDs are not likely going away in the foreseeable future. The role of Micron / Intel’s 3D X-Point, as well as next-generation technologies such as ReRAM, STT-RAM, PCM, etc., should become increasingly relevant over the next 3-5 years.

Discussions at the conference highlighted the view that ARM will play a role in the HPC market over the coming years. E4 Computer Engineering last week announced the public availability of the Cavium ThunderX-based ARKA servers. Gigabyte also announced the availability of ThunderX-based servers.

Top500.org published updated (November 2015) list of Top 500 supercomputers. As summarized below (see Cray section), the list reflected a strong / strengthening position for Cray.

A keynote address by Intel’s Diane Bryant emphasized the need for HPC in commercial environments – noting that less than 10% of U.S. manufacturing firms currently utilize the power of supercomputers.

Intel (INTC: $34.66, Buy, covered by our colleague Kevin Cassidy) officially launched its Gen 1 Omni-Path fabric architecture solutions – positioned against Mellanox’s (MLNX: $45.77, Buy, covered by our colleague Kevin Cassidy) EDR InfiniBand for HPC interconnect technology. As expected, and widely touted by Intel for some time, the solutions leverage the acquisition of QLogic’s TrueScale InfiniBand technology, as well as integrate small pieces of the Aries interconnect technology elements acquired from Cray in 2012. Our discussions with Intel highlighted a key difference in error correction capabilities, or rather what it views to be an advantage in terms of the latency impact of Mellanox’s method of error correction.

There was a lot of discussion around Intel’s next generation Xeon Phi technology codenamed Knights Landing - a product cycle that will be a key focus for Cray and the overall supercomputing market as we look into mid/2HC2016. Cray showcased modules with the Knights Landing Xeon Phi C/GPUs at its booth. Cray continues to highlight Knights Landing as representing an important product cycle for the company looking into 2HC2016; notes that Knights Landing could be considered a competitor to NVIDIA’s GPUs (including upcoming Pascal product launch), though NVIDIA (NVDA: $31.39, Buy, covered by our colleague Kevin Cassidy) is likely to remain the performance leader.

One of the noted biggest stories coming out of the SC’15 supercomputing conference was the announcement of Intel’s OpenHPC foundation. While product announcements are important, the longer-term impacts usually come from standards, mergers, and new groups. It was reported that Intel needs its partners in OpenHPC as much as its partners need Intel – focus on stimulating demand / acceptance of HPC / supercomputing in the commercial market.


3

We believe the role of HPC / supercomputing will continue to become a global topic of national strategic importance. The U.S. has the National Strategic Computing Initiative with an initial allocation of $450 million announced in late July, while China’s President Xhi has continued to highlight the positioning of the Tianhe-2 as the world’s fastest supercomputer. The European Commission announced that it has launched a contractual commitment to bolster its positioning in the HPC market – last week announcing the launch of an IPCEI on HPC (Important Project of Common European Interest on Higher Performance Computing and Big Data-enabled Applications).

Tsinghua Unigroup Chairman Comments on China’s Chip Aspirations – Focus on NAND Flash Production; Implications for WD / SanDisk Combination As we had highlighted in our detailed deep dive report last week on the Western Digital / SanDisk combination, China’s chip (particularly NAND Flash) aspirations and the ties with Western Digital will be a key focus. Early last week Reuters published an article from an interview with Tsinghua Unigroup Ltd’s Chairman, Zhao Weiguo, that highlighted the company’s, and thus China’s, chip aspirations. As published in our report last week, it was announced that Tongfang Guoxin, which is owned by Tsinghua, is planning to spend up to $14 billion in a new NAND Flash fab, estimated to equate to ~120,000 wafer starts per month (wspm) – leaving us / investors to focus on the implications of incremental NAND Flash supply (along with Intel’s investment plans in China) and how China plans on attaining the needed equipment and intellectual property for volume production. Key takeaways from the Reuters article include:

The article notes that Mr. Weiguo stated that the company has been in talks with a U.S.-based company involved in the chip industry – discussions that could culminate in a deal to be finalized as early as the end of this month. He declined in providing specifics, but it was stated that buying a majority state was unlikely as it was too sensitive for the U.S. Government.

The article notes that Tsinghua’s plans to build a fab could be considered a platform enabling the acquiring of intellectual property needed to manufacture chips.

Mr. Weiguo stated that if you can be the top-three giant in the chip industry it would be tough to develop a viable business in the chip industry.

The article notes that the top-5 chip producers control 90%+ of the total NAND Flash industry.

The article notes that China is focused on its reliance on foreign company semiconductors as it seeks to build a modern, digitized armed forces presence capable of matching other advanced militaries, while also looking at the strategic importance of servers and networking given fears of cyber spying – considerations that will certainly continue to leave us / investors to question how China plans to get the technology needed to produce NAND flash and / or other key semiconductors.

The article notes that Tsinghua made an informal $23 billion bit for Micron Technology in August – a deal that was immediately rejected and noted as a concern regarding the endangerment of national security.

The article notes that Tsinghua has no plans for DRAM. Intel Analyst Day – Quick / Relevant Takeaways Last week Intel hosted its Investor Day. While we did not listen to all of the details from the event, we believe the following represent the important highlights for our coverage universe:

Data Center Group (DCG): Intel updated its DCG 2015 CPU-specific revenue growth estimates for enterprise server to be slightly down y/y, HPC / networking / storage to be up approximately 10% y/y, and cloud server to be up +40% y/y. Intel reported that cloud solutions accounted for 20% of total DCG segment revenue in 2013, which is expected to grow to more than 30% in 2016. Overall DCG is expected to grow at a ~15% CAGR from 2015-2019 (consistent with prior estimates) while DCG CPU revenue is expected to grow at a 12% CAGR. Non-CPU related DCG revenue is expected to grow much faster and bring overall DCG revenue growth up to 15% CAGR. Recall that Intel reduced its DCG revenue growth expectations for 2015 to the low double digit y/y range from prior expectations of +15% y/y, reflecting continued macro-related weakness in traditional enterprise (albeit noted as improving in 3Q15), offset by stronger-than-expected demand in cloud. Intel’s Data Center Growth (DCG) segment revenue growth reaccelerated to +12% y/y in 3Q15 vs. +10% y/y in 2Q15.

Intel expects overall DCG revenue growth in the mid-teens range in 2016. Opportunities for growth in Intel’s data center business should continue to be driven primarily by cloud service providers, enterprise IT, telecom providers, and HPC. Intel expects both telecom and cloud service providers to grow at a +20% CAGR from 2015-2019, while gov’t / academia / science is expected to grow at a ~20% CAGR and enterprise IT at <5% CAGR from 2015-2019.

HPC / Omni-Path: Intel highlighted that its Omni-Path interconnect will commence production in December and offer 10% better performance, 60% lower power, and 20% lower system costs than InfiniBand EDR. Intel expects that Omni-Path will target a $1.6 billion TAM in 2020.


4

3D XPoint: Intel reiterated that it expects to commence sampling 3D XPoint DIMMs in 2016 and that 3D XPoint will provide 4x memory capacity at half the cost of DRAM. We continue to believe that 3D XPoint will be complementary to both DRAM and NAND but will be more cannibalistic to DRAM. Regarding silicon photonics, Intel notes that samples are currently shipping and it enables highest port density with +20% cost advantage vs. competitive networking technology. Intel expects 3D XPoint will target a $34 billion TAM in 2020 and that silicon photonics will target a $5 billion TAM in 2020.

3D NAND Flash: Intel reiterated that its investment in its new fab in Dalian, China, will be focused on 3D NAND but that it can be converted relatively easily to 3D XPoint as well if demand ramps more quickly than expected for 3D XPoint. Intel noted that its JV with Micron is continuing to work well but that Intel decided to convert Dalian from older-gen CPUs to 3D NAND because the factory was already in existence and it made sense to transition away from the older-gen CPUs and Intel sees significant opportunities in both 3D NAND and 3D XPoint and wanted to get in position to meet increased demand in future years. Intel continues to expect robust revenue growth in its NAND business, but NAND profitability will decline in 2016 due to increased price competition as well as the significant investments Intel is making into its Dalian fab for instance.

Client Computing Group (CCG): Intel expects the PC market to be slightly down y/y in 2016 vs. what it cites as current IDC / Gartner estimates at flat y/y for 2016. Intel is guiding to flat to low single-digit y/y CCG revenue growth in 2016, driven by mix shift to higher ASPs (i.e., 2015 core mix above 70% or at an all-time high). Intel highlights that its CCG is an important segment to its overall business as CCG generates ~$8 billion in operating profit, enables Intel to achieve economies of scale (CCG accounts for +80% of total units) and CCG funds IP for use across all segments (i.e., +50% of share IP originates from CCG R&D investments). Intel estimates there are ~1 billion PCs (i.e., desktops and notebooks) in the global installed based that are +3 years old; specifically, ~600 million PCs that are 3-5 years old, just under 300 million that are 5-7 years old, and just under 100 million that are +7 years old. Intel estimates that there are just under 300 million PCs that are less than 1 year old.

Capex: Intel expects to spend ~$10 billion in capex in 2016 (+/- $500 million) including ~$1.5 billion for memory (i.e., 3D NAND and 3D XPoint).

Dell Announces New HPC-Focused Solutions Last week we attended the SC’15 conference in Austin, TX (+12,900 attendees), from which Dell made several announcements. The company announced its HPC Systems Portfolio (clustered configurations) including: Dell HPC System for Genomic Data Analysis, Dell HPC Systems for Manufacturing, and Dell HPC Systems for Research. Dell, like others, announced their strategic partnership with Intel, which includes the usage of Intel’s Omni-Path fabric technology (i.e., competing against Mellanox’s EDR InfiniBand solutions). This includes new Dell Networking H-Series switches and adapters for the PowerEdge servers using the Omni-Path architecture. Dell also announced support for OpenHPC that is focused on designing / leveraging a common platform for HPC deployments in commercial environments. Dell also announced its partnership / collaboration with Mellanox. From a systems deployment perspective, Dell announced the Comet petascale system at San Diego Supercomputer Center (SDSC), the deployment of a system at Texas Advanced Computing Center (TACC), and the Jetstream system scheduled for production at Indian a University in early 2016. An interesting note from the conference was the estimate that less than 10% of the 300,000 manufacturing companies in the U.S. currently utilize HPC systems and that the National Center for Manufacturing estimates that 98% of all products will be digitally designed by 2020. China 4G / TD-LTE Subscriber Base Increases by 26.0 Million in October; 25.2% of Total Mobile Phone Subs Late last week China’s October mobile phone subscriber base data was published. Some of the key highlights include:

China exited October with a total of 1.302 billion total mobile phone subscribers, a modest increase from the 1.300 billion reported exiting September and up from 1.286 billion total phone subscribers exiting 2014.

4G / TD-LTE subscribers totaled 323.4 million exiting October, an increase from 302.4 million exiting September (+25.99 million m/m) and up from 97.3 million exiting 2014. This brings total 4G subscriber penetration to approximately 25.2% exiting October. As shown in the chart below, we continue to see a consistent monthly increase in China’s 4G / TD-LTE subscriber base over the past five quarters (+25-27 million per month).

Total 3G mobile phone subscribers totaled 418.9 million exiting October, down from 427.4 million exiting September and versus 485.3 million exiting 2014. This equates to 32.2% of total subscribers and leaves 3G + 4G subscribers at 57.4% of China’s total mobile phone subscribers. 3G + 2G mobile phone subscribers still total 973.4 million exiting October, down from the 1.189 billion exiting 2014.

Gartner estimates that Apple’s iPhone shipments into Mainland China totaled 11.2 million in the September quarter, up 84% y/y. For the first three quarters of 2015 Apple’s estimated iPhone shipments into China totaled ~38.0 million, which would equate to ~16.4% of the total new 4G / TD-LTE subscribers added during 2015 – positive, though we believe some could focus on the fact that this measure of relative share has declined over the past few quarters – see charts below.


5


6

Intel Officially Launches Omni-Path at SC’15 Conference Intel last week officially launched the availability of its 1

st generation Omni-Path next-generation interconnect technology.

The technology leverages the integration of the acquired QLogic TruScale InfiniBand technology, as well as a bit of capabilities from the purchase of Cray’s Aries interconnect technology (most notably error correction, per our discussions at the conference). In conjunction with the announcement, Cray announced its first CS-series clustered system win utilizing Intel’s Omni-Path technology. Per our conversations, and as expected, Omni-Path is becoming positioned against Mellanox’s EDR InfiniBand technology. Our discussions suggest that the technology is focused on reducing some of the latency attributes of InfiniBand when doing error correction.

Our discussions with Cray on the positioning of Omni-Path highlighted: 1.) a cheaper – though unquantified – viable alternative for InfiniBand used in their CS-series solutions, 2.) tighter integration with the Cray Aries interconnect technology would likely come in the second generation Omni-Path offerings (we believe a 2018 story), 3.) Cray is tightly working in conjunction with Intel on these next-generation interconnect capabilities – leveraging Intel’s ability to drive down costs through volume, while maintaining Cray-specific differentiation, and 4.) the first generation Omni-Path could be positioned as a lower performance solution relative to leading edge InfiniBand for clustered configurations. China PC Production Declines 11% y/y in October As a follow-up to the weak Taiwanese PC ODM data points published last week, we would note that data on PC production coming out of our tracking of China data points also highlights weakness in October. For October, China total micro-computer production totaled 29.07 million, down 11% y/y and down 2% when compared to September. This compares to -10%, -22%, -12%, and -7% y/y in June, July, August, and September, respectively. Notebook production totaled 14.6 million, down 48% y/y and down 13% when compared to September. The first chart below illustrates the monthly total China micro-computer production, while the second chart shows the monthly notebook production within micro-computers for the data we have tracked since 2013. While have not yet seen the data for October exports of total portable PCs + tablets, the final chart illustrates these monthly trends through September 2015 (expect October data over the next week or so).


7

Enterprise Hardware, Hard Disk Drive & Mobile Coverage Summary

Price % Ch. % Ch. % Ch. % Ch. % Ch. Forward Forward MC/Sales EV/Sales EV/Sales EV/EBITDA

Ticker 11/20/2015 YTD 12-Months 3-Months 1-Month 1-Week P/E (C2015)* P/E (C2016)** (Trailing) (Trailing) C2016 (Trailing)

Apple (Initiate Buy = 8/8/12) AAPL $119.30 8.1% 4.0% 3.7% 6.8% 6.2% 12.84 11.80 2.88 2.97 2.70 8.14

EMC (Buy to Hold = 10/21/15) EMC $25.33 -14.8% -15.4% -3.1% -8.6% 1.2% 14.36 13.42 1.98 2.04 1.97 10.13

NetApp (Buy to Hold = 11/7/13) NTAP $31.05 -25.1% -25.7% 4.3% -7.8% -0.8% 13.96 10.62 1.57 0.97 0.94 7.35

Nimble Storage (Buy to Hold = 11/19/15) NMBL $10.05 -63.5% -62.2% -60.8% -55.7% -51.8% N/A N/A 2.66 1.95 1.27 N/A

Pure Storage (initiate Buy = 11/2/15) PSTG $14.74 N/A N/A N/A N/A -14.5% N/A N/A 15.95 5.57 1.55 N/A

HP Inc. (Reiterate Hold = 10/30/15) HPQ $14.06 -22.8% -16.7% 11.6% 6.9% 7.2% 8.00 7.68 0.46 0.51 0.55 N/A

Hewlett Packard Enterprise (Initiate Hold = 10/30/15) HPE $14.21 N/A N/A N/A N/A 6.1% 7.58 7.52 0.49 0.38 0.40 N/A

Super Micro Computers (Hold to Buy = 5/15/14) SMCI $23.40 -32.9% -30.0% -14.7% -11.2% -8.2% 11.06 9.73 0.54 0.53 0.44 7.52

Cray Inc. (Initiate Buy = 8/25/14) CRAY $34.24 -0.7% 5.5% 54.9% 58.1% 4.9% 26.82 19.91 1.93 1.63 1.42 15.90

Brocade (Buy to Hold = 2/23/10) BRCD $9.85 -16.8% -13.1% -1.9% -3.0% -1.4% 9.98 9.40 1.83 1.59 1.52 6.09

QLogic (Buy to Hold 7/27/12) QLGC $12.10 -9.2% 6.4% 30.0% 12.2% -0.9% 13.56 13.47 2.10 1.44 1.53 7.86

CommVault (Hold to Buy = 10/31/12) CVLT $40.16 -22.3% -15.9% 7.2% 14.1% 3.5% 52.78 35.57 3.11 2.43 2.26 N/A

Box (Initiate Hold = 9/24/15) BOX $12.90 N/A N/A -12.5% -4.8% 3.1% N/A N/A 7.23 6.33 3.58 N/A

Seagate (Buy to Hold = 8/20/12) STX $34.41 -48.3% -46.8% -30.4% -11.2% 4.1% 10.92 8.59 0.80 0.97 1.04 4.85

Western Digital (Hold to Buy = 4/22/10) WDC $61.71 -44.3% -38.2% -24.5% -23.3% 2.5% 9.45 7.84 1.02 0.82 0.82 4.49

IBM (Buy; David Grossman) IBM $138.50 -13.7% -14.2% -10.0% -7.2% 5.1% 9.27 9.17 1.60 1.98 2.04 7.41

Cisco (Buy; Sanjiv Wadhwani) CSCO $27.57 -0.9% 3.7% -0.8% -2.4% 5.2% 12.33 11.82 2.83 2.12 2.11 7.19

Juniper (Hold; Sanjiv Wadhwani) JNPR $30.46 36.5% 41.9% 8.6% 1.7% 6.1% 15.33 13.95 2.53 2.53 2.14 N/A

F5 Networks (Buy; Sanjiv Wadhwani) FFIV $103.59 -20.6% -18.0% -22.0% -12.6% 1.6% 15.41 14.26 3.78 3.39 2.96 10.61

-20.0% -15.9% -4.0% -3.4% -1.5% 15.39 12.86 2.91 2.07 1.59 8.13

-20.6% -15.9% -2.5% -6.0% 2.8% 12.33 10.62 1.96 1.79 1.53 7.41

S&P500 SP50 $2,089.17 1.5% 2.0% 0.5% 2.7% 3.3%

S&P Hardware, Storage & Peripherals SP024 $139.12 -1.0% -2.9% 2.4% 3.9% 5.7%

S&P Communications Equipment SP749 $128.21 -12.7% -8.6% -6.9% -6.4% 2.0%

S&P Systems Software SP733 $1,118.18 7.5% 7.6% 11.0% 11.2% 3.3%

S&P IT Consulting & Services SP721 $138.74 -1.2% -0.1% -4.9% -2.7% 3.9%

S&P Semiconductors SPT74 $545.75 -1.3% 3.8% 13.5% 3.5% 5.7%

* Forward P/E on C2015 = On Stifel Estimates

** Forward P/E on C2016 = On Stifel Estimates

Source: FactSet; Stifel Estimates

Valuation on Consensus/TrailingShare Performance


8

Noteworthy Company-Specific News/Thoughts Review

Apple (AAPL – NASDAQ - $119.30 – Buy)

Mainland China Non-Android (iOS?) Smartphones +68.5% Y/Y in October In addition to the production data we published in our Weekly report last week, over the weekend our tracking of mainland China smartphone industry data would find / estimate China smartphone shipments at ~32.967 million in October 2015, up 17% y/y, but down 12% from ~37.3 million units in the prior month. China 4G/LTE mobile phone shipments were estimated to have accounted for 87.7% of total mobile phone shipments in October 2015, implying ~32.6 million units shipped, +53% y/y and -11% from the prior month.

Apple Positive: Non-Android Smartphones Increased 68.5% Y/Y. Within the data we track, we would note that Android smartphones are estimated to have accounted for approximately 77% of total October 2015 China domestic smartphone shipments, a decline from 85.1% in September 2015 and 84% in October 2014. This would imply Android shipments up 7% y/y, but down 20% m/m. The implied total non-Android (Apple iPhone = majority?) shipments were ~7.5 million in October 2015, up 68.5% y/y and up 34% from the prior month and marking the first full month of iPhone 6s/+ availability.


9

Apple Looking for Second Watch Manufacturing Partner; Expecting 4M Units in 4Q15 (vs. initial ~6M Unit Estimate) In conjunction with reports that Apple could launch Watch 2 in June 2016 it was reported that the company is now looking for a second manufacturing partner for the Watch. DigiTimes reported that Apple has placed orders for 4 million Apple Watches in the fourth quarter – this was noted as comparing to the company’s initial expectation of 6 million units. The report notes that Apple expects to surpass 10 million Watch sales in 2015. The company is said to be discussing production at Inventec, Wistron, and Foxconn in addition to maintaining its manufacturing relationship with Quanta.

EMC Corporation (EMC – NYSE - $25.33 – Hold)

EMC Announces New Solutions to Improve Integration of On-Premise VMAX and VNX Storage with Public Clouds (vCloud Air, AWS, Azure, and GCP) EMC announced last week that it launched a range of solutions to connect primary storage systems to private and public clouds so that customers can more easily scale workloads. In addition, EMC launched new data protection features to ensure that data is protected regardless of where it resides. Utilizing EMC CloudArray, customers can connect a VMAX to a network switch and immediately connect to storage targets either in an on-premise private cloud or public cloud. The solution works with VNX combined with VPLEX. EMC now supports VMware’s vCloud Air, AWS, Azure, and GCP (IBM’s SoftLayer was not named). EMC’s CloudBoost 2.0, which provides 3x bandwidth and 15x more capacity than the prior version, extends data protection seamlessly to the cloud. EMC’s Spanning solution is a backup / restore solution that now has regional deployment so customers can comply with European data sovereignty laws. Recall that two weeks ago EMC announced its CloudPools, a new feature for Isilon, that enables customers to extend cold data to public clouds without the need for a cloud gateway.

EMC Job Listings Update As of last week, EMC’s job listings totaled 451, which is up 35 listings after declining 969 listings since the announcement of Dell’s acquisition of EMC. Within this total, there were 23 total core EMC listings, down 1 from the number of listings last week. Sales listings totaled 128 openings, up 13 from last week and compared to 228 entering 2015. Total professional services listings were 73, up 8 from the prior week, while software engineer listings totaled 135, up 25 from the prior week. EMC’s Isilon job openings totaled 7, unchanged when compared to the prior period. EMC lists 1 opening for DSSD, down 1 from last week and compared to ~5 listings prior to the Dell/EMC announcement. The figure below highlights EMC’s job listings over the past several months.


10

NetApp Inc. (NTAP – NASDAQ - $31.05 – Hold)

Focus on FCF / Deferred Revenue Trends Continue – Reduce Estimates; Maintain Hold

Conclusion: While NetApp reported F2Q16 EPS upside ($0.61/sh. vs. our $0.58/sh. estimate and street at $0.57; guide at $0.55-$0.60) on in-line revenue at $1.445B (vs. our $1.45B estimate; $1.4-$1.5B guide), and the company’s F2016 guide was only slightly tempered given macro and competitive pressures offset by what looks to be upcoming significant opex realignment efforts, we continue to question some of the moving parts in the NetApp model, most notably: 1.) gradual Clustered ONTAP adoption (17% installed base penetration vs. 15% exiting F1Q16, which includes new customers that grew 95% y/y – suggesting a very slow / declining traditional installed base), product GM% at 51.8% (-6.6pp y/y), 2.) anticipated over the company’s updated F2016 FCF expectation (i.e., 2HF2016 FCF now guided at 15% vs. prior full F2016 guide at 15%, or implying a 20%+ decline in expectation), and 3.) NetApp’s continued decline in deferred revenue (flat y/y, but -$20M seq. albeit considering 14-week period in F1Q16); revenue contributions via deferred revenue recognition stands at 38.2% of NetApp’s total TTM revenue. Maintenance / Support revenue at $326M reflects a sharp y/y growth deceleration; note that seq. decline at 5.8% needs to be adjust for ~$23M impact from 14-week F1Q16.

In-Line F3Q16 Guide; Tempered F2016: F3Q16 guide at $1.4-$1.5B/$0.66-$0.71 vs. our prior $1.49B/$0.71 estimate (street: $1.48B/$0.70). W/ increased macro uncertainties, increased competitive / discounting dynamics, and a gradual long-term Clustered ONTAP installed base uptake, NetApp has slightly tempered its F2016 outlook – now estimating revenue to decline just over 5% y/y vs. prior -5% y/y. Non-GAAP GM% is expected to be down 1-2pp y/y (vs. prior -1pp guide), while operating / EBIT margin is expected to be -2pp y/y (in-line w/ prior guide; ~18% for 2HF2016 vs. prior 18%-20% range). We adjust our F2016 and F2017 estimates from $5.842B/$2.42 and $6.088B/$2.87 to $5.764B/$2.37 and $5.753B/$2.70, respectively.

Additional Noteworthy Metrics: 1.) Clustered ONTAP: 17% of installed base vs. 15% exiting F1Q16, which includes a 95% y/y increase in new-to-NetApp customers; clustered ONTAP accounts for 70% of FAS systems shipped. 2.) All-Flash: Unit ship +445% y/y vs. +140% y/y in F1Q16; EF-series +65% y/y w/ remaining growth driven by all-Flash FAS (note: all-Flash FAS intro was in F1Q15), 3.) FCF: $99M in F1Q16; $795.4M on TTM basis or 13.5% of revenue (lowest TTM % we have seen since F2005). 4.) Cash: $4.814B in total cash and investments (14% held domestic); net cash at $3.325B or $11.23/sh.; leaves us at current ~7.4x EV/FCF (TTM) or 13% yield.


11

Please see our detailed report published on 11/18/15 for additional information.


12

NetApp Highlights All-Flash FAS Pricing Below Disk; Free Controller with 3-Year Support Contract Renewal Similar to Pure’s Evergreen NetApp announced last week that its all-flash FAS storage arrays include new software efficiency technology that can result in data reduction between 5:1 and 30:1 in VDI environments. We assume this includes not only NetApp’s inline and post-process dedupe and inline compression but also thin provisioning. We would also note that NetApp concedes that because of the synthetic nature of the data used to perform the tests, the 5:1 and 30:1 numbers are not typical of real-world savings. Nevertheless, NetApp is highlighting that the new all-flash FAS arrays drive down the cost of the arrays at least below the price of disk. More importantly, NetApp highlighted that its enterprise-grade features make its all-flash systems ready for broad deployment and that NetApp’s all-flash system is the only one in the market to support seamless data management across flash, disk, and cloud resources. NetApp highlighted that its all-flash FAS systems now support 3.8TB SSDs and that eligible customers who purchase a three-year SupportEdge Premium contract by December 31, 2015 now qualify for a new all-flash FAS controller free of charge as part of the renewal process. This is very similar to Pure’s Evergreen / Forever Flash program, although Pure’s offer does not expire at the end of 2015. NetApp Job Listings Update NetApp’s job listings totaled 392, up 1 from the prior week, vs. 471 entering 2015. The company currently has 90 sales openings, down 2 from the prior week. We continue to gauge the secular / architectural changes playing out in the storage industry vs. additional sales capacity relative to NetApp’s ability to reaccelerate growth. The company has 94 openings for software engineers, which is up 10 from the prior period. Customer support engineering listings totaled 53, up 5 from the prior week and vs. 78 openings entering 2015. The figure below highlights NetApp’s employee job listing trends over the past several months.

Nimble Storage (NMBL – NYSE - $10.05 – Hold)

Downgrade to Hold on Big F3Q16 Miss; Meaningful Forward Est. Reduction - Now Consider a 2HC2016 Story

Conclusion – Profitability Off-the-Table; Increased Competitive Pressures = Downgrade to Hold: Nimble’s F3Q16 results and F4Q16 outlook materially missed expectations; more importantly we / investors will likely now question the company’s ability to invest its way out of what appears to be an increasingly competitive landscape (albeit emphasizing consistent commercial win rates). Nimble reported revenue of $80.7M, well below the company’s $86-$88M guidance. Nimble attributed the F3Q16 shortfall and a material change in forward expectations (no longer guiding for profitability) to: 1.) a slower-than-expected ramp in enterprise sales productivity (now taking up to 8 quarters to full productivity ramp vs. our prior 5-6 quarter estimate), driven by competitive intensity – most notably driven by incumbents, but also reflective of increased competitive meet rates versus all-Flash solutions. 2.) A shift in investments from commercial to enterprise; commercial investments not enough to offset increased competitive pressures (i.e., too little channel incentive programs; note implied iSCSI growth deceleration).

With this, and now with a focus on reaccelerating growth over the next few quarters, Nimble is increasing sales and marketing expenses in its core commercial segment (increased incentives = product GM% concern; F3Q16 product GM% at 67.8% vs. LT mid-60% target), while sustaining its focus on enterprise investments. We believe this will leave investors to question whether the answer to the company’s abrupt change is as easy as increased investments vs. what will likely become a more, rather than less, competitive landscape, pushing out profitability expectations. Nimble is now


13

guiding F4Q16 at $87-$90M / ($0.11)-($0.13) vs. our prior $100.8M / $0.00 estimate (street: $101M / $0.00). We are reducing our F2017 and F2018 revenue estimates by 20% and 24% to $369.2M (+15.4% y/y) and $433.8M (+17.5% y/y) as we believe we need to take a prove-it view on Nimble’s ability to reaccelerate growth. While we would note that a 3x EV/sales multiple on our F2017 (C2016) revenue est. would imply ~$15/sh. (using ~$150M in net cash vs. current $210M), and we do expect Nimble to expand its TAM with an all-Flash offering in 1HC2016, we downgrade to Hold as we believe fundamental improvement looks to be a 2HC2016 story.

FC vs. iSCSI Growth: FC totaled 23.5% of bookings vs. 17% in F2Q16; 117 total FC customers in F3Q16 of which 95 were new (vs. 110 / 74 new in F2Q16). We would arrive at an implied iSCSI revenue of ~$62M vs. ~$66.5M est. in F2Q16, or implying a significant y/y growth deceleration at +5% (see charts below).


Nimble Storage Job Listings Update Nimble Storage’s job listings totaled 67, down 4 from the prior week and vs. 48 entering 2015. The company currently has 34 sales openings, up 1 from the prior week and compared to 20 entering 2015. Nimble has 13 openings for engineers, down 2 from the prior week and compared to 15 listings exiting 2014. Support listings were unchanged at 7 openings. As a reminder, Nimble added 71 employees during the July quarter (vs. +66 in the prior quarter), bringing its total headcount to 975 employees. We/investors will continue to focus on Nimble’s ability to expand its sales capacity. The figure below highlights Nimble’s employee job listing trends over the past several months.


14

Pure Storage (PSTG – NYSE- $14.74 – Buy)

Pure Storage Announces Support for 3D NAND, Pure1 Global Insight, Pure1 Capacity Support Offerings Pure Storage made several product announcements, including Pure1 Global Insight, Pure1 Capacity Planner, support for 3D TLC NAND flash, as well as FlashStack Converged Infrastructure support for Oracle and SAP. Highlights of these announcements include:

3D TLC NAND Support; $1.50/GB w/ Software: Pure Storage announced support for 3D TLC NAND in its FlashArray//m product line, supporting 44TB of raw capacity and allowing the ability to mix-and-match TLC and MLC flash in the same system to improve capacity of an existing array to 60TB per rack. Pure notes that Purity FlashCare will ensure TLC NAND has the same lifespan support and performance as MLC. The company notes that the inclusion of 3D TLC NAND will allow a fully scaled FlashArray//m system cost to reach as low as $1.50/GB with all software included. Pure will begin shipping the 44TB 3D TLC NAND Expansion Shelf in 1Q16.

We view this announcement as interesting, though we believe other flash storage vendors are also are beginning to utilize 3D NAND (albeit not announced), including Kaminario in its new K2 v5.5 (fifth generation) all-Flash primary storage arrays (cited at sub-$1/GB price points). We will be interested in the economical evolution of all-Flash storage as the industry looks to more broadly utilize 3D NAND Flash to drive down $/GB cost of flash relative to HDDs and gain broader adoption in secondary storage markets outside of mission critical workloads.

Pure1 Global: The company’s new Pure1 Global was noted as providing predictive cloud-based analytics that allows support personnel to foresee vulnerabilities and pre-emptively protect customers from issues and risks. Pure1 Global takes RT diagnostic data across 1000s of arrays to preform automated scans against a library of fingerprints to detect vulnerabilities to known issues, configurations, and environmental changes – prompting support to reach out to address the issue. Pure notes that it has already found / prevented issues ranging from a host and switch misconfiguration to cabling errors and other potential issues. As a reminder, Pure Storage has noted that 83% of cases are opened proactively by Pure1, while 58% of the first calls to customers related to an issue are initiated by Pure Storage. We believe this service is very similar to Nimble’s InfoSight cloud-based pro-active monitoring system.

Pure1 Capacity Planner: This new feature allows for customers to forecast capacity consumption over time and estimate when they may need to add capacity to their environment. The company notes that since the launch of Pure1 in June 2015, there are more than 2,000 users on the platform across the company’s 1,100 customer base.

FlashStack Oracle & SAP Support: Pure also announced Oracle and SAP support for its FlashStack Converged Infrastructure offering. The company notes that FlashStack for SAP provides a 4:1 average data reduction to traditional SAP ERP landscapes and reduced time spent on common admin tasks by up to 50%.

Hewlett Packard Enterprise (HPE – NYSE - $14.21 – Hold)

HPE Announces Updated All-Flash 3PAR with Support for 3D NAND SSDs at $5/GB Raw HP Enterprise announced last week that it launched updated 3PAR all-flash StoreServ storage solutions. In addition, the new all-flash 3PAR solutions are integrated with HPE’s StoreOnce and StoreEver platforms. HPE highlighted that the 3PAR StoreServ 20850 all-flash array demonstrates up to 75% better performance at half the price of the EMC VMAX 400K. HPE is also now offering a free 12-month license for customers that are migrating off of HDD-based legacy storage systems (including EMC, HDS and IBM XIV systems only). HPE’s new 3PAR systems also now support 3D NAND-based SSDs, similar to Pure’s announcement last week. HPE notes that 3D NAND SSDs will be GA on 12/17/15 with prices starting at $2,075 for a 400GB SSD, which would imply just over $5/GB raw. While Pure notes that it is able to achieve a ~5.5:1 data reduction on average, the company also highlights that HPE’s 3PAR system is only able to achieve less than 2:1 data reduction, which would imply an effective price at $2.50/GB. On the other hand, in contrast to other all-flash arrays, HPE highlights that its 3PAR system has a flash-optimized design (with both file and block protocol support) but that the system also support a secondary tier of HDDs for superior cost effectiveness. Leveraging this architecture, HPE also introduced the 3PAR StoreServ 8200 Converged File and Block Starter Kit, which enables consolidation onto a single unified system that costs 26% less than competitors. We would note that the argument HPE is making with its hybrid arrays is similar to the cost effectiveness of Nimble’s hybrid architecture in which Nimble notes that it rarely competes on price.

Cray (CRAY – NASDAQ- $34.24 – Buy)

Positive Meetings w/ Management; 4x Y/Y Increase in Commercial Interest

In conjunction with the SC’15 conference this week in Austin, TX, Cray hosted its annual investor meeting. Before we discuss some of the key takeaways from what we would view as a positive meeting with Cray management, we believe one of the most interesting comments by Cray’s President and CEO, Peter Ungaro, was that the company saw a quadrupling of its number of commercial customer briefings at this year’s conference versus last year. As a reminder,


15

Cray expects to see commercial accounting for more than 15% of its 2015 revenue; continue to work toward a +30% contribution going forward.

Supercomputing vs. Cloud: A common concern among investors is whether cloud increasingly competes against the supercomputing market. While cloud could compete against loosely coupled / clustered supercomputer architectures, Cray does not view this as a competitive threat to its high-end positioning in the supercomputer market. Cray highlights that it typically sees 90%+ utilization rates in its system deployments; workloads that would find cloud as too expensive.

2016 Outlook Confidence; Growing 2x Addressable Market: Management comments remain very positive on the company’s $825M guide for 2016 – driven by key product cycles: Intel’s Broadwell (1H2016), Intel’s Knights Landing (next-gen Xeon Phi; mid/2H2016), and NVIDIA’s Pascal (next-gen GPU) using HBM2 (16GB of on-silicon memory). With an expectation of continued growth in commercial verticals (energy, manufacturing, fin services), Cray is confident in its ability to continue to grow 2x its addressable market – i.e., 15%-20% per annum range, while also focusing on driving operating leverage (+10%-15% EBIT% remains a key focus). We continue to believe Cray is competitively well positioned to benefit from the aging IBM Blue Gene/Q installed base – this week’s win at the University of Warsaw represents Cray’s first Blue Gene/Q replacement.

Other: 1.) Intel OmniPath Usage: This week Cray announced its first clustered (CS-series) win using Intel’s OmniPath interconnect as a replacement to Mellanox’s InfiniBand. 2.) Storage Hierarchy Changes = Opportunity for Differentiation: Cray’s CTO, Steve Scott, outlined Cray’s focus on differentiating for the rapid architectural changes in the storage hierarchy – data movement across primary storage moving to Flash, next-gen memory (3D X-Point, PCM, ReRAM, STT-MRAM), and on-silicon memory. 3.) Beyond Aires: Cray has been working tightly with Intel on its integration of Aires capabilities into its next-gen. interface for Shasta systems. While Intel has ability to sell this to Cray competitors, there will be hardware-integrated / software differentiation specific for Cray systems (a 2018-2020 discussion).

Please see our detailed report published on 11/17/15 for additional information. Top500 Supercomputer List: Cray Has Leading 25% Share of Performance with Cont’d Top 100 Gains; IBM Adds No New Systems In conjunction with the Supercomputing Conference (SC ’15) in Austin, Texas, the Top500.org released its updated Top 500 global list of supercomputers; we would highlight the following takeaways (positive for Cray): Of the 135 new supercomputers listed for the first time on the November 2015 Top500 list, Cray accounted for 9 or ~7% of them (vs. 12 or 14% of the 87 total new systems on the June 2015 list; 13 or ~19% of the 68 new total systems on the November 2014 list and ~7% or 8 of the 114 new total systems on the June 2014 list). While Cray’s share of new systems declined, this was driven by the 87 new systems installed in China (from Sugon, Lenovo, and Inspur, respectively), only 6 of which were in the top 100. We would highlight that of the 16 new supercomputers in the top 100 that were listed for the first time, Cray accounted for 6 or ~38% of them; this compares to Cray accounting for 7 or 27% of the 26 new top 100 supercomputers on the June 2015 list. Cray increased its share of the top 100 to 33% from 31% on the June 2015 list and 28% on the November 2014 list. Cray increased its share of total performance (as measured by Rmax) of Top 500 supercomputers to ~25%, up from ~22% one year ago and 17% on the November 2013 list. While Cray took the lead from IBM (IBM: $138.50, Buy, covered by our colleague David Grossman) on the June 2015 list, Cray’s lead over IBM is now ~750bps. We would also highlight the sudden rise of not only Lenovo (at 4% performance share) but also Sugon and Inspur (at ~5% and ~2% performance share, respectively). IBM again did not add any new supercomputers to the top 100 and its share declined to 16% (including legacy IBM systems that are now Lenovo), vs. 21% on the June 2015 list and 28% on the November 2014 list. In fact, IBM did not add any new systems at all to the November 2015 list. Lenovo added 24 new systems (all in China: 5 to Alibaba, 3 to China Mobile, 1 each to China Telecom and China Unicom, respectively, 2 to China National Petroleum Corporation, and the remaining to undisclosed service providers); this compares to IBM + Lenovo adding 9 new systems on the June 2015 list. While Lenovo is clearly gaining traction in China (although Sugon had more success with its 45 new systems to the November 2015 list), we believe IBM’s lack of new Top 500 supercomputers continues to highlight the opportunity that Cray currently has in the high end of the supercomputing market. Cray has 69 total systems listed on the updated November 2015 list vs. 71 total systems on the June 2015 list, but the June 2015 list included two systems deployed at Cray itself (i.e., Crystal and Storm1) that are no longer on the November 2015 list. This also compares to 62 total systems on the November 2014 list and 51 total systems on the June 2014 list. Essentially, Cray added 9 new systems to the November 2015 list but 11 systems fell off (including the two systems installed at Cray), which left Cray unchanged overall from the prior list, but with greater share of overall performance and greater share of the more important top 100 systems.


16

Cray had 47 total high-end supercomputer deployments on the updated list (i.e., XC40, XC30, Xe6, XK7, and XT5) with the remaining 22 systems representing the company’s CS400 clustered systems, CS-Storm systems, and legacy Appro Xtreme-X systems. Cray has a total of ~4.46 million total cores deployed across its total systems on the November 2015 list, an increase of 25% y/y, although representing a deceleration from a 42% y/y increase on the prior June 2015 list and vs. +47% y/y on the November 2014 list. Cray’s ~3.90 million total high-end system cores were up 23% y/y vs. +41% y/y on the prior June 2014 list. Cray’s share of the top 100 is 33%, vs. 31% on the prior June 2015 list and 28% on the November 2014 list. Cray had 6 new top 100 systems (vs. 7 new top 100 systems on the prior June 2015 list), all XC40 systems utilizing Intel’s Xeon E5-2690v3 Grantley processors. The six systems (none of which used accelerators / co-processors) include Trinity, ranked #6 with 8.1 petaflops, at the Department of Energy’s Los Alamos and Sandia national laboratories, two systems at the Korea Meteorological Administration (Nuri and Miri at #29 and #30, respectively, with ~2.4 petaflops each), two at the NOAA (Luna and Surge at #41 and #42, respectively, with ~1.6 petaflops each), and one at the Texas Advanced Computing Center/University of Texas (named Lonestar 5 at #78 with ~1.0 petaflops). Cray’s 6 new top 100 systems compares to Lenovo adding 6 new top 100 systems (2 to Alibaba and 4 to undisclosed service providers in China), HP adding 2 new top 100 systems (an Apollo 6000 system to an undisclosed IT provider in the U.S. and a DL360 cluster to an undisclosed trading company in China), Fujitsu adding one new system to Nagoya University in Japan, and T-Platforms adding one system in Germany. IBM’s (including legacy IBM systems now with Lenovo) share of the top 100 now stands at 16%, SGI’s share remains 9%, Fujitsu’s share is at 8%, Bull’s share is at 7%, HP’s share is at 7%, Lenovo’s share is at 6% (including Lenovo-only ThinkServer systems), Dell’s share is 3%, T-Platforms’ share is at 3%, and Sugon’s share is 1%.

Top500.org Summary (November 2015 edition)

Cray Cray Difference from June 2015 edition

Systems # Deployed

Total # of

Cores

Avg. Cores /

System

Sum of

Rmax

Avg. Installed

Life # Deployed

Total # of

Cores

Avg. Cores /

System Sum of Rmax

Avg. Installed

Life

Cray XC 36 2,627,868 72,996 66,085,389 1.07 3 633,504 12,561 19,497,434 0.24

Cray XK 3 600,088 200,029 18,461,100 3.17 0 0 0 0 0.50

Cray XE 8 672,704 84,088 4,954,100 4.38 -4 -293,824 3,544 -2,124,443 0.42

Cray XT 0 0 N/A 0 N/A -1 -20,960 N/A -165,600 N/A

Cray Cluster (CS400) 6 141,300 23,550 2,623,465 1.00 -1 -5,914 2,519 -5,667 -0.07

Appro Xtreme-X 11 244,224 22,202 3,849,700 3.64 0 0 0 0 0.50

Cray CS-Storm 5 175,036 35,007 8,621,200 0.40 1 18,296 -4,178 505,700 0.15

Total 69 4,461,220 64,655 104,594,954 1.90 -2 331,102 6,485 17,707,424 0.05

IBM + Lenovo IBM and Lenovo Difference from June 2015 edition

Systems # Deployed

Total # of

Cores

Avg. Cores /

System

Sum of

Rmax

Avg. Installed

Life # Deployed

Total # of

Cores

Avg. Cores /

System Sum of Rmax

Avg. Installed

Life

IBM BlueGene 19 4,308,992 226,789 47,047,481 3.45 -7 -163,840 54,757 -1,324,702 0.52

IBM iDataPlex 23 807,976 35,129 16,829,167 2.76 -11 -102,382 8,354 -1,916,863 0.66

IBM pSeries (Unix) 3 79,328 26,443 2,010,407 3.33 -5 -89,856 5,295 -2,322,124 0.02

IBM BladeCenter (x86) 0 0 N/A 0 N/A -3 -80,160 N/A -513,547 N/A

IBM xSeries (x86) 3 36,352 12,117 697,893 1.50 -19 -506,288 -12,548 -3,326,445 0.18

IBM Flex Systems 3 49,504 16,501 1,201,135 1.50 -2 -20,512 2,498 -388,064 0.50

IBM NeXtScale 7 201,200 28,743 5,179,055 0.93 -5 -44,376 8,278 -807,575 0.35

Lenovo NeXtScale 1 5,040 5,040 212,841 0.50 -3 -20,448 -1,332 -649,550 0.50

Total 59 5,488,392 93,024 73,177,979 2.63 -55 -1,027,862 35,863 -11,248,871 0.70

Hewlett-Packard HP Difference from June 2015 edition

Systems # Deployed

Total # of

Cores

Avg. Cores /

System

Sum of

Rmax

Avg. Installed

Life # Deployed

Total # of

Cores

Avg. Cores /

System Sum of Rmax

Avg. Installed

Life

HP Cluster Platform 3000 46 1,415,356 30,769 14,962,729 1.28 -4 -101,756 426 1,678,548 0.01

HP Cluster Platform 3000BL 36 742,696 20,630 12,231,652 1.58 -28 -381,212 3,069 -5,011,886 0.22

HP Cluster Platform 3000SL 48 1,013,764 21,120 15,786,982 1.88 -6 -74,376 969 -540,525 0.35

HP Cluster Platform 4000BL 1 64,128 64,128 337,175 1.00 0 0 0 0 0.50

HP Pod 2 55,680 27,840 917,310 1.50 0 0 0 0 0.50

HP Apollo 23 353,452 15,367 10,001,958 0.22 16 236,444 -1,348 6,550,629 0.07

Total 156 3,645,076 23,366 54,237,806 1.38 -22 -320,900 1,085 2,676,766 0.05

SGI SGI Difference from June 2015 edition

Systems # Deployed

Total # of

Cores

Avg. Cores /

System

Sum of

Rmax

Avg. Installed

Life # Deployed

Total # of

Cores

Avg. Cores /

System Sum of Rmax

Avg. Installed

Life

Rackable Cluster 6 144,848 24,141 2,964,576 2.00 0 0 0 0 1.00

SGI ICE 24 1,119,520 46,647 24,735,530 1.81 1 58,144 500 546,962 0.92

Total 30 1,264,368 42,146 27,700,106 1.85 1 58,144 552 546,962 0.94

Source: top500.org; Stifel estimates

Note: IBM + Lenovo category encompasses legacy IBM, or rather excludes 25 Lenovo-only systems (i.e. 2 Lenovo Clusters, 7 NeXtScales, and 16 ThinkServers)


17



18

Biggest Development at SC15 Likely to be OpenHPC; Focus on Potential Competition from ARM over Long-Term The Next Platform published an article last week by Addison Snell, CEO of Intersect360, that the biggest story coming out of the SC15 supercomputing conference was the announcement of Intel’s OpenHPC foundation. Mr. Snell notes that while product announcements are important, the longer-term impacts usually come from standards, mergers, and new groups. Mr. Snell believes that Intel needs its partners in OpenHPC as much as its partners need Intel. The article quotes Richard Stallman in noting that OpenHPC is free in the sense of free speech rather than free beer or rather open competition as opposed to open access. While Intel currently has a dominant lead among supercomputers (i.e., ~90% of supercomputers on November 2015 Top500 list), Mr. Snell notes that a potential vulnerability to Intel is the move to accelerators, and that Intel’s presence in the accelerator market with Xeon Phi has much catching up to do regarding Xeon Phi’s software ecosystem around it vs. NVIDIA’s early-mover lead with CUDA. Mr. Snell believes that Intel is coming from behind in the interconnect market despite the recent launch of OmniPath. While OpenHPC seems to be up against OpenPOWER, the other competitive force is ARM. Mr. Snell notes that while work on ARM is still nascent (i.e., Barcelona Supercomputing Center is doing some work testing ARM architecture), ARM represents a paradigmatic shift in the next few years. And despite both OpenHPC and OpenPOWER being open to any members, Mr. Snell believes it’s unlikely that ARM technologies will join either organization. In sum, Mr. Snell believes that the HPC industry is continuing its migration away from standardization and back toward specialization as end users will be forced to choose a platform (i.e., Intel, IBM, NVIDIA, or ARM) and once down a certain path, lock-in will be significant / it will be very difficult to change tracks later.

Recall that Cray noted at its Investor Day last week that it has been working with Intel for a long time on OpenHPC, which will be targeted at relatively lower-end loosely coupled machines rather than high-end tightly coupled systems. Participants in OpenHPC currently include Allinea Software, Altair, ANSYS, Argonne National Laboratory, Atos (i.e., Bull), Barcelona Supercomputing Center, The Center for Research in Extreme Scale Technologies at Indiana University, Cray, Dassault Systemes SIMULIA, Dell, Fujitsu Systems Europe, Hewlett Packard Enterprise, Intel, Jülich Supercomputing Centre, Lawrence Berkeley National Laboratory, Lawrence Livermore National Laboratory, Leibniz Supercomputing Centre, Lenovo, Los Alamos National Laboratory, MSC Software, NEC, Oak Ridge National Laboratory, Pacific Northwest National Laboratory, ParTec, Penguin Computing, Pittsburgh Supercomputing Center, Sandia National Laboratories, SENAI CIMATEC, SUSE, and Texas Advanced Computing Center. One major vendor noticeably absent from the list is SGI. See our Weekly report two weeks ago for more details. Cray Wins Awards for XC40 Platform; Indicative of Cray’s Leadership Position at the High End of the Market Cray announced last week that it won six awards from the readers and editors of HPCwire, as part of the publication's 2015 Readers' and Editors' Choice Awards, which is the 12

th consecutive year Cray has been selected for multiple

HPCwire awards. Most notably, Cray’s XC high-end supercomputing platform received both Readers' and Editors' Choice awards for Best HPC Server Product or Technology. Receiving both Readers' and Editors' Choice awards is rare. We believe the awards demonstrate Cray’s clear lead at the high end of the supercomputing market. Recall that Cray achieved 25% of total performance across all systems on the November 2015 Top500 list and that IBM did not add a single new supercomputer to the latest list.

Support for Docker Containers on Cray’s Supercomputers Will Improve Ease-of-Use and Will Continue to Drive Increased Commercial Traction Cray announced last week that it will support HPC-optimized Docker containers on its XC line of supercomputers (Docker will be available on Cray’s CS400 as well as XE and XK platforms in 2016). Docker support will enable customers to package an entire software stack (i.e., application code, shared libraries, base OS files, user environment variables and dependencies, etc.) into one container that can be more easily deployed across multiple platforms. The main advantage is that Docker support will increase ease-of-use for commercial / enterprise users of Cray’s supercomputing systems. Recall that in our meetings with Cray management last week at the SC15 supercomputing conference in Austin, TX, ease-of-use is currently a major focal point for Cray’s salesforce and Cray will continue to change the perception in the market so that prospective commercial customers realize that Cray’s systems are no more difficult to use than competing systems. Recall that Cray expects to see commercial accounting for more than 15% of its 2015 revenue; continue to work toward a +30% contribution going forward.

Super Micro Computers (SMCI – NASDAQ- $23.40 – Buy)

SuperMicro 10-Q Review: Revenue Restatement of $10.6M in F1Q16; Reiterate Forward Guide, Purchase Commitments, Ablecom, & More SuperMicro filed its F1Q16 10-Q after announcing last week that it would be delayed by 5 days following internally determining that certain contracts for extended product warranties were recorded as revenue in prior periods instead of being deferred and amortized over the contract of the period. The company provided an update of the impact of this accounting issue, reporting that it will reduce its F1Q16 revenue by $10.6M, related to early recognition of revenue over the past three fiscal years. Additional details / highlights from the filing include:


19

Revenue Restatement; $10.6M Cumulative Adjustment w/ $1.3M Related to F1Q16: The company reported that the cumulative reduction of F1Q16 revenue is $10.6 million, including $9.3 million in the prior three fiscal years ending June 2015 (F2015) and $1.3 million in the September 2015 quarter (F1Q16). Consistent with the company’s previously announced treatment, SuperMicro’s F1Q16 10-Q filing reflects the above cumulative revenue adjustment, restating September quarter revenue at $519.635 million vs. the previously reported $530.235 million. The deferred revenue related to this adjustment will be recognized through January 2019. When adjusting for the September quarter deferred revenue recognition only, we would estimate the company’s non-GAAP gross margin to be ~15.5% vs. the previously reported 15.7%. However, the cumulative $10.6M adjustment would leave adjusted non-GAAP GM% at approximately 13.9% and account for approximately a $0.13/sh. reduction in F1Q16 non-GAAP EPS; adjusted F1Q16 EPS at $0.32/sh. vs, the reported $0.45/sh.

SuperMicro reiterated its previously stated F2Q16 (December 2015) outlook at $580-$630M / $0.54-$0.64. Our estimates remain unchanged at $608.54M/$0.62 (street: $600/$0.59). We would again highlight our conversation with the company last week emphasizing SuperMicro remains confident in its underlying growth drivers going forward – reiterating its expectation of posting more than 20% y/y growth in F2016.

One +10% Customer (IBM/SoftLayer) at 12.2% of F1Q16 Revenue: SuperMicro had previously disclosed that it had one plus-10% customer in F1Q16. The company’s 10-Q discloses that IBM/SoftLayer accounted for 12.2% of total revenue, up from the disclosed 4.0% revenue contribution in the prior quarter; also compares to an ~8.3% contribution in F1Q15. This implies ~$63.4M in revenue during F1Q16, which compares to an estimated $22.99M and ~$36.65M contribution in F4Q15 and F1Q15, respectively. Given that we believe this is IBM/SoftLayer, which would fall in the company’s Internet Data Center/cloud (iDC) vertical, we would note that this would equate to ~49% of total IDC vertical revenue. This also implies non-SoftLayer IDC vertical revenue grew ~172% y/y in F1Q16 (roughly flat sequentially). The company reported that one customer accounted for 10.3% of accounts receivable exiting the September quarter, vs. none accounting for 10% in the year-ago period.

Purchase Obligations – Total ex-HDDs at -3.5% Y/Y. SuperMicro’s purchase commitment balance totaled $391.96 million exiting F1Q16, up from $378.3 million exiting the prior quarter and from $226.53 million exiting the year-ago September quarter (+73% y/y). HDD purchase obligations (paid through December 2016), the company's only long-term purchase commitments, totaled $201.58 million exiting F1Q16, up from the $185.7 million balance exiting the prior quarter. This leaves total purchase obligations, excluding HDDs, at $190.4 million, down from $192.6 million exiting the prior quarter and down 3.5% y/y (see chart below). We would note that this compares to +49%, +61%, +28%, and +16% y/y exiting the September, December, March, and June quarters, respectively. Below we highlight the relatively high historical correlation (R2 = 0.90) between SuperMicro’s purchase obligations, excluding HDDs, versus two quarter forward total revenue.

Inventory Breakdown – Finished Goods Expansion = Forward Indicator? SuperMicro exited F1Q16 with $461.4 million in total inventory. As shown in the chart below, SuperMicro’s 10-Q discloses SuperMicro’s inventory breakdown with finished goods inventory totaling $374.045 million exiting F1Q16, a decrease from $304.2 million exiting the prior quarter and versus $252.36 million exiting the year-ago quarter (+48% yr/yr). Work-in-Process inventory totaled $42.29 million, up from $23.2 million exiting F4Q15 and up from $33.1 million exiting the year-ago quarter. Raw materials


20

inventory totaled $45.04 million exiting F1Q16, down from $55.63 million and $56.1 million exiting the prior and year-ago quarters, respectively. Below we also provide a chart highlighting the historical correlation between SuperMicro’s quarter ending finished goods inventory balance vs. forward two quarter total revenue.

Machinery & Equipment Balance (Gross); Historical Correlation of Machinery & Equipment + Purchase Obligations (ex-HDDs) + Finished Goods. Given that SuperMicro manufactures its servers, we also track the company’s 10-Q disclosure of gross Machinery & Equipment balance as an indicator of the company’s investors for forward revenue. The company exited F1Q16 with a total gross (ex-accumulated depreciation) Machinery & Equipment balance totaling $43.13 million, an increase from $40.689 million exiting the prior quarter and up from $36.34 million exiting the year-ago quarter (+19% y/y). While SuperMicro’s forecasting visibility must always be considered, we continue to find it interesting to highlight the historically high correlation between the forward 2-quarter revenue and the sum of the company’s ending Machinery & Equipment, Purchase Obligations (ex-HDDs), and Finished Goods inventory – historical R

2 = 0.95 (see chart below).

Ablecom Disclosures. SuperMicro’s 10-Q provides several disclosures as it relates to the relationship with Ablecom – the company’s contract manufacturing and distribution partner; Ablecom’s CEO, Steve Liang, is the brother of SuperMicro’s CEO, Charles Liang. Charles Liang & Sue Liu’s ownership of Ablecom continues to stand at 10.5%; Steve Liang’s ownership of SuperMicro shares has remained unchanged sequentially at 0.6%. Disclosures include:

SuperMicro Sales to Ablecom: SuperMicro’s sales of products to Ablecom totaled $5.22 million in F1Q16, which represents a decline from $14.5 million and a record $26.6 million in the prior and year-ago quarters.


21

Ablecom Revenue from SuperMicro: Ablecom’s revenue received from services to SuperMicro totaled $59.26 million in F1Q16, equating to 13.3% of total COGS vs. 11.9% in the prior quarter; averaging 13.6% over the prior five quarters and 15.8% of quarterly COGS over the past three years.

Amounts Owed to Ablecom: SuperMicro’s balance of amounts owed to Ablecom exiting F1Q16 totaled $55.9 million, down from $59.0 million exiting the prior quarter, but up from $45.1 million exiting the year-ago quarter.

Outstanding Purchase Orders to Ablecom: SuperMicro’s outstanding purchase orders to Ablecom totaled $60.95 million exiting F1Q16, a decrease from $67.3 million exiting the prior quarter and $70.455 million exiting the year-ago quarter.

Other Highlights: (1) SuperMicro reported that it saw a $1.9 million FX-related gain during the September 2015 quarter, compared to a $0.5M gain in the year-ago period. (2) SuperMicro’s warranty accrual in F1Q16 remained relatively unchanged at $3.9 million. SuperMicro’s product warranties range from 15-39 months against any defective product. (3) SuperMicro’s 10-Q again discloses that the company is currently in the process of implementing a new ERP (enterprise resource planning) system that is expected to be commencing usage in calendar 2015. The process of implementing this ERP system has been ongoing over the past few quarters and during the September quarter the company began implementation for its Taiwan and Netherlands subsidiaries. (4) The filing notes that in November 2015, SuperMicro extended its credit line with Bank of America to expire on 2/15/16 and is currently negotiating to renew the revolving credit line. SuperMicro had $2.8M remaining of the $65M credit line exiting the September quarter. (5) Other Adjustments. With the above revenue reduction, additional adjustments to the company’s balance sheet included: Accrued Liabilities: Increase from the previously reported $46.967 million to $50.169 million. Income Taxes Payable: Adjusted from $14.866 million to $11.045 million. Other Liabilities: Increase from $17.215 million to $24.629 million.

11/18/15: SuperMicro Participates in Investor Conference Last week SuperMicro’s CFO, Howard Hideshima, participated in an investor conference. The vast majority of the presentation and Q&A centered on an overview of SuperMicro’s business, but we would highlight one comment regarding SuperMicro’s need to continue to invest in working capital to support revenue growth going forward. This is not new, but it does highlight the relatively weak free cash flow metrics that are driven by SuperMicro’s turns business. Specifically, Mr. Hideshima noted that SuperMicro will continue to use working capital because it has to build inventory to support revenue growth. Mr. Hideshima highlighted that the company's goal is to keep inventory days in the mid-80s (which is the company's average over the last several quarters) and SuperMicro will continue to manage its working capital efficiently, but the company is in a high growth phase and will be using cash for working capital. Recall that SuperMicro's FCF has averaged -1% of revenue over the last few years.

CommVault Systems (CVLT – NASDAQ - $40.16 – Buy)

CommVault Job Listings Update CommVault’s job listings totaled 74, down 5 from the prior week and vs. 132 openings exiting 2014. The company had 32 sales openings, 4 openings in professional services, and 16 openings in system engineer. This compares to 33, 4, and 17 openings last week, respectively. CommVault exited the September quarter with total headcount at 2,379, up from 2,305 in the prior quarter. CommVault noted that its Americas sales capacity was up substantially from where it was exiting F2015 (March 2015), while attrition challenges have appeared to stabilize. As such, our tracking of U.S. sales


22

listings would find listings now at 9, down 1 from the prior period and down from a peak of 70 in late May. The figures below highlight CommVault’s employee job listing trends over the past several months.

Western Digital (WDC – NASDAQ - $61.71 – Buy)

W. Digital + SanDisk Deep Dive: Vertical Innovation + Vertical Integration = Strong Long-Term Strategy Stifel Conclusion: As we evaluate the various considerations related to the WD + SanDisk (SNDK: $74.93, Buy, covered by our colleague Kevin Cassidy) combo, we believe it is important to consider what we believe are the two key tenants of Western Digital’s strategic direction: 1.) Vertical Innovation – a media agnostic portfolio positioning in which differentiation can be attained through tight innovation across key layers in the rapidly evolving storage landscape. This entails significant patent and internal portfolio capabilities across controllers and firmware features / functionalities; we believe allowing WD to attain a broader reach in all aspects of storage – including the companies’ continued move up the value stack. 2.) Vertical Integration – while understood, we believe it is important for investors to consider WD’s ability to leverage SanDisk’s industry-leading NAND cost curve capabilities (e.g., SanDisk’s 15nm X3 capabilities). We believe this can result in an enterprise SSD gross margin into the 50%+ range (vs. mid-20% range today); we believe this represents a majority of the guided $500M synergies. Investors could also consider the need for tight NAND Flash / non-volatile memory integration in next-gen high-cap Shingled Magnetic Recording (SMR) drives. We believe the combined company can leverage a balanced view on what appears to be accelerating shifts in the overall storage market, or in some ways the role of an evolution / innovation influencer.

Additional Topics / Considerations:

1. Cost Curve of NAND Flash: Although the HDD industry will continue down the cost curve driven by SMR and eventually HAMR, we continue to see a declining $/GB for NAND Flash – we estimate SanDisk’s COGS/GB has declined by a ~30% CAGR, or 2x the decline in WD’s COGS/GB decline from 2009-2015.

2. 3D NAND / SanDisk’s Positioning; 2D NAND Longevity: SanDisk/Toshiba have taken a different approach to transitioning to 3D NAND, waiting to volume ramp 3D NAND until it can obtain cost efficiency vs. planar – an approach widely considered, we believe wrongly so, as competitively behind; production yields and die size remain key considerations. While the progression toward 3D NAND will remain a key focus, we believe investors should also consider the potential long tail of cost effective 2D NAND (e.g., SanDisk’s 15nm and the possibility of further shrinks).

3.) The Intel Tax: Based on our math, we believe the vertical integration of NAND supply alone could equate to +$300M in savings for WD’s enterprise SSD business – estimating WD pays +$0.50/GB for Flash from Intel vs. SanDisk’s production at sub-$0.20/GB. It is also important to note that the HGST / Intel joint development partnership only pertains to enterprise SAS SSDs.


23


24


Seagate Technology (STX – NASDAQ - $34.41 – Hold)

Seagate Launches Three New Versions of ClusterStor in addition to New HPC HDDs with Superior Bandwidth and Latency vs. Competitive HDDs Seagate announced last week that it launched three new ClusterStor solutions (i.e., L200, G200, and A200) in addition to Seagate’s new ClusterStor HPC HDD. The new ClusterStor L300 is now the highest-performing version of Seagate’s ClusterStore Lustre storage architecture, with up to 112GBps bandwidth per rack when utilizing the company’s new ClusterStor HPC HDD, which supports up to 4TB capacity and 300MBps bandwidth, which Seagate highlights is the highest sequential bandwidth of any HDD on the market (as well as 35% faster random performance than any other 3.5” HDD on the market). We would note that this compares to HGST’s 8TB Helium Ultrastar HDD with 205MBps bandwidth. Seagate notes that the bandwidth of the new L300 is 77% higher than prior-gen ClusterStor solutions. In addition to the eL300, Seagate also launched its ClusterStor G200 Spectrum Scale (Spectrum Scale is the new name of IBM’s General Parallel File System), an HPC-class storage system targeted at the enterprise. The G200 also utilizes Seagate’s new ClusterStor HPC HDDs. Seagate highlights that HPC is no longer limited to just academia, defense, or intelligence applications, but rather a growing number of industries as they grapple with how to manage massive amounts of data. Seagate believes that its G200 provides a purpose-built solution, i.e., including everything from disk drives to the enclosures (84 slots for 3.5-inch HDD and SSD in a 5U enclosure with embedded storage controllers) and software tuned for HPC performance and scalability. The G200 system comes pre-configured, tested, and ready to deploy in a matter of hours. The new A200 is targeted at archive workloads and was designed to complement the rest of the ClusterStor family of scale-out storage solutions. The A200 enables customers to non-disruptively migrate designated data off of the performance-optimized, primary storage tiers while keeping it online for fast retrieval. The A200 consists of up to seven scalable storage units per rack with additional storage units and racks easily added to the storage pool. The ClusterStor L300, G200, and A200 will be GA in C1Q16. The ClusterStor family now includes L300, G200, and A200 in addition to previous versions such as the ClusterStor 9000, ClusterStor 1500, ClusterStor Secure Data Appliance, and Hadoop Workflow Accelerator for ClusterStor’s architecture.

Brocade Communications Systems (BRCD – NASDAQ - $9.85 – Hold)

Brocade Job Listings Update As of last week, Brocade’s overall job listings totaled 208, up 16 from the prior week and compared to 212 openings entering 2015. There were 31 total sales listings, up 2 from the prior week. Services listings totaled 14, up 1 from the prior week, while support listings totaled 12, up 1 from the prior week. We would highlight the company’s headcount stood at 4,626 exiting the July quarter (vs. 4,553 exiting the April quarter). The figure below highlights Brocade’s job listings over the past several months.


25

All prices are as of 11/20/15 market close. All statements in this report attributable to Gartner represent Stifel's interpretation of data, research opinion or viewpoints published as part of a syndicated subscription service by Gartner, Inc., and have not been reviewed by Gartner. Each Gartner publication speaks as of its original publication date (and not as of the date of this report). The opinions expressed in Gartner publications are not representations of fact, and are subject to change without notice.


26

Enterprise Storage Industry Indices

Source for All Charts: FactSet and Stifel Past performance should not and cannot be used as an indicator of future performance


27

Hard Disk Drive Industry Indices

Source for All Charts: FactSet and Stifel Past performance should not and cannot be used as an indicator of future performance


28

Important Disclosures and Certifications I, Aaron C. Rakers, CFA, certify that the views expressed in this research report accurately reflect my personal views about the subject securities or issuers; and I, Aaron C. Rakers, CFA, certify that no part of my compensation was, is, or will be directly or indirectly related to the specific recommendation or views contained in this research report. Our European Policy for Managing Research Conflicts of Interest is available at www.stifel.com. For applicable current disclosures for all covered companies please visit the Research Page at www.stifel.comor write to the Stifel Research Department at the following address. US Research Stifel Research Department Stifel, Nicolaus & Company, Inc. One South Street 16th Floor Baltimore, Md. 21202 The equity research analyst(s) responsible for the preparation of this report receive(s) compensation based on various factors, including Stifel’s overall revenue, which includes investment banking revenue. Our investment rating system is three tiered, defined as follows: BUY – We expect a total return of greater than 10% over the next 12 months with total return equal to the percentage price change plus dividend yield. HOLD – We expect a total return between -5% and 10% over the next 12 months with total return equal to the percentage price change plus dividend yield. SELL – We expect a total return below -5% over the next 12 months with total return equal to the percentage price change plus dividend yield.

Occasionally, we use the ancillary rating of SUSPENDED (SU) to indicate a long-term suspension in rating and/or target price, and/or coverage due to applicable regulations or Stifel policies. SUSPENDED indicates the analyst is unable to determine a “reasonable basis” for rating/target price or estimates due to lack of publicly available information or the inability to quantify the publicly available information provided by the company and it is unknown when the outlook will be clarified. SUSPENDED may also be used when an analyst has left the firm. Of the securities we rate, 54% are rated Buy, 40% are rated Hold, 1% are rated Sell and 5% are rated Suspended. Within the last 12 months, Stifel or an affiliate has provided investment banking services for 19%, 7%, 5% and 5% of the companies whose shares are rated Buy, Hold, Sell and Suspended, respectively.

Additional Disclosures

Please visit the Research Page at www.stifel.com for the current research disclosures and respective target price methodology applicable to the companies mentioned in this publication that are within Stifel’s coverage universe. For a discussion of risks to target price please see our stand-alone company reports and notes for all Buy-rated and Sell-rated stocks. The information contained herein has been prepared from sources believed to be reliable but is not guaranteed by us and is not a complete summary or statement of all available data, nor is it considered an offer to buy or sell any securities referred to herein. Opinions expressed are subject to change without notice and do not take into account the particular investment objectives, financial situation or needs of individual investors. Employees of Stifel, or its affiliates may, at times, release written or oral commentary, technical analysis or trading strategies that differ from the opinions expressed within. Past performance should not and cannot be viewed as an indicator of future performance. As a multi-disciplined financial services firm, Stifel regularly seeks investment banking assignments and compensation from issuers for services including, but not limited to, acting as an underwriter in an offering or financial advisor in a merger or acquisition, or serving as a placement agent in private transactions.

http://www.stifel.com/




29

Affiliate Disclosures “Stifel”, includes Stifel Nicolaus & Company (“SNC”), a US broker-dealer registered with the United States Securities and Exchange Commission and the Financial Industry National Regulatory Authority and Stifel Nicolaus Europe Limited (“SNEL”), which is authorized and regulated by the Financial Conduct Authority (“FCA”), (FRN 190412) and is a member of the London Stock Exchange. Registration of non-US Analysts: Any non-US research analyst employed by SNEL contributing to this report is not registered/qualified as a research analyst with FINRA and is not an associated person of the US broker-dealer and therefore many not be subject to NASD Rule 2711 or NYSE Rule 472 restrictions on communications with a subject company, public appearances, and trading securities held by a research analyst account. Country Specific and Jurisdictional Disclosures United States: Research produced and distributed by SNEL is distributed by SNEL to “Major US Institutional Investors” as defined in Rule 15a-6 under the US Securities Exchange Act of 1934, as amended. SNEL is a non-US broker-dealer and accordingly, any transaction by Major US Institutional Investors in the securities discussed in the document would need to be effected by SNC. SNC may also distribute research prepared by SNEL directly to US clients that are professional clients as defined by FCA rules. In these instances, SNC accepts responsibility for the content. Research produced by SNEL is not intended for use by and should not be made available to retail clients, as defined by the FCA rules. Canadian Distribution: Research produced by SNEL is distributed in Canada by SNC in reliance on the international dealer exemption. This material is intended for use only by professional or institutional investors. None of the investments or investment services mentioned or described herein is available to other persons or to anyone in Canada who is not a “permitted client” as defined under applicable Canadian securities law. UK and European Economic Area (EEA): This report is distributed in the EEA by SNEL, which is authorized and regulated in the United Kingdom by the FCA. In these instances, SNEL accepts responsibility for the content. Research produced by SNEL is not intended for use by and should not be made available to non-professional clients. In jurisdictions where Stifel is not already licensed or registered to trade securities, transactions will only be affected in accordance with local securities legislation which will vary from jurisdiction to jurisdiction and may require that a transaction carried out in accordance with applicable exemptions from registration and licensing requirements. Non-US customers wishing to effect transactions should contact a representative of the Stifel entity in their regional jurisdiction except where governing law permits otherwise. US customers wishing to effect transactions should contact their US salesperson.

Additional Information Is Available Upon Request © 2015 Stifel. This report is produced for the use of Stifel customers and may not be reproduced, re-distributed or passed to any other person or published in whole or in part for any purpose without the prior consent of Stifel. Stifel, Nicolaus & Company, Incorporated, One South Street, Baltimore, MD 21202.

Documents

November 22, 2015 IT Hardware, HDD/SSD, & Mobile Weekly