MIPS Technologies MIPS
August 08, 2008 - 4:04pm EST by
repetek827
2008 2009
Price: 3.62 EPS
Shares Out. (in M): 0 P/E
Market Cap (in $M): 161 P/FCF
Net Debt (in $M): 0 EBIT 0 0
TEV (in $M): 0 TEV/EBIT

Sign up for free guest access to view investment idea with a 45 days delay.

Description


    MIPS  is the world's second largest  semiconductor design IP company, second only to ARM Holdings (ARMH) in  microprocessors and number one in analog and mixed-signal IP. The net present value of MIPS's after tax royalty stream, which is growing and  adding new licensees every quarter, is worth two thirds of the present TEV ($127mm vs  $170mm). On top of this MIPS has a newly acquired  analog licensing biz (Chipidea)  which they acquired for $147mm,  running at $11mm/ quarter revenue run rate and growing at 25%yoy. Chipidea is a gem of an asset and truly  unique in the semiconductor IP world. As MIPS works through the acquisition and returns to its historical 20% operating margins, investors get  the company  at one of its lowest valuations ever with excellent long term growth prospects.

THE INDUSTRY  
     The processor/microcontroller market can roughly be divided into three segments. At the high end in terms of  processing power is Intel and its direct competitors These companies sell into the personal computer, workstations, servers, mainframes, and mini computer markets. At the bottom of the market in terms of processing power,  you have the microcontroller companies like Microchip that are basically putting their devices into  everything you can imagine.  This market primarily consists of 4-bit, 8-bit, 16-bit, and 32bit microcontrollers embedded into  low-cost consumer products such as home appliances, fax machines, printers, and automobiles.
    Somewhere in the middle between the high end and the low end you find MIPS and ARMH.  From the standpoint of their IP, ARM and MIPS are  very similar. Both have been in the business for decades, have a roster of who's who among their licensees, and support their architectures with the highest level of software and 3rd party IP. The good news for MIPS is that  there is minimal overlap in their end markets.  ARM has carved out a dominant position in wireless telecommunications and mobile phones.  Their architecture is the de-facto standard for the mobile semiconductor industry(~90%share).   Although ARM does play  in some non mobile end markets, especially in hard disk drives,  MIPS has laid claim to the "digital consumer" portion of the market. MIPS dominates  Digital TV (68%), Mobile TV  (60%), set-top boxes (cable - 70%; IPTV - 77%), and  DVD recorders (blue laser - 77%; DVD-R - 72%). 

THE MOAT
    Sandwiched between big guys at the top and bottom and a tiny fraction of  the size of ARMH, it would appear that MIPS is in a very precarious competitive position. But the nature of processor technology,  unlike other chips  that are so vulnerable to commoidization, is  that each processor architecture once designed in almost never gets designed out.  The secret sauce in this business is what is referred to as the ecosystem, the hundreds and hundreds of software programs developed to support  a processor.  MIPS provides system developers with a  a broad array of software engineering  development  tools such as compilers, debuggers, tests, middleware, and reference designs.  As might be expected, design engineers are human and get used to designing with a certain architecture.  The availability of this supportive technology is a huge incentive for anyone building a new system to stay with the standard and an obvious barrier to entry to anyone attempting to create  a new standard. It's a classic network effect  created  by adding more licensees (>200) who in turn give third parties  more of an incentive to come in and provide their supporting software which is what the design engineers crave in the first place.   This is how MIPS and ARM have found a way to coexist and enjoy near complete dominance in their respective end markets.
    There has been a lot of buzz surrounding Intel's new Atom processor
http://online.wsj.com/article/SB121686601364079915.html which  is Intel's attempt to diversify beyond its traditional business. Initially Atom was only seen as a threat to ARM and its near monopoly in the high volume / high margin mobile handset market. But in their latest announcement, Intel says that its new chips can be used in a wide range of devices, from communications hardware to automotive applications to TV set top boxes. While no one in their right mind would want to compete against Intel,  there is no reason to think that Intel  will be able to come in and win share overnight given the fact that ARM has been trying to take share from MIPS in these markets for years with minimal success. It is definitely something to watch but it is important to remember that it is not enough for Intel to convince potential licensees that they should ditch their current architecture (not just the microprocessor but  the entire MIPS system built around the chip). In addition, Intel has to convince   the 3rd party software designers that currently support development  to support a new architecture. Intel counters that since its x86 design is so well known, developers will have an easier time creating applications.   At the end of day the microprocessor is a commodity and the ecosystem is the franchise and the nature of the business is that switching costs are very high.  

FUTURE DRIVERS OF GROWTH

·       Chipidea

    In August of 2007, MIPS acquired Chipidea for $147mm,  the world's leading provider of outsourced analog  IP. Analog IP is one of the last frontiers in IP outsourcing. Unlike processor and memory development that is almost entirely outsourced to a small group of vendors,  analog components are largely done in-house.  If you look at semiconductor IP outsourcing opportunities in the system on Chip (SoC) market, there are four basic applications: 1) Control Plane Processor; 2) Embedded Memory; 3)Data Plane; and 4) Analog Converter. The data plane  is where customers differentiate with their own IP and processor and memory functions are almost entirely outsourced. That leaves the Analog plane which  involves power management (mobile handsets), analog to digital and digital to analog , data conversion, signal conditioning, and connectivity (USB, HDMI). All of these constitute a small portion of the system cost and seldom differentiate one electronic device to the next, making them ideal candidates for outsourcing as outsourcing analog components to field proven third party IP can speed time to market and remove a key bottle neck in migrating to smaller process geometries.
    In spite of the rationale for outsourcing, there are relatively  few analog IP vendors. Traditionally the  knock against analog IP has been the perception of it being customized for every application and semiconductor product. Consequently anyone wanting to sell  IP on a large scale would   need a large number of IP blocks to achieve any meaningful operating leverage.   Chipidea has assembled the industry's largest number of IP blocks (>400) and the largest number of analog engineers (the scarcest resource in the industry), giving them the greatest  ability  to mix and match between various applications. 
    Mixed signal and standard analog ICs are a $48 billion market (CY08). Assuming a 20% R&D /sales ratio and that 25% of a R&D budget goes to analog, that leaves a $2.4 billion addressable market,  where only only 13% is currently  being outsourced. The opportunity is huge and no one is positioned better than Chipidea given the proven sales synergies between MIPS's processor business and Chipidea.  MIPS targets a 20-25% operating margin for Chipidea as they migrate the business model  from an engineering design service   based business model to a model more  similar to the processor business group.  Over the long term as the price to design SoCs goes up and becomes more complex, more designers are going to outsource the  analog piece.

·       Microchip

    Two years ago MIPS licensed  its 32-bit microcontroller technology to Microchip, the WORLD'S  LARGEST microcontroller vendor (>60,000 customers). This was a coup d'etat in the industry that cannot be underestimated, especially since ARMH has been  aggressively pursuing the 32-bit  space. Currently there is a massive upgrade cycle involving lower end microcontrollers as users migrate from  8/16 bit to 32-bit controllers. Microchip will be standardizing its 32 bit products around MIPS which means that a whole new customer base  -- thousands and thousands of programmers -- will be exposed to the MIPS architecture.   Revenue from this deal should start ramping in  two years as volume shipments begin in CY08. MIPS's management estimates that Microchip ships 500MM 16-bit products every year.  Granted that 32-bit controllers are still in their infancy but they will replace 16-bit controllers as costs shrink and customers want more performance.  Microchip is launching with 5 products that contain the MIPS processor and ASP's  are in the $2-5 range.  The royalty rate is typically 1% (that is low end and conservative) of revenues which means that   the royalties from Microchip could reach $3mm/ year a in the next two years and be on a trajectory for >$10mm a few years from now as16 bit completes its transition to 32bit. In the context of Microchips's overall biz which ships 5-7 billion 8/16 bit  units a year, 100mm chips with MIPS architecture is not  an unreasonable assumption that should ramp to 200-300mm over the next 3-5 years.

 
VALUATION
    There are a few ways to value MIPS. One metric management talks about is to go through the theoretical exercise of modeling what would happen if  MIPS were to stop all spending on Opex today  and just  milk its current and future royalty pipeline.  Obviously this is  a unrealistic assumption  as MIPS will never  do this. In fact, MIPS's biggest licensees will never let this happen as they are beholden to MIPS for their next generation of chips and frequently come to MIPS requesting new designs. Nevertheless, going through the exercise yields a NPV of after tax royalties equal to ~$127mm.  Since we assumed in this example that MIPS discontinues investing in the business, we don't even assign a terminal value at year 5 but you could safely assume that the royalties would  continue well after year  5 but  decline.

($ millions)
Year                                      2009      2010     2011      2012     2013 
Royalty Revenue (1)               49.02     50.49    52.00    53.56    55.17   
After Tax Free FCF(2)           31.86     32.82    33.80    34.82     35.86   
DCF                                       28.97    27.12     25.40    23.78    22.27 
NPV of Royalties =$127mm                     
(1) We have royalties growing 3% yoy given the fact that only 1/3  or the 120 licensees have rolled out royalty generating products (licensees take 3-5 years after licensing in the MIPS platform to design and ship products in large volumes).Theses royalties have been growing steadily for the last five years and  drop almost entirely to the bottom line (98% gross margins)
(2) 35% tax rate
(3) We used a 10% discount rate since the royalty stream is essentially an annuity payment. 
   
 
    MIPS has 44mil shares outstanding, $15mm in cash,  and just refinanced a $25mm facility. The stock is trading at $3.64 giving it a TEV of $170mm. The last several quarters have been challenging for MIPS, partly due to the  inherent lumpiness in its processor contract licensing business  and partly  because of  IFRS to GAAP accounting issues related to the acquisition of Chipidea. Needless to say, MIPS's Q308 pro forma operating margin of 9% was not even close to the 20% margin management was guiding to for Q408 and a far cry from the 25% margin MIPS used to earn before its acquisition of Chipidea. Incrementally things are getting better and there is operating leverage in the business model as management has been hacking away at Opex for the last few quarters and doesn't see it growing going forward.  For Q408 MIPS has guided for 10-16% operating margins.   As of Q308 (FY ends in the August quarter) and at its current run rate, MIPS will exit 2008 with ~$105mm in revenue and nothing spectacular happening on the bottom line. It is only when one goes out to 2009 that MIPS becomes a real value. Per management's internal guidance, we expect the FY2009 top line to grow by  20% given continued strength in processor royalties and the completion of the Chipidea integration. At this level of revenue, MIPS can easily achieve 20% operating margins resulting in $.38-.41/ share in EPS.  That means you are getting a durable semiconductor  franchise with high barriers to entry  for 10x FY09 earnings.  Historically this is one of the lowest multiples MIPS has ever traded at.   After reporting  several clean quarters and  demonstrating the long term growth profile of Chipidea , we expect MIPS will trade at  a premium to its current valuation. At 13x FY09 EPS, for example,  MIPS is worth  $5 a share. MIPS has sold off recently because of its exposure to the consumer , the messy integration of Chipidea, and  the market pricing in a liquidity premium for smaller capitalization companies.  While these are concerns, they  have been more than priced into the stock and are totally ignoring MIPS's annuity-like royalty stream and the uniqueness of Chipidea

RISKS
-customer concentration -Broadcom represents over 10% of revenues
-correlated to the  SOXX
-weak consumer markets
-retaining intellectual capital
-competition form ARM and Intel




Catalyst

-report clean quarter without acquisition related charges
-sign more licensees
-operating leverage
    show   sort by    
      Back to top