technology, media & telecom key points

© 2014 FBR CAPITAL MARKETS & CO. Institutional Brokerage, Research, and Investment Banking

David [email protected]

Christopher [email protected]

Scott [email protected]

Technology, Media & Telecom

March 4, 2014 Industry UpdateMWC 2014 Insights and Key TakeawaysSummary and RecommendationLast week, we attended Mobile World Congress (MWC) in Barcelona, the largest annual globalwireless conference. The conference was bigger than ever in FY14, with more than 80,000attendees. One of the key conference attractions may have been the desire to escape from poor,inconsistent voice and data network user experiences in countries such as Spain, France, the UnitedKingdom, and Ireland, which are plagued by under-investment due to low incremental returns oninvested capital associated with heavy-handed regulation and a four-player construct. Regulatorsappear to be taking note, and share prices, at least in Europe, are priced for consolidation.

Key Points■ Takeaways for the telecom services sector. Developments in focus included data usage,

incremental returns, M&A, regulation, devices, spectrum, networks, and interconnection.Key points: (1) The spectrum challenge is intensifying in the U.S. relative to the rest of theworld, positioning cable Wi-Fi-based networks well; (2) LTE Advanced is a differentiator, butincremental ROIC is low, which should drive industry consolidation, where we are bullish; and(3) the appeal of SDN and NFV is growing outside the datacenter, but service providers facestructural hurdles.

■ Takeaways for the equipment sector. As the worlds of networking and compute collide,carriers and Web scale cast a new vision for IT. MWC further opened our eyes to theirreversible trends in the networking industry. The networking and compute process, whencombined in ways that Web-scale operations have been perfecting for several years, is likelyto deliver significant savings over traditional methods and architectures. At the heart of thistransformation are software-defined networking and network function virtualization. WhileSDN and NFV solutions are well into their commercial phases and gaining momentum, webelieve it may take some time before these solutions become truly scalable. It is clear that theindustry is headed into a more virtualized world. That world is likely to become much morecompetitive as traditional communications equipment vendors begin to innovate at a pacethat more closely resembles the overall IT industry.

■ Takeaways for the semiconductor sector. The most evident theme for Mobile World Congress2014 was the ongoing commoditization of smartphone hardware, particularly those newmodels in the low and mid tiers. Overall, we believe this commoditization places evengreater emphasis on a competitive bill of materials for handset OEMs and increasingpressure on semiconductor ASPs. Additionally, as the pace of smartphone features andinnovation has slowed, it has served to lengthen the consumer handset refresh rate. Whilesmartphone penetration continues globally, it has been partially offset by this lengthening inthe consumer purchasing cycle, slowing overall volume growth. Also fitting within the theme ofcommoditization and component ASP pressure is the increased availability of LTE in 2014 froma variety of vendors (besides QUALCOMM), including Intel, Broadcom, Marvell, MediaTek,and NVIDIA. That said, innovation abounds in telecom infrastructure with opportunities forsemiconductor companies to disrupt traditional networking equipment incumbents.

Important disclosures can be found on pages 20 - 23 of this report.

MWC Takeaways: Telecom Services Sector David Dixon

703.469.1282

Summary Thoughts—Bullish on Industry Consolidation

Throughout the week, we focused on key developments across each area highlighted on page 1. The key inflection points within these areas that showed themselves at MWC 2014 included the following: (1) There is an intensifying spectrum challenge in the U.S. market relative to the rest of the world, which should drive industry consolidation; (2) LTE Advanced is here in response as a network differentiator, and applications such as LTE Broadcast have to work given the spectrum challenge, but with incremental returns on invested capital becoming more challenging, integrating Wi-Fi with LTE networks to better manage traffic is key, despite better QoS on LTE networks that are driving a short-term device trend away from Wi-Fi to LTE—this also bodes well for industry consolidation, in our view; and (3) the appeal of SDN outside the datacenter is growing, and NFV is about to happen—providing the potential for a standardized compute platform, but major structural challenges exist that may inhibit momentum over the next five years. As a means for the future, 5G divergence from 4G was interesting, with Kumu Networks and pCell suggesting significant disruption potential. The forking of Android was unexpected, with implications for Nokia, Google, and Samsung, but with Samsung and Huawei not known for software excellence, they may be exposed. For each focus area we provide the following specific insights for investors:

Big Switches, Little Servers—thesis developments: (1) There is a broader appeal of SDN and NFV beyond the datacenter. (2) The threat of a shift to commodity hardware platforms is driving steep discounts from legacy providers, and network operators cited two examples: (a) Cisco in the datacenter and (b) VMWare—95% discounts achieved on licenses, coupled with requests not to move forward with open-source standards based on Openstack. (3) Steep discounts are driving operators back to legacy providers for now. Savings generated are being reinvested into the architecture shift. This data appears to be consistent with EMC’s recent results whereby the company lowered license growth revenues and lowered expectations for NSX revenues from Nicira. Our checks with network operators confirm significant challenges with the Nicira platform, though it is early days. (4) Virtualized network functions are being bundled into network solutions offered by Ericsson for (close to) no cost, suggesting software revenue momentum may be lower than expected in the early transition phase. (5) Major changes to organizational structures are required to align network, compute, and storage functions. Current silo-based organizational structures are inhibiting momentum amid a clash of legacy culture and risk-averse mindsets from an aging work force. (6) Aggressive hiring of software engineering talent with different compensation structures is needed—AT&T is trying to build momentum in this area. (7) Leading indicators from Telefónica are not good. The CTO role (Caytano Martin) is taking greater significance as the company positions to sell cloud and network services to enterprises based on an internal smart network that is reapplied to external customers by undergoing a reorganization to align compute, network, and storage. Applications released to date have been diplomatically less than stellar. (8) Telecom service providers are progressively migrating internal UNIX-based applications to LINUX as a precursor to virtualization. Depending on which version of Unix, certain internal applications cannot run on virtual machines. This is ahead of SDN 2.0 controller deployment in the datacenter based on standardized OpenFlow technology to communicate between the software controller and all existing commoditized hardware and the migration of key internal network platforms. (9) There is an effort to build out more distributed cloud architectures to optimize mobile video and potentially differentiate from less distributed third-party cloud and CDN providers—the challenge is the limited datacenter space available to create further distribution in key areas. However, service providers may be able to leverage existing CO locations. AT&T plans to deploy 4,600 cloud nodes across its network but we see established datacenter providers such as Equinix as attractive assets due to scarcity in key locations. (10) Core Network Virtualization is occurring first. (11) The NSA effect is

Institutional Brokerage, Research, and Investment Banking© 2014 FBR CAPITAL MARKETS & CO.

driving requests to build localized data centers as DT pushes for a European data network, which could slow the pace of SDN adoption in some areas.

Data usage, network performance, capex spending, and economic returns. (1) The insatiable appetite for 4G data is causing a greater-than-expected capex squeeze as incremental investment returns are diminishing. (2) 4G LTE data usage surpassed six-year peak on 3G data in six months. (3) The Internet of Everything is impacting Wi-Fi performance—the proliferation of Wi-Fi devices is lowering service quality on complementary Wi-Fi network platforms. There is new evidence that users are increasingly switching from Wi-Fi to LTE networks for improved quality of service. This should exacerbate the spectrum scarcity problem in the U.S. market. (4) Capex outlook—network operators cite a lifelong focus on network quality management. We see a macro LTE focus continuing in FY14 followed by a small cell densification focus in FY15 as radio networks shift to being built from an indoor out perspective to manage the majority of consumption, which has shifted indoors. (5) While indoor DAS deployment continues, the industry is approaching questionable economics for additional cell densification indoors. The cost to move indoors could be pushed to the enterprise as indoor consumption pressures Wi-Fi networks. Integrating Wi-Fi with LTE is key—we profile the Ericsson managed Wi-Fi/LTE DOT solution, which has stopped network operators deploying sub-scale small cell solutions. DOT service provides a voice coverage solution for the enterprise and a data capacity management solution for the network operator. LTE capacity scales better with the Ericsson DOT service because it provides more capacity without affecting outdoor coverage and capacity, a compelling solution for the enterprise, in our view, with implications for Ruckus Wireless, Aruba Networks, and Cisco. (6) LTE networks with solid foundation layers are seeing dramatic improvements in network quality, which is serving as a differentiator. Telstra network deployment and performance results provide useful insights for Verizon Wireless and the U.S. industry as Verizon adds equipment based on 1.7 GHz 4G spectrum to its 700 MHz 10 MHz x 10 MHz network. Recall that Telstra deployed 3G on the 850 MHz cellular band and has now deployed 4G LTE on the 1,800 MHz band to major cities equivalent to 85% of PoPs in either 20 MHz x 20 MHz or 15 MHz x 15 MHz configuration. Similar to Verizon Wireless, Telstra is benefiting from (a) a healthy device ecosystem—the proliferation of 1,800 MHz–based 4G devices and (b) dramatic network performance improvement as the 3G network quickly de-loads. As a result, Telstra expects to progressively shut down the 3G network more quickly, which bodes well for 1.9 GHz 3G spectrum re-farming at Verizon Wireless. (7) The ability to differentiate and monetize class of service may be on the horizon—enterprise demand for QoS-based wireless 4G service is substantial. This includes the potential for spectrum to be leased to certain industry verticals to create robust private networks in low-density regions (e.g., mining vertical) that AT&T sees an opportunity to differentiate through the creation of 4,600 distributed cloud network compute and storage nodes that can provide differentiated classes of services relative to approximately 50–100 peering interconnection points for providers such as Amazon Web Services or Akamai. These distributed network hubs are likely to provide lower latency capabilities especially for wireless users. For non-U.S.-based network operators, peering interconnection points are likely to be even more limited. We see constructive signs that OTT mobile video and Internet applications seek paid peering agreements that should progressively increase over time, but the key question will be whether a network operator’s revenue generation from API-based OTT applications (e.g., Netflix on a deeply cached video platform or a WhatsApp voice service on a prioritized, virtualized VOLTE platform) can offset the additional network investment required, especially to manage a proliferation of video services. Bottom line, we recommend investors discount capex savings potential. (8) Real-time charging platforms suggest potential service disruption to Ericsson and Alcatel. For example, Matrixx provides a real-time online charging system deployed as an efficient virtualized customer care application on one rack of servers. The application shows the consumer in real time what amount of data is being used. It also provides significant flexibility to the network operator by simplifying the process to change plan tariffs and structures. Trials are underway at Swisscom and Telstra. Recall that traditional postpaid and prepaid databases cannot update usage real time and are deployed as a platform on floors full of a hundred racks of servers. Third-party applications such as these suggest prepaid and postpaid billing system disruption to Alcatel and Ericsson.


Industry consolidation update and M&A perspectives. Softbank/Sprint–T Mobile US. We held very constructive discussions with regulators, SoftBank, Sprint, and T Mobile US that reinforce our views on consolidation. Interestingly, MWC management invited Softbank, Sprint, and T Mobile US to participate in the conference, but all declined. We are bullish on this M&A opportunity despite negative rhetoric, which we believe is being taken out of context. We believe Verizon is concerned about a deal primarily because it is worried about integration risk and regulatory oversight increasing on dominant players (if the integration is poor) at a time when (following the Supreme Court ruling on net neutrality) the industry appears to be at an inflection point where there is a reemerging nexus between access and class of service, along with more constructive direct relationships with application service providers. In contrast, we think AT&T is concerned about the disruptive impact of SoftBank/Sprint/T Mobile on its free cash flow outlook (though we see major differences between Japanese and U.S. markets). We were encouraged to hear equipment vendors including Ericsson, NSN, Samsung, and Alcatel agree with our thesis that the current investment cycle to stabilize assets in the short run with four players is unsustainable.

Thoughts on AT&T/Vodafone/Equinix. During the week, investors in attendance in meetings with AT&T management highlighted that the company indicated a willingness to do an M&A deal in FY14 that could impact its dividend in the short run should an M&A deal be sufficiently accretive. AT&T management also confirmed a plan to build a distributed cloud infrastructure to 4,600 wireless network nodes served by owned versus leased fiber over the next three to five years but is challenged by the fact that it is difficult to build in key locations. This suggests Equinix could be a strategically significant asset to AT&T. Interestingly, MWC also invited AT&T management to participate, but the company declined despite the fact that CEO Randall Stephenson and CFO John Stephens were in attendance.

Thoughts on DISH Networks. Charlie Ergen spent time at the conference meeting with equipment vendors to better understand where the spectrum industry was headed and what options may be available. Recall that Charlie Ergen owned debt in LightSquared and while there may have been bylaw issues, the Glasnost community (only 15 MHz away from LightSquared uplink spectrum below 1.6 GHz) has exerted pressure on DISH regarding interference concerns, and the device community does not think it can build 120db of isolation. LightSquared believes it can be overcome, but conversations with L2, a Mexican satellite company, suggest there is much uncertainty there. The good news is that once a band plan is determined, a network hardware spin has shortened from 12–18 months to 9 months, and the limiting factor is the lack of a device ecosystem to leverage the band. We believe options for DISH to monetize spectrum may be enhanced by industry consolidation between Sprint and T Mobile US. Aside from any regulatory conditions imposed to sell Clearwire spectrum, if Sprint does not create a unique culturally aware solution through network deployment and a major marketing and branding campaign, it will not load up the spectrum. It would then not be surprising that it would sell it, as according to vendor checks, the need is so big and the amount of spectrum is so massive. Furthermore, to combine Sprint and T Mobile, an enormous amount of money is required to change the business model and merge the two companies. This costly challenge would be balanced against the benefit of selling some spectrum. Note that the company is deploying a split mode radio in Band 41 (i.e., 2x60 MHz TDD), which suggests the potential is there to sell spectrum. Such a move would bode well for DISH networks, in our view.

Thoughts on Sprint/NTELOS. We confirm site preparations are underway, which confirms our overbuild thesis. Our checks at CES 2014 and MWC 2014 in Barcelona confirm that Sprint is likely to move forward with the planning and entitlements phase of the overbuild of sites in the NTELOS area. Web sites following the progress of Sprint’s Network Vision upgrade now refer to RF engineers under contract with Sprint to provide installation and maintenance of Sprint's wireless and wire-line networks who confirm that Sprint is aggressively planning an 1,200 LTE site build in the wholesale area starting in early summer 2014. Site preparations have already begun. Areas would include eastern Kentucky (Ashland area); Huntington/Charleston, West Virginia; and Beckley. The company expects this process to take 18–24 months before the cell site is "on air." Note that Alcatel-Lucent has missed its deadline for the delivery of an 8T8R 2.5


GHz radio to Sprint's line in a split mode 2x60 MHz configuration as required by the specifications laid out by Sprint. This is a chipset issue, which Alcatel expects to be fixed by early FY15. We believe Sprint is skeptical of an interim software fix. These radios should be included in the overbuild of the NTELOS region.

Regulatory insights and updates—Chairman Wheeler focused on strong third provider and concerns about duopoly in enterprise and postpaid segments. FCC Chairman Wheeler attended MWC 2014 and participated in a series of closed-door sessions. (1) Regulatory frustration at MWC switched from public to private this year as major CEOs engaged with their respective regulators in private closed-door sessions. Interestingly, feedback suggested constructive dialogue, with an emphasis on a strong third competitor. Interestingly, there was a continuation of U.S. regulatory antipathy toward the European regulatory approach with limited interest in meetings between the respective camps. (2) FCC Chairman Wheeler acknowledged that the industry is at an important juncture, the marriage of high-speed computing and delivery systems. As a result, the focus is on access interconnection, public safety, national security, and customer protection. (3) U.S. perspective on M&A—Chairman Wheeler is clearly focused on making the right industry policy decision, concerned about the duopoly status of the U.S. postpaid and enterprise segment and the need for a strong, sustainable third wireless operator. We believe the chairman appreciates the difference between vibrant versus relevant and viable networks needed to serve the entire wireless market, not simply the less profitable prepaid sector that is less relevant to U.S. industry and GDP growth. (4) European perspective on M&A—the European regulatory (DG Comp) position on industry consolidation has shifted from impossible in FY13 to “very difficult” in FY14, in our view. Interestingly, the heavy-handed European regulatory oversight was seen by non-Europeans as a reaction to imperialistic-based strategies by European “network operators” trying to control OTT technology versus a non-European “service provider” philosophy of connecting customers to their technology of choice. (5) Peering interconnections versus net neutrality—the Comcast–Netflix agreement was seen by the telecom services industry as a very significant development. The FCC reiterated its light-touch approach by endorsing a monitoring role. The FCC sees the Internet peering interconnection market as still evolving—laying somewhere between the spectrum of market abuse by content providers/intermediary networks and anticompetitive behavior by end-user access networks. Network operators point to a progressive increase in paid peering costs over time. We see limited potential for regulatory intervention. (6) Leading indicators on Incentive auction structure and the uncertainty of participation by the market exposes a lack of effective U.S. industry policy on spectrum versus the rest of the world. Wireless industry points to the superior 700 MHz ATP band providing 2x45 MHz of spectrum and expects the 600 MHz auction to be a delayed failure. Interestingly, a delay in auction timing of this important priority item is not a negative for Chairman Wheeler, as he has yet to own this initiative from ex-Chairman Genachowski. Chairman Wheeler sees a chance to establish a positive legacy by making the right decision on U.S. industry policy to incent sustained investment by approving a Sprint–T Mobile deal, in our view.

Device ecosystem—Is a subsidy discount from high-end LTE handset providers likely? (1) We see increased penetration of low-end Android smartphones with pricing in the $33–$50 range. Globally, smart phones only represent 2 billion of a 6 billion phone base, so this represents a significant opportunity to penetrate legacy feature phones. Phone profiles include the $33 K Touch T619, Spreadtrum 8810 SoC, Android 2.3 Gingerbread, ARM Cortex A5 1.0 GHz, and ARM Mali 300 GPU. Applications are driving smartphone momentum but also hardware application extensions such as wearable technology (e.g., Samsung Gear Fit) and we see different categories emerging. Service providers suggest they were underwhelmed by Nokia's first low-end Android phones—Nokia X, X+ and XL. The UI Interface is closer to Windows, i.e., a Windows-style with Android applications, and oddly enough, the launch was not coordinate, with Microsoft unaware of these Nokia device launches despite the devices having Microsoft cloud one drive access and Microsoft services (e.g., Nokia store—No Google playstore but reference to third-party stores, Microsoft mail and no Gmail) There is no HD screen or high-end camera, and the devices are sold for €89, €99, and €109, respectively. (2) China Mobile is driving device costs down to the $130–


$150 range for low-end TD LTE devices, compared to Samsung at $300, which bodes well for equipment subsidies over time in an increasingly commoditized environment. (3) Qualcomm continues to dominate high-end LTE devices—service providers cited significant network efficiencies, but the cost of 4G LTE devices is the major impediment to profitability (recall that the LTE chipset is $130). (4) Higher-speed Cat 6 devices such as the Samsung Note 3—using the Snapdragon 805 processor and Gobi 9x35 modem—demonstrated speeds at 300 Mbps using a 2x20 MHz channel using two antennas in the device. This chip supports FDD LTE and TDD LTE bands, which suggests the device will be deployed across Verizon, AT&T, Sprint, and T Mobile US, but network differentiation will be a function of 4G LTE coverage. With the introduction of Category 6, adoption of LTE Advanced is expected to accelerate, with a key factor being the availability of commercial infrastructure and global chipsets for smartphones. SK Telecom will be first to deploy this technology. (5) On the device front, while the industry is pointing to commoditization of high-end devices, perspectives on high-end device differentiation differed by service provider. Many saw signs of device commoditization, but some saw contrasting camera performance (particularly low light) and battery life as key hardware differentiators, which they expect to continue. Service providers are not renowned for picking device winners, but to this end, they cited a potential switch in consumer preference from the Samsung Galaxy G5 to the Sony Xperia Z2, which may be a trend to watch. Upgrade cycles are a concern for the equipment and semiconductor sectors, but upgrade cycles will see a boost from U.S. service providers responding to spectrum scarcity by more aggressively migrating customers to new technology to recapture 3G spectrum, which the industry does not believe will be enough to address spectrum scarcity. (6) The forking of Android among Nokia, Google, and Samsung was the biggest surprise at MWC 2014 and is the biggest unknown, with potentially negative implications for Samsung and Huawei, not known for software excellence.

Mobile payments—service providers lose to Google on mobile payments opportunity; just like the loss of control of location-based services. MWC 2014 highlighted that the industry has changed yet again in this important, unregulated segment. Service providers have touted Project ISIS as a potentially positive venture aligned with Visa and MasterCard to capture revenue from the vibrant mobile payments arena. However, at MWC 2014, our discussions with Visa and MasterCard suggest a major shift is underway that should benefit Google and Apple (via iBeacon). Visa and MasterCard announced support for a new NFC-based payment using smartphones. Cards are stored in the Google cloud for NFC payments, and consumers can check out of stores with a phone tap without the need for a secure element in the device that stored the card information. Android KitKat supports this technology, along with BlackBerry 10. The secure element of the device is usually controlled by the service provider, so this development suggests service providers will no longer be able to limit the applications that can access it. To date, Google Wallet has seen limited adoption because of service provider restrictions. NFC payments remain in the early development stages, and it is likely that Apple will do something more with iBeacon, released last year.

Spectrum and Radio Access Network developments. (1) Spectrum challenge intensifying—Spectrum scarcity in the U.S. continues to be more intense than in the rest of the world and is intensifying as incremental returns diminish. Absent industrial policy regarding the 700 MHz band, the FCC (after converting broadcasters from analog to digital) led with a suboptimal band plan (diplomatically put), while the rest of the world now looks like it is enjoying a wider, clean 700 MHz ATP band that offers 2x45 MHz bands. Furthermore, the industry views the incentive auction structure for the 600 MHz band (where auctioned spectrum varies by region) as likely to be a failure and be delayed. Wi-Fi will continue to play a critical complementary role with managed Wi-Fi displacing basic Wi-Fi solution providers. (2) Low band spectrum is being relied on by 3G network more than expected—Leadership in Energy and Environmental Design (LEED) is a set of rating systems for the design, construction, operation, and maintenance of green buildings, homes, and neighborhoods. LEED-compliant windows adversely impact the ability of a radio signal to be received indoors, especially in bands above 1.5 GHz. While low bands get indoors, the attenuation drops, and coverage needs to support a wide enterprise area. Network operators have evidenced a greater-than-expected reliance on low band spectrum as a result. (3)


Integrating Wi-Fi into the LTE experience through managed Wi-Fi solutions is key, and the consolidating cable industry now has the third-largest Wi-Fi-based wireless network and should be able to continue to price up this lower-cost advantage. (4) In a scenario where Sprint and T Mobile combine, there is potential for SoftBank to sell (or be mandated to sell) CLWR spectrum, which would be snapped up by AT&T or Verizon, along with DISH spectrum, in our view. DISH should have a clearer path beyond the AWS 3 auction and Sprint T Mobile combination. (5) LTE Advanced is here and being deployed for service differentiation, but LTE Broadcast needs to work to conserve spectrum. All Qualcomm devices support this technology, and we expect business models to quickly evolve, as this is needed to help manage peak usage; other applications include software downloads from companies such as Google Apple and Microsoft. (6) Cloud RAN—meetings with network operators cited major performance gains possible from Cloud RAN networks to help differentiate at the network layer. As a prerequisite, fiber-based backhaul circuits are required, and owner economics are key. We see limited major cloud RAN deployments at this time, even at SK Telecom. However, this will be a necessary platform for 5G services. (7) Sprint network update—the network has very good signal strength, but major interference and significant optimization is needed. Alcatel’s 2.5 GHz radio solution has been delayed until FY15. A chipset issue is limiting the 8T8R radio to 2x20 MHz of capacity versus 2x30 MHz of capacity. Softbank is moving management resources to the U.S. to take over network management in key markets and will adopt a user-based approach to network performance, in the same way that it used Twitter in Japan, which saved SoftBank. We see potential for Ericsson to displace NSN in existing Ericsson markets due to the complexity of Sprint’s radio solution that will require a single vendor solution by the market to optimize features and performance between bands. (8) T Mobile network update—Industry checks suggest two years of network capacity, but LTE usage profiles are driving an upward capex bias, in our view. Unlike Sprint’s plan for a combined Sprint and T Mobile, which is to leverage the density of a combined grid to drive revenue synergies, the company is focused on opex efficiencies by turning down MetroPCS cell sites aggressively as it adds small cells and densifies the network as part of its capex spend plan in key areas today. (9) U.S. service providers cited intensifying concerns regarding the outlook for Alcatel-Lucent, particularly as the architecture shift gains momentum. Globally, leading network operators are generally supported by leading RAN vendors: Ericsson, NSN, and Huawei.

2020/5G networks update—Is wireless power on the horizon? MWC provided an update on fifth-generation mobile networks, which are the next major phase of mobile telecommunications standards beyond current 4G standards. No specification has been published by any standardization body to date, but new capabilities are being grouped. Two companies highlighted examples of techniques that are being proposed for upcoming 5G systems—Kumu Networks and pCell. While the current 4G family of technologies focuses on minimizing interference to maximize output and optimize the user experience, these technologies embrace interference and focus on interference cancellation technology to eliminate interference in real time to provide much faster, more consistent speeds and with lower latency. Required 5G infrastructure includes (1) base station radios, (2) cloud (C-RAN) infrastructure, and (3) backhaul fiber, which could take five to 10 years to develop. We summarize each below:

Kumu Networks has created a potentially disruptive 5G spectrum technology (but can also work within current 4G standards) based on interference cancellation technology that allows radios to send and receive data at the same time on the same frequency. The technology, known as Full Duplex wireless, applies to every radio (Wi-Fi, LTE, etc.) in every frequency band. Because of its potential to double the capacity of wireless networks, address spectrum fragmentation issues, and unify TDD and FDD standards into one coherent standard, this technology is currently being considered as a foundation for next-generation 5G and High Efficiency Wi-Fi standards. Even without any changes to the standards, Kumu’s Full Duplex technology can solve 4G LTE and Wi-Fi interference and spectrum fragmentation challenges, which today cost the industry billions of dollars. Specifically, Kumu allows existing radios to use any set of spectrum fragments simultaneously (carrier aggregation) without limitations of fixed hardware filters. The technology also enables high-capacity mesh networks, full


duplex microwave backhaul links, high-capacity multi-channel Wi-Fi APs, and other valuable use cases that are completely within the boundaries of the current standards. By eliminating all self-interference from inside the device, a radio can transmit and receive at the same time on the same channel—effectively doubling the efficiency.

pCell has also created a potential 5G spectrum technology called Coordinated Multipoint technology. Each user can utilize a dedicated frequency channel and can use it up to the full Shannon limit, in contrast to 4G technology, where frequency bands are shared between users and multiple antennas are used to increase throughput.

5G wireless power and implications? Higher throughput suggests 5G wireless solutions could displace both cable and DSL connections or drive a technology upgrade cycle. Wireless power a game changer? Through the synthesis of a tiny radio-wave bubble in real time software, pCell technology could open up a new wave of applications. Specifically, wireless power transmission could be feasible with microwave beams using a multitude of less expensive beam-forming antennas (versus few expensive beam-steering antennas) to transfer non-harmful DC power electricity and convert this to RF energy in end-user devices in real time at multiple locations across a wireless network area, versus between two places within line-of-sight distance. While 5G is a big leap in performance from 4G technology, it does not provide any fundamentally new capabilities. But ubiquitous wireless power has major implications for social utility, including the consumer electronics industry, the auto industry through increased electric vehicle usage, and the oil industry through reduced oil demand.

Wired access network—Google fiber update. Checks suggest Google’s cheaper access network design may indicate it is more interested in gaining leverage over telecom and cable sector versus scaled-up deployments. We see evidence of an interesting engineering conflict between Google and network operators. While software engineering talent at Google dwarfs the service provider segment, our MWC checks have network providers suggesting that Google may be naïve as the company is using low-cost enterprise technologies to build a carrier network to end users. For example, the company is using active electronics and equipment to serve end-user customers that will lead to higher opex costs at scale. Google may have a model that justifies initial capex investments, and certainly the lack of regulatory oversight and use of non-unionized labor is a big advantage. But such deployments could also be justified to help form the basis for greater leverage in interconnection discussions, as opex costs will be too high to justify a scalable deployment over the next five to 10 years. Bottom line, MWC checks suggest Google is more likely to be seeking to exercise more leverage over cable and telecom network providers versus providing a material threat to these respective industries from a scaled-up deployment—which was similar to its strategy to gain more leverage in the 700 MHz auction. Service providers are clearly seeking to engage with application service providers on a rational basis. A rational basis means how much it costs to build and operate a carrier access network as Google will not be able to create a total cost of ownership that is lower than Comcast. Fiber access is a game of utilization and scale economies. Companies such as Facebook have been more constructive with respect to paid peering arrangements but are a long way from covering the incremental network costs, which will likely drive the architecture shift and a progressive increase in interconnection costs.

Metro and backhaul network update—a move to regain owner economics from leased fiber infrastructure. MWC checks confirmed a push by AT&T and Verizon Wireless to regain owner economics in fiber backhaul as LTE data consumption is tracking higher than expected. We see both companies seeking to serve a more distributed hub of network nodes (4,000–5000) with owned fiber. Additionally, both companies cite major performance gains from 4G advanced and 5G ecosystems, which will require infrastructure that includes (1) base station radios, (2) cloud (C-RAN) infrastructure, and (3) backhaul fiber, which could take five to 10 years to develop. Specifically, we see Verizon Wireless planning to deploy dark fiber through Zayo Bandwidth to 10 out-of-region states covering 10,000 miles. Similarly, we see AT&T deploying fiber to its wireless network hubs, which is well underway. Interestingly, Cat 6 (LTE Advanced) devices already have network operators dialing backhaul VLAN pipes up from 130 Mbps to 300 Mbps. This suggests that the lack of a fiber-rich wireline capillary network at Sprint and T Mobile will impact their


ability to scale the businesses and provide differentiated performance over time, further marginalizing these franchises absent consolidation or re-regulation of fiber backhaul services, which we do not expect.

Datacenter networking update—the challenges of virtualized machines are driving upward pressure on capex in the near term. New architectures create a standardized compute environment and should provide operational flexibility and improve customer service: MWC checks confirm that the major challenges for the next five years will be the changes to service provider organizational structures needed to execute these changes to create standardized compute platforms across networks and datacenters.

As detailed in the next section, AT&T rolled out its architecture for a next-generation network—the user-defined network cloud with three objectives: (1) A call to the industry to accelerate pace of change; (2) a request for outside help—recruitment of distributed software and network skills; and (3) broadcasting the plan—i.e., getting AT&T leadership, getting people behind it, and getting designs done. We applaud the initiative from AT&T, which to some extent is forced on the company. This is because in the event that the cloud is dark (i.e., dominated by over-the-top service providers), the revenue opportunity for network operators will be very limited, and the network expense pressures will be very high due to the challenge of virtualized machines being accessed dynamically (repeatedly) across the network in multiple geographically dispersed datacenters.

Service providers are having mixed results at turning inside out. Results at Telefónica Digital suggest avoiding a dark cloud is challenging. Telefónica Digital is the global digital services and innovation arm of Telefónica, one of the world’s largest telecom companies with operations across Europe and Latin America and over 320 million customers. Telefónica Digital was created to lead Telefónica’s transformation into a digital telco as the engine for bringing digital products and services to the market, including developing new technologies for consumers to communicate with friends and family through to helping businesses and governments address new opportunities, improve operations, and increase efficiencies. The company cited an entrepreneurial approach to identify and accelerate the latest technology trends, incubating new disruptive digital businesses, delivering end-to-end digital products and services in B2B, B2C, and B2B2C in a number of domains: future communications, machine-to-machine, financial services, media services, cloud computing, and information security. The company also cited the development of the right partnerships to be the single source of the best digital experiences for its customers. However, while the company combined assets under Telefónica Digital, then created a series of over-the-top API-based applications, our MWC checks indicate that these applications were absolute junk.

The major impediments to momentum we see at AT&T (as a proxy for the service provider sector) are as follows: (1) The company cannot build datacenter fabric with end-to-end visibility where needed due to lack of space, which suggests acquisitions may be required; (2) AT&T has a legacy networking culture and an insufficient set of change agents such as John Donovan; (3) AT&T network management has a risk-averse approach to new technology; (4) AT&T has a significant lack of software engineering skill sets; and (5) AT&T employees have a heightened concern regarding job loss.

For FY14, we see AT&T benefiting from major price reductions from its legacy vendors (Cisco, Juniper, VMWare, etc.) due to the threat of the migration to next-generation networks. Saving are being invested into attempts to accelerate this migration, but major organizational structure changes will be required to align datacenter and networking silos under change agents similar to John Donovan, along with support from the senior leadership teams.

Meanwhile the challenge of Virtual Machines in the datacenter is creating havoc with networks. Inside the datacenter, many companies are shaping how SDN is used to build next-gen datacenter networks for hyperscale companies, telecom service providers, and enterprises. Essentially the datacenter is being redesigned as a highly scalable, standardized compute platform using commoditized hardware and a standardized protocol (OpenFlow) to manage this


hardware, along with APIs to manage the interface between applications and virtualized network functions on openstack software platforms.

Major enterprise CIOs want to automate all their datacenter assets to provide very rapid deployment of business line applications to limit competition from Amazon Web Services and other distributed cloud service providers. They also seek comprehensive policy control of compute and network architecture to authenticate access across all datacenters and service templates for different applications that can be done once and spread out many times—e.g., setting up QoS, filters compliance rules with permissions on what to access. So, if there is an issue with the network or a change needed, it is done once. Adding programmability and automation to turn virtualized machines instantly in the datacenter is key. Solutions that can address virtualized and non-virtualized assets in the datacenter and can solve other networking issues will gain the most traction, in our view.

Outside the datacenter, telecom service providers are seeking to build their compute and networking infrastructure using white boxes built on top of merchant silicon with open-source control plane and value-added applications. We saw this trend first applied in the datacenter (OCP), and SDN and ONF have then gained traction and are being extended from datacenter network to enterprise and service provider networks. ODL is trying to build an open-source SDN control plane but trying to preserve legacy networks. Unconstrained network virtualization has six challenges: (1) It is hard to federate and scale architecture; (2) routing protocols introduce management complexity; (3) overlay with underlay visibility is not clear; (4) policy and abstraction tools for end users is unclear (What APIs are needed to allow end users to configure SPF [to hang themselves], or does the vendor provide abstractions to allow end users to focus on applications?); (5) visibility and assurance (thousands of tunnels created for virtual networks); (6) physical end points—the reality is that not all the world is VMs, and there are many physical endpoints, interconnections to hosted and co-located end points.

The multidimensional network virtualization solution starts with a simple core and intelligent edge design to decouple services from transport—a simple core allows the core to scale. This can be achieved by starting an overlay network using hypervisors and letting them talk to each other. In this way, any party’s IP transport is the fabric. Tunneling (VXLAN encapsulation) is used for two reasons: (1) IP address—namespace isolation to keep enterprises’ IP addresses moved in separately—and (2) mobility. With regard to intelligent edge, there is complexity associated with access controls, ACL, and QoS out at the edge. But the interesting question is what the control plane is (what tunnels are used is less interesting). The Internet approach is to have a control plane on every VMM/hypervisor run with a protocol in between—it has Internet properties but does not scale well (bare metal hypervisors run direct on the hardware, hosted hypervisors run as second level software similar to conventional o/s software). A second SDN view is the use of a single SDN controller to run all hypervisors—this has a limit to scale and is subject to software bugs (protocol is OSTP and OpenFlow). MWC checks cite a middle way—with domains and federated controllers. This can be achieved using IP VPNs (BGP MP—to support multiple services) to create a scalable server architecture. Visually, a big router has a control plane, line cards, and fabric. Lobotomizing the router is achieved by taking the control plane out of deployed routers and making it a virtualized machine—federated, so now instead of hardware line cards in chassis, they are hypervisor switches that instead of an IP fabric create a “big router” on a scalable IP network. As we have discussed, these developments have substantial implications for incumbent vendor relationships with Cisco and Juniper.

Peering and interconnection update. Online video and internet application providers are entering into more constructive, direct relationships with end-user networks as it becomes apparent that low latency services do not scale as well as expected due to poor quality of service (e.g., Skype). We see an increased likelihood that over time low-latency services such as WhatsApp voice will seek a quality of service relationship with end-user networks on their virtualized VoLTE platforms. We see negative implications for intermediary networks such as Cogent and Level3 Communications.


MWC Takeaways: Equipment Sector Scott Thompson

646.885.5441

Facebook’s Mark Zuckerberg reveals carrier-like ambitions. For the second time in a month, Facebook's Mark Zuckerberg used the keynote pulpit to reinforce his enthusiasm for deploying basic Web-based services to underserved corners of the world. He revealed that Facebook would like to work with two or three large carriers to deploy minimal application services over the next year, then launch into longer-term, larger-scaled trials over the next few years. Facebook may not be the only Web scale platform pushing traditional telco carriers in uncomfortable directions that combine IT infrastructures, focusing more on compute and storage and less on networking. The rest of the week revealed several other drivers of substantial change across the communications landscape. Web scale business models are setting the pace for change across the telco and communications equipment landscape. It is important that investors understand the architectural changes that Facebook and Google are driving in the networking space. The changes that they are pushing for are dramatic and will have a lasting effect on the sector’s landscape.

AT&T’s user-defined network cloud accelerates shift from traditional telco architecture. AT&T’s John Donovan revealed his vision for delivering the network Facebook needs to achieve its lofty goal of connecting the world. The carrier unveiled its “user-defined network cloud,” the next step in its well-orchestrated Domain 2.0 program. The company also announced the first group of four companies selected to work on the strategy: Ericsson, Tail-F Systems, Metaswitch (all three engaged in design and development functions), and Affirmed Networks (to work on a virtualized evolved packet core). Ericsson will also be involved in integration and transformation services. Donovan hinted that AT&T would rely on small vendors for innovative solutions, mid-size vendors to coordinate end-to-end solutions, and larger vendors to achieve cost benefits through scale. While AT&T has stopped short of providing any view on changing capex trends, we expect that the first wave of AT&T’s Domain 2.0 could be targeting as much as 20% declines as early as 2016. Traditional large-scale networking vendors could struggle to keep pace with AT&T’s changing capex trends. We expect that CIEN, INFN, ERIC, and ALU could benefit, while vendors such as CSCO and possibly JNPR could struggle in the new ecosystem. Mr. Donovan published a white paper on his company Web site providing details on how AT&T is shifting its business lines, including virtualizing both residential and enterprise edge devices (carriers spend approximately 60% of communications equipment capex on the edge).

AT&T Domain 2.0 white paper provides rough outline of carrier’s path to network innovation. AT&T’s SVP of technology and networks, John Donovan, made public AT&T's plan to transition its network to a user-defined cloud ecosystem in a white paper posted on his blog. One key aspect of the program, from a networking perspective, is the open and ongoing procurement process that the company intends to adopt. AT&T's white paper reinforced the company’s goal of replacing the existing network elements that currently perform single functions with servers that support (1) NFV, (2) packet forwarding capabilities on merchant silicon (i.e., “white boxes”) at the datacenter and the network edges, and (3) the virtualization of the network edge (enterprise and residential) by running virtual instances of the logical devices on aggregation routing platforms in the network (MX and 7950). We believe companies whose products are well suited to deliver on these requirements, particularly those in the optical-layer space, can benefit from AT&T’s infrastructure transformation in the future.

The difference between a router and a switch is up for debate by industry experts. We were shocked to hear that one of the most hotly debated points by industry experts was what constitutes the real difference between a router and a switch. Today, new developments in the realm of data networking have relegated the difference between the two somewhat inconsequential. The battleground for this debate begins in the datacenter where Layer 3 ECMP ports are the fabric that hold the transport layer together, while engineers work diligently to make the network riding on top if it as flat and simple as possible. This network ends up looking


more like a switching network than a routed network and, due to aggressive price compression, is priced at switch-like levels. The result is that the industry is moving to eliminate many of the differences between the switching and routing layers, creating a more simplified transport layer. We expect this debate will be key in truly understanding the recent decline in Cisco’s core business. Investors should take note.

Broadcom’s Trident II is a BIG deal. We have become more impressed by the simplistic brilliance that Trident II might possess to turn the networking world on its head. The Trident II ASIC can combine traditional routing and switching layers into one platform, performing many common Layer 2 and Layer 3 protocols in the same chipset. This enables the white-box platforms to both collapse and enable OEMs to commoditize both layers of the OSI model. We believe that the Trident II ASIC platform could actually be more near-term disruptive than SDN or NFV. It is a must-know component of the evolving communication equipment ecosystem.

Intel expects to begin monetization of networking in 2014 through 2016. Intel's Rose Schooler revealed that Intel intends to pursue the networking sector through potentially disruptive cloud-based technologies but on a much smaller scale than we had at first expected. The company appears enamored with the strategic implications of the worlds of compute and networking combining. Intel suggested that investors keep a close eye on how vendors in both networking and traditional compute will defend existing market share while working to gain share in nontraditional ways (i.e., moving from compute into networking and vice versa). Intel suggested it might be able to capture 5% of a $16 billion networking TAM by 2016. This does not include much of the rackscale architecture with which Intel has been making waves.

HP adding arrows to the SDN quiver. No longer a distant player in the routing and switching market, HP has evolved its product to something different. The company is talking about (and selling) solutions like NFV, SDN, virtual routers, virtual switches, virtual firewalls, virtual EVC, SBCs (via SONS), RAN, etc. Over the past two years, HP has been busily virtualizing the edge of, and many of the services delivered by, the network. It has moved beyond the traditional networking model and into a world that revolves around software, x86-based compute platforms, and only an essential amount of network to make it all communicate. The booth was buzzing about solutions for both enterprise and service provider. Why is HP betting so heavily on these disruptive solutions? Is the solution a pipe dream that will never scale? With the buzz around the products and the size of the accounts at the booth, we doubt it.

Tail-F. What do we say? It is named after a Linux command, and AT&T loves them for it. We spent a considerable amount of time at the Tail-F booth. Here’s the gist, after running a few checks and looking in-depth at the product: Tail-F is a service orchestration company. It basically takes every command-line interface and recreates it to allow for service provisioning over multiple hardware platforms in a drag-and-drop format. While the product is obviously what AT&T is interested in, we remain concerned that the time lag involved in recreating every additive feature, fix, and function for every device across every vendor in the communication equipment ecosystem would be difficult to manage successfully. Then again, maybe carriers only use 20% of the features included in various OSs across the industry to deliver services. Automating 20% of the CLI for multiple platforms is much more manageable and reasonable. Our checks reflect Tail-F deploying a large-scale proof of concept deployment with AT&T in 2H14 in Texas. We will be watching closely to see if that project rolls out as planned.

Alcatel-Lucent's momentum in the NFV category appears to be accelerating. A quick conversation with Alcatel left us pleasantly surprised with the momentum ALU is gaining with its NFV products. After announcing its virtualization trial with China Mobile (virtual base station, EPC) on 2/19, ALU's CloudBand unit showcased its virtualized IMS, RAN, and EPC products at Mobile World Congress. General availability of the products will take time; however, we expect that if ALU maintains its current momentum, it could be one of the first credible vendors to market with some game-changing virtualized technologies.

Affirmed looks to be the real deal in the EPC; other virtualized functions are on the way. We also spent some time kicking the tires at Affirmed, another of AT&T’s early picks in the Domain


2.0 RFP process. This company, at first glance, appears small but solid. It has 30 trials, 18 POCs around the world, and is live and scaling its virtual EPC at one carrier. The live carrier has a cap-and-grow strategy and is transitioning new traffic off of Ericsson. Affirmed emerged from stealth mode in November 2012 and has been working closely with carriers to develop virtual services on x86-based platforms for nearly four years. This stealthiness creates a challenge to judge how long it takes (normally three to five years for a start-up to make a dent) for newly announced and disruptive technologies to gain traction.

NEC the quiet SDN giant of the show. While we did not make it to the booth this year, we had multiple vendors point us in the direction of NEC. Accolades generally revolved around NEC having the most SDN and NFV solutions that are getting real-world traction today. Apparently, NEC has a platform that is gaining more traction across the globe than nearly any other. Our checks also show that they have an integrated solution that spans from the datacenter to the metro edge. For a preview of what Cisco’s product may look like when fully baked, we recommend investors take a look at what NEC has been able to deliver. We think this quite giant could be making a move to gain serious share in the new world where networking and compute collide.

MWC Takeaways: Semiconductors Christopher Rolland

646.885.5405

GS5 was the crown jewel of MWC handset announcements but lacked differentiation. Samsung’s flagship refresh offered only modest improvements over its predecessor. For our semiconductor companies, the phone will feature Broadcom’s just announced BCM4335 combo connectivity chip featuring 802.11 ac Wi-Fi 2x2 MIMO (and the company’s discrete GPS), which we believe will drive $315 million in revenue. Additionally, non-Exynos versions of the GS5 will sport Qualcomm’s Snapdragon 801. Other hardware improvements include a swipe fingerprint sensor (likely Synaptics) and heart-rate monitor (perhaps Maxim). Earlier this week, we introduced our first estimates for production and sales of the Galaxy S5, which we are now targeting at approximately 70 million builds and about 67 million in sales (above our estimates of 67 million and 64 million respectively, for the GS4).

Nice show for Qualcomm, but we grow increasingly concerned given new LTE entrants in 2014; however, valuation dictates our Outperform rating. Marvell, Intel, Broadcom, MediaTek, and NVIDIA all have announced new LTE platforms coming to market throughout 2014. Reports of a GS5 built around and Exynos processor and Intel’s LTE also made the rounds. That said, valuation remains favorable, as QCOM trades at 11x our 2015E EPS (pro forma including stock comp expense, excluding net cash and interest income) and below the industry average of 13x.

Intel announces Cat 6 LTE-A baseband part but interestingly lacks CDMA. Intel revealed more details surrounding its XMM7260 baseband featuring Cat 6 LTE-Advanced with carrier aggregation and two downlink channels for up to 300 Mbps data speeds. That said, it does not offer backwards compatibility for CDMA, which we believe limits its overall TAM (for example, handsets at both Sprint and AT&T will require this until they move to an all-LTE network).

Marvell’s cellular story marches on; lawsuit keeps us sidelined for now. Marvell announced the new 64-bit Armada SoC, which will feature ARM’s Cortex A53 processors coupled with a 5-mode LTE modem. The PXA1928 will begin sampling in March and helps solidify Marvell’s market position in 2H14 by adding a powerful AP to its 5-mode baseband, including LTE. We would like to be more constructive on the name but are compelled to wait until the risks from the lawsuit are quantifiable.

Tegra announcements fall flat. There were only two Tegra 4i announcements at this year’s show: (1) LG G2 mini, primarily for the South American market, and (2) the Wiko “WAX” for some European carriers, which will begin shipping in April. While we expect more design win announcement as the year progresses, the start was disappointing. We continue to believe that


NVIDIA should discontinue Tegra’s cellular development and focus purely on embedded opportunities if it fails to generate significant interest with this generation.

For semiconductor stocks overall, the SOX was up more than 30% in 2013, and we believe the early stages (call it the first six innings) of the recovery are already reflected in share prices. While supply chain inventories are lean, they are not as lean as their 2009 trough; therefore, this “cycle” may lack the benefits of aggressive inventory restocking, increasing the likelihood of a stop-and-start expansion for chip companies. Despite the near-term choppiness, we believe the recovery looks rather typical, as it appears we have entered the expansion stage of the cyclical recovery. Within this context, our favorite longs are ARMH, AMD, AMCC, BRCM, QCOM, INTC, and ONNN.

Intel Corporation (Outperform – $27 Price Target) Intel’s tablet wins march on, but phone wins absent; LTE chip appears competitive. Intel

demonstrated incremental design win traction for both tablets as the company works toward its goal of 40 million tablets in 2014. Intel announced multiyear tablet and smartphone processor agreements with several OEMs: Dell (expanding upon the win in the Venue line last year); Lenovo, which introduced plans for a new Intel-powered mobile device this year (including LTE connectivity in the Ultrabooks); Foxconn (remember it will be manufacturing Blackberry handsets); and Asus (also announced a few models will be refreshed with Intel’s baseband). That said, there were no immediate design win announcements for either Merrifield or the older Medfield platforms in Barcelona this year. According to CNET, Intel told its reporter that it had won an LTE socket in a Samsung GS5 SKU but later retracted the statement and said it was a later variant of the model expected for release later in the year.

MWC 2014, INTC: High-Profile Smartphone and Tablet Announcements

Source: FBR Research

New smartphone platforms Moorefield and Merrifield interesting but waiting for Broxton. Intel provided more details around its 22nm Merrifield and Moorefield platforms. However, discussions with the design channel revealed lukewarm reception with more interest in the company’s Broxton upgrade in 2015. Perhaps more important than the company’s mobile processor announcements was its LTE roadmap, which includes the XMM7260 featuring Cat 6 LTE-Advanced with carrier aggregation and a pair of downlink channels for up to 300 Mbps data rates, expected to come to market in the first half of 2015. Like Tegra 4i, we note that this part however, does not offer backwards compatibility for CDMA, which we believe will therefore limit adoption, particularly in North America as carriers including Verizon and Sprint utilize the technology. While the lack of CDMA is somewhat discouraging, as it decreases the overall addressable market (and particularly for high-volume SKUs), Intel expects several design wins

Device Name Announced Date Part AP

Sample -HP Elite Pad 1000 2/24/2014 Bay Trail Atom Z2760

Dell, Lenovo and Foxconn tablets 2/24/2014 ? ?

Asus ZenFone 2/24/2014 Clovertrail+ Atom Z2580

Prestigio Multipad Thunder 2/15/2014 Clovertrail+ Atom Z2580

Cloudfone CloudPad 800w 1/30/2014 Clovertrail+ Atom Z2580

VivoTab Note 8 1/6/2014 Bay Trail Atom™ Z3740

The HP Pavilion x2 11t 12/26/2014 Silvermont PentiumN3510

Hp 810 EliteBook Revolve G2 12/12/2013 Haswell Core i3,i5,i7

Sony VAIO Tap 11 11/14/2013 Haswell Core i5-4210Y

Surface Pro 2 10/22/2013 Haswell Core i5 4th gen

Venue 8 & 11 Pro 10/8/2013 Bay Trail Atom Z2760

Asus fonepad 7 LTE 2/24/2014 Clovertrail+ Atom Z2560

Asus Padfone mini 2/24/2014 Clovertrail+ Atom Z2560


this spring and should help the company move one step closer to its stated goal of becoming number two in market share for LTE chips by the end of 2014.

Intel gives update to its roadmap. Below is FBR’s comprehensive product roadmap for Intel, which includes an update of all mobile products released at MWC.

FBR’s Perspective on Intel’s Roadmaps


We rate INTC Outperform with a $27 price target. While headwinds in PC have been discouraging, computing form factors will change over time. However, we believe that Moore’s Law is considerably more durable than any one form factor and presents as good a business plan as any in the technology industry. Moving forward, we are increasingly confident that Intel can opportunistically extract value from the extra transistors afforded to it through the best silicon manufacturing operations in the world. We rate INTC shares Outperform with a $27 price target, based on a 13.5x target multiple of our 2014 EPS estimate (including stock compensation), a low but appropriate multiple given peer valuations and investor concerns about future growth.

Marvell Technology Group (Market Perform – $15 Price Target) Marvell makes waves with new 64-bit chip; company pushes a more complete phone

platform. Marvell announced the new 64-bit Armade SoC, which will feature ARM’s Cortex A53 processors coupled with a 5-mode LTE modem. The PXA1928 will begin sampling in March and helps solidify Marvell’s market position by adding a powerful applications processor to the 5-mode baseband, including LTE. We are increasingly encouraged by the prospects for this platform, and it should provide a reacceleration in Marvell’s mobile shipments in 2H14.

Below is our comprehensive roadmap of product launches as highlighted by management.

FBR’s Perspective on Marvell’s Roadmaps


While 64-bit announcement was exciting, there was a lack of Marvell wins announced at the show. A comprehensive list of near-term wins can be found below. However, there were no wins announced at MWC.

Intel Roadmap 1Q12 2Q12 3Q12 4Q12 1Q13 2Q13 3Q13 4Q13 1Q14 2Q14 3Q14 4Q14 1Q15 2Q15 3Q15 4Q15

Premium Desktop

Desktop Platform

Premier Desktop (i7) SB

Mid-range Desktop (i5) SB

Entry Desktop (i3)

Value Desktop (Celeron)

Value Desktop Atom Cherry Trail Broxton

Mobile Platform

High End Mobile

Mobile Premium (i7) SB

Mobile Midrange (i5) SB

Mobile Entry (i3)

Mobile Pentium/Atom

Mobile Atom

Atom

Atom Core Architecture Saltwell

Smartphone Merrifield 64b [email protected]

Smartphone - Value Lexington SoFIA 3G

Tablets Bay Trail - 64 bit

Comm. Proc. / Other Rangelely - 64 bitMicroserver CPUsBaseband XMM 7260 LTE Advanced - CarrAg + Cat 6

Server

Server Platform Romley (Patsburg Chipset) Brickland for E7 (3 generations IB, H, B) Grantley for E5

Server One/Dual Socket Ivy Bridge - EP

Server Multi-Socket (>4) Haswell - EX

HPCMicroserver CPUs

Process Technology 45/40 32 28 22/16FF 14 10 7Note: Production dates are not always offered by companies and the information above often represents FBR's best estimate

SB

SB

SB

SB

Centerton Avoton - 64 bit

XMM 7160 LTE

Cherry Trail

Airmont

Moorefield 4 core 2.3GHz

Haswell (4/2 core)

Haswell (4/2 core)

Cherry TrailDenverton

Denverton

Sandybridge - EP

Sandybridge

Westmere - EX

Goldmont

Broxton

Shark Bay (Lynx Point Chipset)

Ivy Bridge

Ivy Bridge Haswell

Cherry TrailBay Trail - M

Silvermont

Skylake (4/2 core)

Haswell (2 core)

Haswell (2 core)

Skylake

Ivy Bridge (2 core) Broadwell (2 core) Skylake (2 core)

Medfield - 64 bit

SoFIA w/ LTE BB

Avoton - 64 bit

Clovertrail

Ivy Bridge (2 core) Broadwell (2 core) Skylake (2 core)

Bay Trail - D

Skylake (4/2 core)

Ivy Bridge (4/2 core) Broadwell (4/2 core)

Broadwell

Shark Bay (Lynx Point Chipset)

Haswell

Broadwell

Denverton follow on

Ivy Bridge (4/2 core)

Huron River (Cougar Point Chipset)

Haswell - EP

Haswell

Broadwell / Broadwell SoC

Centerton

Clovertrail Plus

Haswell

Haswell

Ivy Bridge - EX

Ivy Bridge (on Knights Corner)

Broadwell (4/2 core)

Chief River (Panther Point Chipset)

Broadwell

Skylake

Ivy Bridge

Denverton follow on

Broxton

Chief River (Panther Point Chipset)

Ivy Bridge Skylake Broadwell

Broxton

1Q12 2Q12 3Q12 4Q12 1Q13 2Q13 3Q13 4Q13 1Q14 2Q14 3Q14 4Q14 1Q15 2Q15

Marvell

AP + BB PXA1928, 64b, 4-A53, 1.5GHz, LTE, 5 mode

Baseband PXA1802 LTE thin-modem LTE Cat 4

Connectivity Avastar 88W8777 WiFi.n, BT 4.0, FM Avastar 88W8897 .ac, BT 4.0, FM, NFC

PXA988 dual A9, 3GPXA920 PXA1088LTE

PXA1801 LTE thin-modem, 5 modes

PXA1088 4-A7s 3G 5 mode


MWC 2014, MRVL: High-Profile Smartphone and Tablet Announcements


We rate Marvell Technology Group Ltd. Market Perform with a price target of $15. While we have recently warmed to the name, we remain on the sidelines until resolution of the CMU versus Marvell litigation. The lawsuit aside, we applaud Marvell’s success in the Chinese cellular platform market but do acknowledge this may be a more challenging market over time. Additionally, the company gained strong share in both HDDs and SSDs throughout 2013, but those share gains will be difficult to replicate in 2014. For the time being, we remain on the sidelines but may be compelled by a lower stock price and a reduction in CMU headline risk. Our $15 price target is based on a 14x target P/E multiple (pro forma CY 2014 EPS, including stock compensation, adding net cash).

NVIDIA Corporation (Market Perform – $17 Price Target) A few low-profile wins for Tegra 4i but disappointing overall. Remember, many of the struggles

for Tegra 4 were blamed on the decision to pull in Tegra 4i, which caused the company to miss the natural design period for Tegra 4. This time around, Tegra 4i was ready in time for the show but failed to generate significant design traction. The most high–profile device using Tegra 4i was probably the LG G2 mini, primarily for the South American market. Additionally, the company announced a win in the Wiko “WAX,” which will begin shipping in April. While these wins are a nice start, we do not view them as high-volume SKUs. A full list of recent Tegra design wins can be found below.

MWC 2014, NVDA: High-Profile Smartphone and Tablet Announcements


Tegra K1 may help drive further Tegra penetration through 2014. The 32-bit version of NVIDIA’s K1 will begin shipping soon, followed by the 64-bit version later in the year. The first K1 design win announcement, the ThinkVision 28 from Lenovo, was released just prior to the show. Below is a more comprehensive look at the Nvidia product cycles.

Device Name Date Part BB Combo

China Mobile 4G LTE Hotspot 1/3/2014 Armada PXA1802 2x2 MIMO ?

Yulong Coolpad MagView 8970L 1/6/2014 Armada PXA1802 2x2 MIMO ?

Blu-Castle connected Home 1/7/2014 G.hn 88LX3142

Skyworth-TVS 1/7/2014 Armada 1500 Plus (88DE3108)

SFR-TV's 1/14/2014 Armada 1500 Plus (88DE3108)

Sk Broadband-TV Smart 1/15/2014 Armada 1500 Plus (88DE3108)

Galaxy Win Pro Smartphone 2/18/2014 Armada PXA1088

Part Device Name Announce Date Ship Date

Tegra 4i LG G2 Mini 2/25/2014 Q1 2014

Tegra 4i Wiko Wax 2/25/2014 Q1 2014

Tegra K1 Lenovo ThinkVision 28 1/13/2014 Q2 2014

Tegra 4 HP SlateBook x2 9/19/2013 Q4 2013

Tegra 4 Coolpad Grand 9/23/2014 Q4 2013

Tegra 4 Cintiq Companion 8/20/2014 Q4 2013

Tegra 4 Xiaomi Phone 3 9/5/2013 Q4 2013

Tegra 4 Microsoft Surface Pro 2 9/23/2013 Q4 2013


MWC 2014, FBR’s Estimate of NVDA’s Roadmaps


Overall, we rate shares of NVDA Market Perform, with a price target of $17. Regarding NVDA shares, bears say NVIDIA faces long-term secular PC headwinds, lower GPU attach rates, and a lack of mobile traction. Bulls will say NVIDIA is making great strides in diversifying outside of PCs and into Quadro, Tesla, GRID, Shield, and Tegra. We are neutral on NVDA and rate the stock a Market Perform, given ongoing demand weakness in notebook GPU sales, a gap in Tegra uptake, and a lack of meaningful catalysts for now. That said, we think NVDA is one of the most attractive PC chip stocks in the sector today, given mobile and datacenter growth opportunities, the profitable model, and almost $5.00 of net cash per share.

QUALCOMM Corporation (Outperform – $79 Price Target) Announces Snapdragon improvements at MWC; up the ante in baseband. Qualcomm

announced the Snapdragon 801 sporting significantly faster performance. That evening, it was announced that the 801 would power certain models of the GS5. Additionally, Qualcomm added the Snapdragon 610 and 615 chipsets for high-end mobile devices, with both having third-generation LTE modems, supporting LTE-Broadcast and LTE Dual SIM Dual Active (DSDA). The Snapdragon 610 and 615 chipsets enable OEMs to ship a single 5-mode global LTE SKU across all major bands. They also unveiled the 20nm Gobi 9x30, an LTE platform that will bring cellular capabilities to the auto space. The company is also integrating dual-stream 802.11ac Wi-Fi and Bluetooth 4.1 chipset into the platform. Qualcomm clearly had the most robust list of design wins announced at the show, which can be found below.

MWC 2014, QCOM: High-Profile Smartphone and Tablet Announcements


Below is a more detailed look at the Qualcomm roadmap from 2012–2015.

MWC 2014, FBR’s Estimate of QCOM’s Roadmaps


We rate QCOM an Outperform with a price target of $79. Stepping back, QUALCOMM remains the gold standard cellular technology provider, with 4G LTE and Apple/Samsung catalysts ahead of it.

1Q12 2Q12 3Q12 4Q12 1Q13 2Q13 3Q13 4Q13 1Q14 2Q14 3Q14 4Q14 1Q15 2Q15 3Q15 4Q15

NVIDIA - Mobile

Mobile Apps Processor

Mobile AP + Baseband

Baseband

NVIDIA - GPU

Architecture

Tegra K1 64b 4/A15

Tegra K1i (?) LTE

Tegra K1 32b 4/A15Tegra 3 quad A9s

i410/i450 discrete LTE SRD

Tegra 4 quad ARM A15 Tegra 5 (?) 64b Maxwell GPU (est)

i500 discrete SRD - LTE Cat 3-4

Tegra 4i quad A9s w/LTE up to 2.3 GHz

Kepler Maxwell (transitions into 16nm FinFET)

Device Name Date Part AP

Sony Xperia Z2 Tablet 2/24/2014 Snapdragon 801 MSM8974AB

Sony Xperia Z2 2/24/2014 Snapdragon 801 MSM8974AB

Samsung Galaxy S5 2/24/2014 Snapdragon 801 MSM8974AC

HTC One (2) 2/24/2014 Snapdragon 800 MSM8974AB ?

Sony Xperia M2 2/24/2014 Snapdragon 400 MSM8926

LG (G2 Mini) 2/24/2014 Snapdragon 400 MSM8226

Google (Nexus 6) 2/24/2014 Snapdragon 810 ?

BlackBerry (Z30) 2/24/2014 Snapdragon S4 MSM8960

Nokia Lumia (1520) 12/6/2014 Snapdragon 800 MSM8974

1Q12 2Q12 3Q12 4Q12 1Q13 2Q13 3Q13 4Q13 1Q14 2Q14 3Q14 4Q14 1Q15 2Q15 3Q15

Qualcomm Snapdragon Series 800, 600 (LTE)

Mid-range

Connectivity

Front End

Auto

WCN3680 WiFi.ac, BT 4.0, FM

Snapdragon 602A

801, 805 Quad Krait 2.7 GHz Series 801, 610, 615 (8 core, 64 bit, LTE)

Series 410, (64 bit, LTE)

RF360

Series 400, 200 (LTE)

S4 - 8960


The stock is one of our preferred, mega-cap semiconductor longs, with favorable risk/reward dynamics. The long-term QUALCOMM story seems intact, with the firm a solid play on global growth in 3G and 4G technologies and with opportunities to ramp new silicon content, ride the 4G LTE wave, grow TD-SCDMA opportunities in China, and reap annual QTL device shipments as they approach 2.0 billion units.

Broadcom Corporation (Outperform – $34 Price Target) Broadcom ups the stakes with new 2x2 Combo MIMO chip. Last week, Broadcom introduced

the industry's first 5G Wi-Fi (802.11ac) 2x2 multiple input multiple output (MIMO) SoC for smartphones. While the 5G aspect has been announced, the 2x2 MIMO is the next step in the evolution of its combo chip market, thus creating distance from QUALCOMM. MIMO is the use of multiple antennas at both the transmitter/receiver to improve data speed performance without increased power needs. MIMO works by spreading the data flow through multiple ports, allowing for more rapid transfers without exponential increases in power usage. We believe that this chip will be used in the GS5, and it will have a chance to power the iPhone in 2H14 (although Apple has favored prior-generation parts from Broadcom in the past). Additionally, Broadcom announced that Datang Mobile would be using the dual-mode TD-SCDMA/TD-LTE SoC for its enterprise and residential gateway solutions.

MWC 2014, BRCM: High-Profile Smartphone and Tablet Announcements


Below is a more comprehensive display of Broadcom’s roadmaps. We continue to believe that LTE will be the biggest driver of incremental mobile and wireless revenue for Broadcom in 2014 and 2015.

MWC 2014, FBR’s Estimate of BRCM’s Roadmaps


We rate BRCM Outperform, with a $34 price target. We believe BRCM can move into the mid $30s over time, as integrated cellular platforms coalesce around the company and others (e.g., QCOM). Additionally, we now believe we are on the precipice of a massive movement toward SDN that should drive accelerated networking growth rates for arguably the proprietor of the world's best switch-silicon portfolio. That said, we acknowledge the near-term headwinds but are willing to take a longer-term perspective, as we believe BRCM is a long-term core holding in the sector, owing to the company’s competitive advantages (speed to integration), product cycles, and strategic importance with customers.

ATML (Market Perform – $9.25 Price Target) Atmel had few announcements at MWC this year. The company did announce a partnership

with Movea’s Smartmotion platform to better address the mobile market. It also announced an expansion of its ARM Cortex MCU portfolio to help deliver power efficiency and connectivity, particularly for smaller devices.

We rate ATML Market Perform, with a price target of $9.25. Overall, we think the Street underestimates all, or a portion, of the following: (1) new sensor hub opportunities; (2) a strong

Device Name Date Part Combo

Datang Mobile 2/27/2014 5G WiFi 2x2 MIMO BCM4354

Samsung Galaxy 6 2/24/2014 5G WiFi 2x2 MIMO BCM4354

Apple iPhone 6 ? 5G WiFi 2x2 MIMO BCM4354

Broadcom 1Q12 2Q12 3Q12 4Q12 1Q13 2Q13 3Q13 4Q13 1Q14 2Q14 3Q14 4Q14 1Q15 2Q15 3Q15 4Q15

AP + BB BCM28155 2-A9 3G BCM28550 4-A9 3G M320 LTE SoC (dual core) M340 LTE SoC (quad core, FD/TD Cat 4)

Baseband

Connectivity

Switch Silicon

BCM4334 WiFi.ac, BT 4.0, FM

LTE Advanced Thin Modem Cat 6 (est)

BCM21553 3G Single Core

BCM4354 WiFi.ac 2x2 MIMO, BT 4.1, FMBCM4334 WiFi.n, BT, FM

Trident II


http://www.broadcom.com/products/features/5gwifi.php

cyclical recovery in the core MCU business, also benefiting the ASIC and RF units; (3) Win8 touch PC wins; and (4) growing XSense demand in 2014. Additionally, the unwinding of the take-or-pay Foundry agreement and higher utilizations should drive gross margin expansion of approximately 400 bps over the next six quarters, benefiting the bottom line. That said, valuation keeps us sidelined, as shares of ATML trade at 29x our 2014 estimate and 19x our 2015 (EV/pro forma EPS including SBC expense).

Sector Stance: Chip Stocks Partially Discount Five to Six Innings of the Cyclical Recovery For semiconductor stocks overall, the SOX was up more than 30% in 2013, and we believe the early stages (call it the first six innings) of the recovery are already reflected in share prices. While supply chain inventories are lean, they are not as lean as their 2009 trough; therefore, this “cycle” may lack the benefits of aggressive inventory restocking, increasing the likelihood of a stop-and-start expansion for chip companies. Despite the near-term choppiness, we believe the recovery looks rather typical, as it appears we have entered the expansion stage of the cyclical recovery. Within this context, our favorite longs are ARMH, AMD, AMCC, BRCM, QCOM, INTC, and ONNN.

Valuation Summary for FBR's Semiconductor Devices Coverage Universe

Note: % return to target is potential.


Semiconductor Industry Risks Risks to this sector include (1) demand or macroeconomic risks that could weaken end-user consumption of chips; (2) cyclical risks, such as excess inventory or capacity; (3) chip price declines; (4) risks of excess competition driving pricing and margin weakness; and (5) technology design or implementation risks.

Price Return Market Current Target Current Target Current Target Current Target Current Target Current Target

3-Mar-14 Ticker Rating Price Target to Target Cap ($m)

2014

EV/S

2014

EV/S

2015

EV/S

2015

EV/S

2014

P/E

2014

P/E

2015

P/E

2015

P/E

2014 EV/

EBITDA

2014 EV/

EBITDA

2015 EV/

EBITDA

2015 EV/

EBITDA

ADVANCED MICRO DEVICES AMD OP $3.67 $5.50 50% 2,844 0.7 0.9 0.6 0.9 12.8 17.6 11.2 15.3 7.4 10.2 7.3 10.0

APPLIED MICRO AMCC OP $11.32 $14 24% 892 3.6 4.5 2.8 3.4 NM NM 32.4 40.1 39.1 48.3 12.5 15.5

ARM HOLDINGS ARMH OP $49.26 $59 20% 23,448 15.8 19.0 13.9 16.7 64.3 77.2 42.1 50.5 29.6 35.5 24.9 29.8

ATMEL CORP ATML MP $8.03 $9.25 15% 3,453 2.1 2.4 2.0 2.3 28.1 33.0 18.1 21.2 10.8 12.7 8.3 9.8

BROADCOM CORP BRCM OP $29.76 $34 14% 18,197 1.6 1.9 1.5 1.8 11.8 14.1 10.2 12.1 7.2 8.5 6.4 7.7

FAIRCHILD SEMICONDUCTOR FCS MP $13.89 $14 1% 1,707 1.0 1.0 0.9 0.9 18.7 18.8 10.4 10.5 4.2 4.2 3.5 3.5

INTERNATIONAL RECTIFIER CORPIRF MP $26.88 $26 -3% 1,956 1.2 1.2 1.1 1.0 15.5 14.7 9.7 9.2 7.1 6.8 5.5 5.2

INTEL CORP INTC OP $24.50 $27 10% 123,652 2.1 2.4 2.1 2.3 12.0 13.4 11.2 12.4 5.6 6.2 5.2 5.8

MARVELL TECHNOLOGY GROUP MRVL MP $15.43 $15 -3% 8,029 1.5 1.4 1.4 1.4 14.8 14.2 13.0 12.5 8.5 8.1 8.0 7.7

NVIDIA CORP NVDA MP $18.29 $17 -7% 10,560 1.7 1.6 1.7 1.5 17.3 15.6 16.0 14.4 9.9 8.9 9.6 8.7

ON SEMICONDUCTOR CORP ONNN OP $9.35 $11 18% 4,144 1.4 1.6 1.3 1.5 11.3 13.3 9.2 10.9 6.9 8.2 6.2 7.3

QUALCOMM INC QCOM OP $73.63 $79 7% 130,633 3.4 3.8 3.1 3.4 12.5 13.8 11.0 12.1 8.3 9.1 7.3 8.0

TEXAS INSTRUMENTS INC TXN MP $44.58 $40 -10% 48,414 3.9 3.5 3.7 3.4 21.2 19.0 18.0 16.2 11.1 10.0 10.4 9.3

Median (ex-ARMH) $15.43 $15 7% 6,118 1.7 1.8 1.6 1.6 14.8 14.5 11.2 12.4 8.3 8.9 7.3 7.9

Average (ex-ARMH) $22.34 $23 9% 27,738 2.0 2.2 1.7 2.0 16.0 17.1 14.2 15.6 10.5 11.7 7.5 8.2

All P/E multiples include stock compensation expense and back out net cash and related interest income (or the opposite for debt)


Important InformationFBR is the global brand for FBR & Co. and its subsidiaries.

This report has been prepared by FBR Capital Markets & Co. (FBRC), a subsidiary of FBR & Co.

FBRC is a broker-dealer registered with the SEC and member of FINRA, the NASDAQ Stock Market and the Securities InvestorProtection Corporation (SIPC). The address for FBRC is 1001 Nineteenth Street North Suite 1100, Arlington, VA 22209.

All references to FBR Capital Markets & Co. (FBRC) mean FBR & Co. and its subsidiaries including FBRC.

Company-Specific DisclosuresFBR acts as a market maker or liquidity provider for the company's securities: Applied Micro Circuits Corp, Inc., AdvancedMicro Devices, ARM Holdings plc, Atmel Corporation, BlackBerry, Broadcom Corporation, Ciena Corporation, CiscoSystems, Inc., DISH Network Corporation, Fairchild Semiconductor International, Infinera Corporation, Intel Corporation,Juniper Networks, Inc., Marvell Technology Group Ltd., NVIDIA Corporation, ON Semiconductor Corporation, QUALCOMMIncorporated, Sprint Corporation, AT&T Inc. and T-Mobile US, Inc.For up-to-date company disclosures including price charts, please click on the following link or paste URL in a web browser: www.fbr.com/disclosures.aspx

General DisclosuresInformation about the Research Analyst Responsible for this report:

The primary analyst(s) covering the issuer(s), David Dixon, Christopher Rolland and Scott Thompson, certifies (certify) thatthe views expressed herein accurately reflect the analyst's personal views as to the subject securities and issuers andfurther certifies that no part of such analyst's compensation was, is, or will be, directly or indirectly, related to the specificrecommendations or views expressed by the analyst in the report. The analyst(s) responsible for this research report hasreceived and is eligible to receive compensation, including bonus compensation, based on FBR’s overall operating revenues,including revenues generated by its investment banking activities.

Information about FBR's Conflicts Management Policy:

Our Research conflicts management policy is available at: http://www.fbr.com/conflictsmanagementpolicy.asp.

Information about investment banking:

In the normal course of its business, FBR seeks to perform investment banking and other services for various companiesand to receive compensation in connection with such services. As such, investors should assume that FBR intends to seekinvestment banking or other business relationships with the companies.

Information about our recommendations, holdings and investment decisions:

The information and rating included in this report represent the long-term view as described more fully below. The analystmay have different views regarding short-term trading strategies with respect to the stocks covered by the rating, optionson such stocks, and/or other securities or financial instruments issued by the company. Our brokers and analysts may makerecommendations to their clients, and our affiliates may make investment decisions that are contrary to the recommendationscontained in this research report. Such recommendations or investment decisions are based on the particular investmentstrategies, risk tolerances, and other investment factors of that particular client or affiliate. From time to time, FBR, its affiliatedentities, and their respective directors, officers, employees, or members of their immediate families may have a long or shortposition in the securities or other financial instruments mentioned in this report.

We provide to certain customers on request specialized research products or services that focus on covered stocks from aparticular perspective. These products or services include, but are not limited to, compilations, reviews, and analysis that mayuse different research methodologies or focus on the prospects for individual stocks as compared to other covered stocksor over differing time horizons or under assumed market events or conditions. Readers should be aware that we may issueinvestment research on the subject companies from a technical perspective and/or include in this report discussions aboutoptions on stocks covered in this report and/or other securities or financial instruments issued by the company. These analysesare different from fundamental analysis, and the conclusions reached may differ. Technical research and the discussionsconcerning options and other securities and financial instruments issued by the company do not represent a rating or coverageof any discussed issuer(s). The disclosures concerning distribution of ratings and price charts refer to fundamental research and


www.fbr.com/disclosures.aspx

www.fbr.com/disclosures.aspx

http://www.fbr.com/conflictsmanagementpolicy.asp

do not include reference to technical recommendations or discussions concerning options and other securities and financialinstruments issued by the company.

Important Information Concerning Options Transactions:

This discussion is directed to experienced professional investors with a high degree of sophistication and risk tolerance.

Options transactions are not suitable for all investors. This brief statement does not address all of the risks or other significantaspects of entering into any particular transaction. Tax implications are an important consideration for options transactions.Prior to undertaking any trade you should discuss with your preferred tax, ERISA, legal, accounting, regulatory, or other advisorhow such particular trade may affect you.

Opinion with respect to options is distinct from fundamental research analysis. Opinion is current as of the time of publication,and there should be no expectation that it will be updated, supplemented, or reviewed as information changes. We make nocommitment to continue to follow any ideas or information contained in this section. Analysis does not consider the cost ofcommissions. Research personnel may consult Options Sales and Trading personnel when preparing commentary concerningoptions. Supporting documentation is available upon request.

Please ensure that you have read and understood the current options risk disclosure document before entering intoany options transactions. The options risk disclosure document can be accessed at the following Web address: http://optionsclearing.com/about/publications/character-risks.jsp. If this link is inaccessible, please contact your representative.

Risks

Some options strategies may be complex, high risk, and speculative. There are potentially unlimited combinations of hedgedand unhedged options strategies that expose investors to varying degrees of risk. Generally, buyers establishing long optionspositions risk the loss of the entire premium paid for the position, while sellers establishing short options positions haveunlimited risk of loss. There are a number of commonly recognized options strategies, that expose investors to varying degreesof risk, some of which are summarized below:

Buying Calls or Puts--Investors may lose the entire premium paid.

Selling Covered Calls--Selling calls on long stock position. Risk is that the stock will be called away at strike, limiting investorprofit to strike plus premium received.

Selling Uncovered Calls--Unlimited risk that investors may experience losses much greater than premium received.

Selling Uncovered Puts--Significant risk that investors will experience losses much greater than premium income received.

Buying Vertical Spreads (Calls--long call and short call with higher strike; Puts--long put and short put with lower strike) Sameexpiration month for both options. Investors may lose the entire premium paid.

Buying Calendar Spreads (different expiration months with short expiration earlier than long). Investors may lose the entirepremium paid.

Selling Call or Put Vertical Spreads (Calls--short call and long call with higher strike; Puts--short put and long put with a lowerstrike, same expiration month for both options.) Investors risk the loss of the difference between the strike prices, reducedby the premium received.

Buying Straddle--Buying a put and a call with the same underlying strike and expiration. Investors risk loss of the entirepremium paid.

Selling Straddle--Sale of call and put with the same underlying strike and expiration.) Unlimited risk that investors willexperience losses much greater than the premium income received.

Buying Strangle--Long call and long put, both out of the money, with the same expiration and underlying security. Investorsmay lose the entire premium paid.

Selling Strangle--Short call and put, both out of the money, with the same expiration and underlying security. Unlimited riskof loss in excess premium collected.

Important Information about Convertible & Other Fixed-Income Securities and Financial Instruments:

This discussion is directed to experienced professional investors with a high degree of sophistication and risk tolerance.

Opinion with respect to convertible, other fixed-income securities and other financial instruments is distinct from fundamentalresearch analysis. Opinion is current as of the time of publication, and there should be no expectation that it will be updated,supplemented, or reviewed as information changes. We make no commitment to continue to follow any ideas or informationcontained in this section.


http://optionsclearing.com/about/publications/character-risks.jsp

http://optionsclearing.com/about/publications/character-risks.jsp

Research analysts may consult Credit Sales and Trading personnel when preparing commentary on convertible and fixed-income securities and other financial instruments. FBR may be a market maker in the company’s convertible or fixed-incomesecurities. FBR Capital Markets LT, Inc. may be a market maker in financial instruments that are not securities.

Securities and financial instruments discussed may be unrated or rated below investment grade, may be consideredspeculative, and should only be considered by accounts qualified to invest in such securities.

Securities and financial instruments discussed may not be registered or exempt from registration in all jurisdictions.Nonregistered securities discussed may be subject to a variety of unique risk considerations, including those related toliquidity, price volatility, and lack of widely distributed information.

Rule 144A securities are sold only to persons who are Qualified Institutional Buyers within the meaning of Rule 144A, underthe Securities Act of 1933, as amended.

Information about our rating system:

FBR instituted the following three-tiered rating system on October 11, 2002, for securities it covers:

■ Outperform (OP) — FBR expects that the subject company will outperform its peers over the next 12 months. Werecommend that investors buy the securities at the current valuation.

■ Market Perform (MP) — FBR expects that the subject company’s stock price will be in a trading range neitheroutperforming nor underperforming its peers over the next 12 months.

■ Underperform (UP) — FBR expects that the subject company will underperform its peers over the next 12 months. Werecommend that investors reduce their positions until the valuation or fundamentals become more compelling.

A description of the five-tiered rating system used prior to October 11, 2002, can be found at http://www.fbr.com/disclosurespre10702.aspx.

Rating. FBR Research Distribution 1 FBR Banking Services in the past 12 months1

BUY [Outperform] 49.07% 14.35% HOLD [Market Perform] 48.24% 4.72% SELL [Underperform] 2.69% 0.00% (1) As of midnight on the business day immediately prior to the date of this publication.

General Information about FBR Research:

Additional information on the securities mentioned in this report is available upon request. This report is based on dataobtained from sources we believe to be reliable but is not guaranteed as to accuracy and does not purport to be complete.Opinion is as of the date of the report unless labeled otherwise and is subject to change without notice. Updates may beprovided based on developments and events and as otherwise appropriate. Updates may be restricted based on regulatoryrequirements or other considerations. Consequently, there should be no assumption that updates will be made. FBR and itsaffiliates disclaim any warranty of any kind, whether express or implied, as to any matter whatsoever relating to this researchreport and any analysis, discussion or trade ideas contained herein. This research report is provided on an "as is" basis for useat your own risk, and neither FBR nor its affiliates are liable for any damages or injury resulting from use of this information.This report should not be construed as advice designed to meet the particular investment needs of any investor or as an offeror solicitation to buy or sell the securities or financial instruments mentioned herein, and any opinions expressed herein aresubject to change. Some or all of the securities and financial instruments discussed in this report may be speculative, highrisk, and unsuitable or inappropriate for many investors. Neither FBR nor any of its affiliates make any representation as tothe suitability or appropriateness of these securities or financial instruments for individual investors. Investors must maketheir own determination, either alone or in consultation with their own advisors, as to the suitability or appropriateness ofsuch investments based upon factors including their investment objectives, financial position, liquidity needs, tax status, andlevel of risk tolerance. These securities and financial instruments may be sold to or purchased from customers or others byFBR acting as principal or agent.

Securities and financial instruments issued by foreign companies and/or issued overseas may involve certain risks, includingdifferences in accounting, reporting, and registration, as well as foreign currency, economic, and political risks.


http://www.fbr.com/disclosurespre10702.aspx

http://www.fbr.com/disclosurespre10702.aspx

This report and the securities and financial instruments discussed herein may not be eligible for distribution or sale in alljurisdictions and/or to all types of investors. This report is provided for information purposes only and does not represent anoffer or solicitation in any jurisdiction where such offer would be prohibited.

Commentary regarding the future direction of financial markets is illustrative and is not intended to predict actual results,which may differ substantially from the opinions expressed herein. If any hyperlink is inaccessible, call 800.846.5050 and askfor Editorial.

FBR utilizes a tiered approach to service its clients. The services provided by FBR’s research analysts to clients vary, basedupon a variety of factors including, but not limited to, client preferences and the extent of a client’s total relationship withthe Firm. FBR does not provide any of the Firm’s clients with access to unpublished research opinion. FBR provides clientsacross all tiers equal access to research reports.

Information for Clients of FBRC:

This publication has been approved by FBR Capital Markets & Co. (FBRC), which accepts responsibility for its contents and itsdistribution to our clients. Any FBRC client who receives this research and wishes to effect a transaction in the securities orfinancial instruments discussed should contact and place orders with an FBRC Sales representative or a representative of FBRCapital Markets LT, Inc. for financial instruments that are not securities.

Copyright 2014 FBR Capital Markets & Co.


technology, media & telecom key points

Documents