Post Re-Org Equity: Spansion (CODE)

Every few months, I post an idea from the Distressed Debt Investors Club to give potential guests and members a feel for the ideas on the site. Remember, we have unlimited spots for guests that have a 50 day delay access to the database. We only have about 50 member spots left which give you access to the entire database and the forum (which I post on 2significantly more than this medium).

Last week, a member wrote up Spansion (CODE), a post-reorg equity with a target of $30/share with potential upside from there on additional royalty sales. Enjoy!

Company Name / Ticker: Spansion / CODE

The core thesis is that you have a hugely cash generative business, that is characterized by stable predictable orders, benefitting from a global trend towards increased electrical automation and increased content in all manner of industrial applications. EBITDA growth is continuing to grow steadily yoy each quarter in the core business, and now with the licensing revenues, CODE should have nearly double digit rev growth for the next two years and over 20% growth in EBITDA this calendar year. Incrementally you are long optionality based on litigation strategies and licensing opportunities on the IP portfolio. At $30, Spansion would be at 5.2x ’11 EBITDA and trading at an 11.6% ’11 FCF yield and we have yet to add value from royalties from Elpida sales or other incremental value from future licensing opportunities.


Business Description:

Spansion is a leading nonvolatile flash manufacturer. The primary focus of the company is NOR flash, which has superior reliability and fast read times, which makes it used widely for code storage, boot up instructions, and execute in place (XIP) applications in all manners of electronics. The company has ~36% mkt share in the embedded market. Spansion's segments are approx. 41% consumer and gaming, 22% auto and industrial, 12% computer and comms, 19% wireless and M2M, 5% other. The company will benefit by continued electronic content and active safety control features in autos, smart meters, increased industrial automation, and 3D TVs and gaming etc. They also are licensing their technology to Elpida in an important JV to build NAND. The embedded markets are characterized by steady demand and ASPs, high customer retention, little interest in advanced node technologies, and 5 to 10yr product lives.


Spansion was originally formed by AMD and Fujitsu. In the years leading up to its restructuring, it was a company run by engineers for engineers. Tremendous amounts of money spent on R&D, non-core business lines, and capital spending while primarily competing with Samsung and Numonyx to provide NOR flash for the hyper competitive handheld market. Low margins (if positive at all), tremendous cyclicality, and huge capital intensity were constants. Unlike many modern restructurings, Spansion actually went through both a financial restructuring and a true business restructuring. Some key aspects of the business restructuring included the following:
  • Exiting the bulk of the wireless business to focus on embedded markets
  • Reducing headcount ~60% and focusing R&D on core business & product extensions
  • Abandoning SP1 their new fab built in Japan, and creating a hybrid model whereby Spansion utilizes there fab in Austin, TX for ~65% of their product. This keeps it at running at near 100% capacity with an “asset-lite” model for the rest of their wafers. They utilize the same model with assembly and test facilities (2 owned but utilize 3rd parties as well).
As a consequence of the restructuring, Spansion cut sales approximately in half, but took gross margins from 4% to mid 30s and they are on their way to north of 40%. This was accomplished by focusing on the embedded market. By focusing on the embedded market, they are selling into customers whose product life is 5-10 years. The orders are stable and predictable and once they are designed in, it tends to be very sticky. They have excellent visibility as they do not produce much of any spec product, and this is corroborated empirically by 8 consecutive quarters that have met or exceeded guidance. Due to the longer term nature of the products made by the embedded market, the business has low capital intensity. In fact, the company suggested that these attributes were always attributes of the embedded business, but it was invisible to outsiders based on poor segment level disclosure and the cash generated by the embedded business was swamped by the cash furnace that was the wireless division.

Capital Structure as of 3/31/11:

Cap leases $1.7
Secured TL $251
Unsecured Bond $200
Cash and securities $308

Investment Thesis:

Spansion spent over $1B on R&D in the 3 years leading up to the bankruptcy. They already have process technology and IP for the next 2 smaller node sizes. This gives them a roadmap to produce more advanced flash products (transitioning from 110 to 90 to 65 to 45nm) for the better part of this decade. Spansion can now in effect, monetize much of the expenses incurred by previous investors. This is driving all sorts of new products specifically targeting existing and new markets. They brought 4 to market in 4Q10, with 17 expected this year. All at 65nm (at the forefront of NOR flash) and designed with the embedded market in mind. This is driving significant design wins that is expected to boost their embedded market share from 36% up into the 40%+ area (Numonyx is #2 with ~20%, and at the higher densities this duopoly is even more significant). These new products have driven numerous design wins which is one reason Spansion has quantified $15-30mm in top line as the estimate for 2011 from the Japanese tsunami and still haven’t lowered their guidance. Incremental demand in 2012 for the rebuilding efforts should further bolster pricing and volume. Also, Spansion now believes they will sell $75-100mm in NAND flash from their JV with Elpida in 2012. They see this as a ~$500-600mm mkt which think they will get ~approx. 1/3 of the share. This does not include any possible licensing royalties from NAND sales by Elpida. Spansion also recently settled all of their outstanding litigation with Samsung, which is providing for $150mm in revenues over 6 years. Also as part of the settlement, they stipulated an outstanding claim in the bankruptcy estate. Spansion agreed to buy back that claim and retire the shares.

Their products are known for their reliability, durability and ruggedness. This is one reason Spansion does so well in the auto industry, where products must meet not only strict specifications, but also be qualified and supplied by reliable suppliers. The content per vehicle is going up. The designs are done. All new safety and security systems, infotainment, and in fact every time you see an electronic screen in the cabin, it generally requires NOR. It’s necessary because it has instant boot up time and it doesn’t drop bits like other memory (NAND). These same traits are why NOR is going into industrial automation, advanced Set top Boxes, communications equipment, smart meters etc. etc. In fact, the same “drivers” of business that people like in companies like FSL, ADI, LLTC, TXN etc. all generally benefit Spansion.

In addition to the core business, which is growing with expanding margins, there are new opportunities to monetize Spansion’s 1696 US patents (873 foreign, as well as 502 and 840 applications are pending both in the US and abroad respectively). The royalties from Samsung will bolster the top line and cash flow by $25mm and ~$20mm per year (NOLs shield in US, some leakage in Korean sub) respectively. I believe we will see Spansion attempt to go after some or all of the following: MU, Toshiba, Hynix using the Samsung settlement as a guide to try to extract further value going forward.

Another example of Spansion being able to “monetize” its IP is with a JV with ELPIDA. ELPIDA is a significant DRAM manufacturer and had interest in entering the NAND market. ELPIDA and Spansion have formed a JV with a licensing agreement that provides royalty payments and product near cost to Spansion for its non-exclusive use of its Mirror-bit technology which is slated to ship product before the end of 2011. This NAND is not intended to compete against Samsung, Toshiba or Micron in the hyper competitive NAND found in handhelds, computers or tablets. This is “ruggedized” 1-2GB 43nm SLC NAND that is targeting niches in industrial applications where increased performance and reliability are needed. It appears to be a complimentary product and doesn’t appear to be likely to cannibalize NOR given early indications from design engineers. In discussions with Micron, they corroborate CODE’s contention that there is real demand for this new aspect of the NAND market. The $75-100mm target will be very backend loaded in 2012, but should ramp from there assuming the product timeline doesn’t slip. CODE should be able to keep low to mid 30’s on the GM on this (lower, b/c they will have to pay some small margin to Elpida above cost. Elpida is working on 3x and 2x nm NAND to compete with the big guys. The TAM for NAND is growing rapidly as is estimated at ~$25B. If Elpida can carve out even a piece of this mkt, and with a mid single digit royalty, this could provide a solid tailwind just on royalties alone. Spansion has no capital commitments to the JV. Spansion holders have optionality on this new product which I do not believe is discounted today.

All microcontrollers use NOR. Historically microcontroller manufacturers have utilized in-house NOR exclusively for consumption embedded into their microcontrollers. This NOR is bulky, not technology advanced, and there is likely to be opportunities to license Spansion’s technology to microcontroller companies looking for superior performance and smaller form factors. It’s unclear and probably unlikely that there is anything imminent on this front, but this has been confirmed to me by large microcontroller companies as being a truly sensible partnership opportunity. Spansion has actually indicated that they are now starting to field incoming interest from other companies on this front. This optionality is not widely understood and certainly not discounted.

The core thesis is that you have a hugely cash generative business, that is characterized by stable predictable orders, benefitting from a global trend towards increased electrical automation and increased content in all manner of industrial applications. EBITDA growth is continuing to grow steadily yoy each quarter in the core business, and now with the licensing revenues, CODE should have nearly double digit rev growth for the next two years and over 20% growth in EBITDA this calendar year. Incrementally you are long optionality based on litigation strategies and licensing opportunities on the IP portfolio.

These are based on conservative estimates and do include an estimate for what the NOLs are worth. There is good disclosure and given the bulk of the PV of the NOLs are usable within 3 years I get a PV of only the Federal NOLs of $165mm.

Multiples are primarily a reflection of three things: FCF conversion rate, volatility of operating results, and growth. Spansion has excellent FCF conversion based on its low capital intensity and its model. This is not a DRAM or NAND company! The volatility of cash flows has been very low over the last 8 quarters. No big surprises and no misses. It’s an annuity like business. Public comments from MU corroborate this when they discuss the nature of the embedded markets at Numonyx, and Spansion’s flexible model reduces operating leverage cyclicality. Spansion is growing, despite multiples lower then those of the yellow pages and newspapers. iSuppli suggests 3% CAGR for the embedded NOR market through 2014, which Spansion suggests does not pick up the whole market. Spansion is growing embedded NOR revenues (81% of revs) north of 10%, with the wireless business coming down (19% of revs) 10-15%. This has the core business up 6-7% on the topline prior to the Samung royalties. Add in the new royalty stream, and you have sales growth of 8.8% this year and 8.7% next year.

MU is a memory company and has a division that is ostensibly Spansion’s best comp. But in aggregate, nothing about MU’s business looks similar from a financial perspective. Its hugely capital intensive, exhibits tremendous cyclicality, is hyper competitive with companies with more resources, has poor if not negative ROIC, and short product life cycles. It’s almost the mirror opposite of Spansion. Spansion’s business attributes much more resemble the analog companies. ADI, LLTC, MCHP, ATML, ONNN trade with EV/EBITDA multiples ranging from 8-11.5x and FCF yields ~4.5-7.5%. Spansion will probably never fully escape its memory stigma, and is not well followed as it operates in a weird little niche that regular tech analysts don’t understand or spend any time on. So it will realistically probably never close the entire gap of relative value, but as the cash continues to pile up, which it has been doing for over 2 years, it will become harder to ignore. At $30, Spansion would be at 5.2x ’11 EBITDA and trading at an 11.6% ’11 FCF yield and we have yet to add value from royalties from Elpida sales or other incremental value from future licensing opportunities. This seems to me to be a rather conservative PT still at a significant discount to the analogs, if they continue to grow mkt share or gross margins, there is upside to this number.

  • Stock sold off hard after Japanese earthquake, and despite solid guidance quantifying the effects, skepticism in the market will fade over next couple of quarters - even more so 2012 demand to rebuild should be a tailwind
  • Removal of the overhang from the resolution of the claims pool and final distribution of reserve pool shares
  • Significant earnings growth as gross margins climb up over 40% bolstered by the transition to 65nm products and increased overall sales. This should help sentiment as this arbitrary level is used as a benchmark by many tech investors and might help change perception and thus the multiple.
  • I actually thought the stock would get a much better reaction to the positive Samsung outcome. All of the street estimates are too low for ’11 and ’12 as they have not updated numbers to incorporate the Samsung royalties or the NAND sales later next year.
  • The run off of fresh start accounting will drive GAAP gross margins and earnings higher as increased D&A resulting from inventory step up, cycle off and higher cost inventory no longer boosts COGS - street coverage have been focused on these and i think will be forced to revise targets higher
  • If SPI or serial peripheral interface NOR (primarily made by Macronix and Windbond) can somehow come up market to more directly compete with parallel interface NOR at higher densities which is where Spansion and Numonyx really dominate, that could change the stable ASP that characterizes the embedded NOR market. While possible, nothing currently suggests this is likely.
  • If Spansion does something ill advised with its cash as it continues to accumulate. This is a bigger risk to me, but hopeful that the 2 board members from Silverlake can act somewhat to mitigate this. Spansion has done accretive buybacks of claims in the bankruptcy, and this most recent retirement from Samsung is a continuing this.
  • MRAM, FRAM, phase change or some other memory technology comes along that displaces NAND and/or NOR. This has been a discussed risk for 20 years with no demonstrated increased likelihood of displacing current flash memory technologies. Adoption of NAND with error correction in industrial applications is a risk as well. However there is no sign that industrial chip set designers view that as a superior solution as they continue to use NOR.


Anonymous,  7/20/2011  

I followed this case closely. One of the main reasons why this business went into Ch 11 was because CODE's particular technology was replaced by a better one for use in cameras and other consumer electronics.

Does the writer understand the technology well enough to assert that this wont happen to whatever other platforms they are in now?

The other reason why I thought that this stock wasnt all that attractive was because when MU, which trades gazzillions of share each day has a cheaper valuation, what is so great about CODE in comparison?

Come to think of it, arent Intc, MSFT, CSCO and thier ilk trading at 5-6x ebitda? Why invest in this little guy when such large, liquid, wide-moat names are trading so cheaply?

1U Rackmount LCD 4/21/2012  

Keep up the good piece of work, I believe that your blog is really interesting and holds lots of wonderful information. I have to say, it's just an interesting and refreshing view.


hunter [at] distressed-debt-investing [dot] com

About Me

I have spent the majority of my career as a value investor. For the past 8 years, I have worked on the buy side as a distressed debt and high yield investor.