Wedbush Morgan analyst Craig Berger recently sent a note to clients on Pixelworks Inc. (PXLW), the Oregon-based designer, developer and marketer of semiconductor devices and software for advanced multimedia products. Past Wedbush notes have been bullish on the company but after another quarter of dismal earnings (or lack thereof), Wedbush has decided to lower their rating for shares of PXLW from BUY to HOLD. The note follows:
• We were wrong to have been constructive on PXLW over the past year and are downgrading shares to HOLD given our view that the firm will not recover as much lost TV business in 2007 as we were previously hoping for. Design win traction with meaningful customers appears limited in its TV business, and we do not believe holding out hope that PXLW is acquired is reason enough to recommend the stock. We do still believe the firm has a compelling technology position and product roadmap, though we favor Trident Microsystems (TRID) in this space given their strong execution and earnings power. For those still in PXLW, there seems to be potential acquisition appeal from Broadcom (BRCM), NVIDIA (NVDA), TI (TXN), STMicro (STM), or LSI Logic (LSI).
• Pixelworks reported Q2 revenues below our own and consensus estimates, though EPS was a tick better than consensus driven by OpEx spending restraint. Pixelworks recorded Q2 revenues of $30.9 million (-16% QoQ, -25% YoY) and pro forma EPS of ($0.16), worse than consensus revenues but one penny ahead of consensus estimates as management limited operating expense spending. Pro forma gross margins of 40.0% decreased -80bps QoQ and -60bps YoY. Operating margin of –26.1% was an all time low since becoming a publicly traded company.
• Management guided Q3 revenues in the range of $33 million to $35 million, worse than consensus estimates of $35 million; EPS guidance was generally inline with consensus estimates on restrained OpEx spending. Pro forma earnings is expected to be in the range of ($0.09) to ($0.14) per share, in line with consensus estimates of ($0.11). Management is reporting a ‘run-rate’ book to bill of 1.3x for its TV business and 1.1x for its projector business, indicating some sequential growth in these segments is possible.
• Reducing 2006 EPS estimate from ($0.49) to ($0.53), 2007 EPS estimate from ($0.37) to ($0.45), and price target from $5.50 to $2.75 on lower sales forecasts. Our $2.75 price target is based upon a 1.0x multiple of 2006 enterprise value to sales, an appropriate trough valuation, we believe, for chip firms losing money. Historically, Pixelworks has traded within a range of 1.5x to 7.0x enterprise value to sales, with a median multiple of 2.9x. Our target EV/S multiple is below the company’s historic median EV/S multiple because of our concerns about execution and market share losses, in order to discount the very high multiples the company enjoyed during the early part of its history, and to offer a more tempered view of valuations relative to our expectations for increased competitive pressures in 2006.
• Risks to attainment of our share price target include: weaker than expected projector, ATV, or IPTV chip shipments, late delivery of production quantities of newly announced chips, the inability to garner significant customer design wins following chip production slips in mid-2005, greater than expected chip price declines, new competitive threats, and other execution issues.
Background – Pixelworks Launched Slate of New, Competitive Products in Jan’06
Pixelworks’ Chip Offering Seems to Be Competitive – Pixelworks product announcements at CES (1/5/06) ended a four-month stretch since its last new product announcement. These solutions nearly refresh Pixelworks’ entire product offering and give us significantly improved confidence that the firm is designing compelling products, can win material customer designs in 2006, and will generally ship its new silicon on-time. Versus its competition, we now believe Pixelworks’ product roadmap is competitive with that of Genesis Micro and Trident Micro in product quality, exceeds that of Genesis Micro (GNSS) and Trident Micro in depth of product offering within the TV signal path, and trails that of Genesis Micro and Trident Micro in chip integration, specifically in the integrated MPEG analog image processor space, a still-nascent market in 2006. We now follow with an overview of Pixelworks’ important new products announced at CES, in order from greatest potential revenue impact to least potential revenue impact:
• ‘Pearl’ Image Processor - Pearl, the firm’s latest generation image processor, integrates an HDMI receiver, a 12-bit video decoder, a 10-bit analog to digital converter, a scaler, a deinterlacer, the on-screen display controller, and other image enhancement circuitry. We believe this product is a very competitive analog image processor and is on par with what Trident Micro and Genesis Micro are offering for 2006. Like these firms’ 2006 offerings, this chip contains an integrated HDMI receiver, circuitry that would otherwise be a $2-3 discrete chip. This is the latest block of integration into the analog image processor market for 2006.
• ‘Opal2’ Image Processor – Pixelworks announced an enhanced version of its highly integrated Opal processor, which finally went into production in Q3’05. Opal2 includes an improved analog video decoder and specific improvements aimed at the European TV market. This chip also includes weak signal enhancement, a plus for TV manufacturers.
• MPEG Decoder PWM2020 and PWM2030 – Pixelworks’ new MPEG decoder chips are the PWM2020 and the PWM2030, targeted at Europe and North America. The PWM2020 is a low-cost, single channel MPEG decoder for standard definition digital TV. This chip uses only 32MB of memory versus 64MB of memory required in most solutions, thus saving BOM costs. The PWM2030 is a higher-end, dual channel MPEG decoder chip. Dell is a current Pixelworks customer that uses the PWM2000, the firm’s first-generation MPEG decoder solution. The indigo software solution paired with the PWM product family is a linux-based solution currently used in production TVs, and contains approximately five million lines of software code.
• New Projector Chips (not-named) – Pixelworks discussed a couple of new projector chips at its product demonstration. One chip is a low-cost solution containing significant integration including system memory, driving BOM cost savings. The second chip is a higher-end projector chip including extreme keystone correction and other features.
• PW50 and PW60 Advanced CRT Chips – For the first time in a couple of years Pixelworks has refreshed its lineup of advanced CRT chips for regular tube TVs. These chips are essentially an Opal image processing chip that allows tube TVs to display digital content with higher quality. These chips contain an ADC, a video decoder, a scaler, and other image enhancement circuitry.
• ‘Songbird’ Audio/Video Demodulator – Pixelworks launched a new family of demodulator chips at CES, codenamed Songbird. This chip contains an audio demodulator, a video demodulator and an audio baseband processor for output to a speaker system. Competitors Trident Micro and Genesis Micro are both working on this technology but to our knowledge do not yet have working silicon solutions. This product has been in development for more than a year and represents a major thrust forward in Pixelworks’ long-stated effort to move upstream in the TV signal processing chain towards the tuner. Songbird actually allows customers to buy less expensive single-stage tuners by integrating the intermediate frequency demodulation functionality within more expensive multi-stage tuners. Demodulators separate the audio signal from the video signal so that both signals can be processed further in the signal processing chain.
• ‘DreamStream’ IPTV Reference Platform – Pixelworks launched a software reference design platform to be used with the firm’s BSP-15 Babelfish IPTV chip solution, and represents another step forward for in the firm’s pursuit of garnering IPTV set top box revenues in this nascent market (small $70 million TAM in 2006). This software platform includes functionality for audio and video decoding, wired or wireless network connectivity interface, internet browsing, and digital rights management functionality.
• ‘Peanut’ and ‘Cashew’ Timing Controller Chips – Pixelworks has been working with development partner Samsung to develop programmable timing controllers, a new type of timing controller that helps synchronize the pixels on an LCD panel itself. Pixelworks said they shipped about 100,000 units of this product in Q2’06 to its lead customer Samsung. This product sports a mid-single digit average selling price and could potentially go into many LCD TVs in the future if the chip’s price and performance are compelling versus competing non-programmable solutions available in the market today.
Background: Some Hope Remains for (Very) Patient Investors
Some Hope Remains for patient investors with new product launches, restructuring program, and executive changes indicating that management is not operating with a ‘business-as-usual’ mentality:
1. Pixelworks’ product roadmap is competitive heading into 2006 as chip integration with Opal2 and Pearl now comparable with products from Trident (SVP-EX/LX) and Genesis (Cortez/Hudson). Also, we believe the firm has the most robust software stack and reference designs of the three competitors, thus enabling the ‘lowest-touch’ solution launch for its customers. Finally, Pixelworks’ HDTV platform is in production and ready to ramp with other customers.
2. Pixelworks replaced its previous head of engineering with a former Toshiba executive in order to speed time to market for new product development. Other structural changes have been implemented within the engineering team to drive better efficiencies between working groups.
3. Pixelworks hired two veteran sales executives to head the Korean sales team and the Japanese sales team, with a particular focus on penetrating key OEM accounts. One of these executives came from a top image processing competitor.
Investment Thesis and Valuation
Pixelworks Investment Thesis: We rate Pixelworks a HOLD with a $2.75 price target. We believe Pixelworks will not recover as much lost TV business in 2007 as we were previously hoping for. Design win traction with meaningful customers appears limited in its TV business, and we do not believe holding out hope that PXLW is acquired is reason enough to recommend the stock. We continue to believe that with a low enterprise value of around $125 million, PXLW could be an acquisition target for Broadcom, Texas Instruments, Intel (INTC), Zoran (ZRAN), or LSI Logic. In addition to our concerns over market share losses and product ramp missteps, we also have concerns about the firm’s high operating expense spending compared to revenues and see the need for tight spending control or even spending reductions.
Pixelworks Valuation Methodology: Our $2.75 price target is based upon a 1.0x multiple of 2006 enterprise value to sales, an appropriate trough valuation for chip firms losing money. Historically, Pixelworks has traded within a range of 1.5x to 7.0x enterprise value to sales, with a median multiple of 2.9x. Our target EV/S multiple is below the company’s historic median EV/S multiple because of our concerns about execution and market share losses, in order to discount the very high multiples the company enjoyed during the early part of its history, and to offer a more tempered view of valuations relative to our expectations for increased competitive pressures in 2006.
• ATV Demand and Design Win Risk – Advanced TV chip revenues will account for nearly 60% of total revenues in 2006. If end demand or design wins fail to materialize as expected then Pixelworks’ financial performance will be negatively impacted. Conversely, if the company sells more chips than we expect then our estimates and target price on the stock could be too conservative.
• LCD and Plasma Panel Supply Risk - Pixelworks’ expected revenue ramp is very dependent upon availability of LCD and plasma panels. If either LCD or plasma panel production growth should slow or stall, then Pixelworks’ financial performance would be negatively impacted.
• Price Declines — Given the historical precedent of same-part price declines in the LCD monitor market, we expect aggressive price declines in the company’s ATV image processor chips as volumes start to ramp more significantly in 2H’06. Steep ASP erosion would negatively affect Pixelworks’ financial performance. ASP declines in excess of –15-18% per year will have negative ramifications on Pixelworks’ earnings power.
• Competition — Significant competition is already competing in, or will compete, in the ATV chip market. Chinese or Taiwanese cost leaders may be able to capture significant share. Alternatively, well-funded companies like Broadcom, ATI Technologies (ATYT), and Intel are entering this market and may exercise their substantial resources to take market share from Pixelworks and other incumbent suppliers.
• Integration Risk – Pixelworks has not historically been a leader in chip integration. If the company fails to integrate leading features into its chips, than its business would suffer. Additionally, if the company cannot design and then subsequently integrate a high quality MPEG decoder chip for ATVs then the company’s financial performance could lag in 2007 and beyond.
PXLW 1-yr chart: