The Starboard-powered bull-case for Tessera ($18.73, $1 Bn market cap)

The Starboard-powered bull-case for Tessera ($18.73, $1 Bn market cap)

As Starboard emerged victorious from the Proxy battle in May 2013, Tessera’s roadmap to value realization appears clearer to all investors: what had been dragging the company in the past 5 years had been (1) its inability to renew licenses / collect revenues and (2) unsuccessful attempt to vertically integrate its digital optics business. Starboard looks to reverse course, bring on litigation and IP licensing experts to milk the IPs, and aggressively cut OpEx / CapEx via shifting the DOC business to an outsourced, fabless model.

Via a SOTP approach, I estimate that currently TSRA could be worth at least $14.5 per share in the bear case ($7 cash + $2 Amkor litigation + $5.5 base case IP Licensing), $ 18.5 per share in the base case ($7 cash + $3 Amkor & Micron + $7.5 base case IP Licensing with Micron + $1 in DigitalOptics), and potentially $25-30 in the bull case (35-60% upside) if the optionality in litigation, IP licensing rejuvenation, and MEMs migration for camera modules all play out. Given Starboard’s ~6% stake and a new board with ample IP licensing and litigation experience, TSRA can be a compelling opportunity should there be a pull-back to the $14-16 range.

Why this opportunity exists

-             Tessera is a small company with ~$1 Bn market cap that cannot generate much trading revenue for sell-side firms.

-             IP Licensing for semiconductor products is a highly specialized field that has poor disclosure, earns unpredictable revenue streams, and can be subject to massive technological changes. Given how difficult it is to understand and predict, no sell-side firm covers it as a priority, and most generalist hedge funds and analysts place it in the “too hard” bucket.

-             A simple screen shows TSRA has declining revenue and negative EBITDA and EPS.

-             Most of whom I speak with either have not heard of the company, views it as a legacy technology company, or view it as merely a semiconductor-packaging IP company.

-             While informative pitches do exist on various investments sites such as the VIC and Sumzero, I have seen no updates post May 2013 after Starboard’s board victory.

SOTP Valuation

I believe it makes the most sense to value the company on a SOTP basis: (1) Cash, (2) IP portfolio, (3) Litigation optionality, (4) DigitalOptics (DOC hereon) optionality. The framework is set up such that one can infer, by conservatively estimating the worth of TSRA’s IP portfolio, whether the optionality on litigation and DOC is worth paying for.

Cash $7 / share: As of 2Q13, TSRA has $380 mm in cash and equivalents, which translates to ~$7 / share, or ~40% of market cap. Note that the risk of imprudent cash-burn on low ROIC projects is materially mitigated now that Starboard is in the helm.

Litigation can yield ~$2 – 4 per share: Tessera expects ~$200 mm of episodic revenue over the next 12 months as per Richard Hill, Chairman of board, during the May 15, 2013 JP Morgan conference. In particular, the Tribunal of ITC issued a Partial Award from Amkor of at least $125-130 mm as of February 20, 2013 (which TSRA had won similar cases time and time against Amkor in the past). Timing remains uncertain, but good source of optionality here especially when TSRA cites that its return on litigation spend historically averaged out to be $2.7 damage award per $1 of litigation fee spent. Note that this $200 mm figure may also include Micron’s ~1 year unpaid episodic

The IP Licensing business is likely worth ~$5.5 – 7.5 per share as a run-off case with 5-6 year life-span and materially more if TSRA successfully extends royalty drop-off and migrates to 3DIC IP.

Tessera’s IP segment holds ~1500 patents for semiconductor packaging and licenses these patents to over 50 companies in DRAM and related industries w/ 99% gross margin on its royalties. In mid-2012, it lost 2 key licensees in Powertech and Micron (25% and 19% of revenue in 2011 respectively).

I agree that (1) Micron’s renewal, (2) non-DRAM opportunities, and (3) extension of IP licensing business beyond 2018 can add substantial value to this segment, but the speed of technological evolution within semiconductor package dissuades me in counting any of the above arguments into my base case valuation (see appendix for background and rationale).

With only a 5-6 year life-span, no major new contract wins (~120 mm run-rate annual revenue vs. 200 mm a few years ago), no Micron renewal, and a lean corporate structure (likely given Starboard’s involvement), I estimate that TSRA’s IP business can run on a 70% EBITDA margin and is worth ~$5.5 / share with a DCF as my base case.

With various other assumptions, TSRA’s IP business is worth $4.5 - $ 9 / share. Note that the top-line within this mental framework tops out at ~$175 mm per annum, so if Tessera manages to successfully and meaningfully expand to TSV and other lines of licensing business and reach beyond 200-300+ mm in revenue, a $10-15 / share valuation is not impossible.

In fact, several points may get one more comfortable about a higher valuation:

-             The emergence of a DRAM oligopoly (Samsung, Hynix, Micron) with 90% market share can be very good for Micron’s pricing (as seen in many long MU cases). Given Samsung and Hynix’s acceptance to TSRA, it may not make sense strategically for Micron to forgo Tessera’s IP to save on a royalty that is <1 o:p="" of="" revenue.="" their="">

-             Invensas represents Tessera’s more nascent IP portfolio focusing on 3DIC. Given Hynix’s 8-year deal (vs. traditionally 5 year), we can hypothesize on the (1) staying power, (2) cost benefit, and (3) longer R&D cycle for TSV and next-gen packaging solutions.

-             Tessera’s acquisition of Allvia, a TSV development company, in November 2011 may be slipped many, but note that Allvia’s founder Sergey Savastiouk actually introduced the term TSV in 2000 and built the 1^st TSV foundry. Having him on TSRA’s side can give us a bit of comfort on the R&D side.

-             According to TSRA, the sales lead cycle is long and can take 12-18 months, so perhaps expecting anything on the non-DRAM side before mid-2014 can be difficult.

So at this point, we have $7 in cash, $2 in litigation put-back, and let’s say $5.5 assigned to the IP business with 5 year life-span and no Micron renewal. With Micron renewal, TSRA should get another $2 in value to its IP business, and ~$0.5 in episodic revenue for Micron’s 2Q12-2Q13 usage, leaving us $17  / share, or ~$2 / share of implied value for (1) Other litigation upside, (2) IP business upside, and (3) DigitalOptics upside.

DigitalOptics can be worth $5 / share

To combat the potential shortfall in semiconductor packaging and for growth purposes, Tessera had integrated 6 photography-related companies over the past 8 years (see below) to tackle the mobile camera frontier. In particular, the key Siimpel acquisition granted TSRA the chance to displace VCM autofocus, an incumbent technology that has been around for > 100 years and has 95%+ market share.

Traditional autofocus camera modules employ voice coil motors (VCM) to move the lens module along the optical axis of the camera. This technology – originally patented in 1874 – has reached a point of diminishing returns, where further reduction in size or cost creates unacceptable performance compromises. Tessera’s MEMs (microelectromechanical system) autofocus provides a solution. Much similar to other MEMs-based technology like gyroscope and accelerometers found in many mart phones, TSRA’s MEMs autofocus is smaller, focuses 2-5x as fast, generates no heat and has no degradation in reliability, and uses 99% less power but cheaper than the VCM.

Given the promise this product demonstrates along with other patents and parts, it is hard for TSRA not to get excited and try to make all the money—thus its 2008-2012 attempt to become a vertically integrated camera module supplier that would supply the $9 billion camera module market (especially when they can sell a camera module at $20-25 a pop). Unfortunately, this model proves to demand significant scale and manufacturing expertise that Tessera simply does not possess, and contributed to >$300 mm worth of operating losses and capital expenditure between the period. The management and Starboard had both realized this mistake and, into the future, DigitalOptics will operate as a fabless service and license provider much similar to its IP licensing business. I believe this laudable strategy sets TSRA on the right track to mass adoption – they need OEMs of scale and relationship to (1) dramatically ramp production if need be, (2) push this technology and associated software to become the new industry standard, and (3) effect significant cost reduction. Yole estimates that it typically takes 16+ years to bring MEMS from lab R&D to fruition (see below), and maybe we are finally at that point for MEMs autofocus.

Additionally, the largest global camera module manufacturer Lite-On (30% share according to the company) signed a contract with DigitalOptics on August 2013 to ramp initial production in the fourth quarter of 2013, and add high volume capacity in 2014 for its mems|cam. Coincidently, 2 months later in October 2013, China’s 6-8^th smartphone OEM Oppo has placed the first mems|cam volume production purchase order.

Looking forward, TSRA’s DigitalOptics business may be run and valued as a very similar fashion as its IP licensing business (50-60% EBITDA margin at 5-10x forward multiple). Assuming (a) 25% adoption on a base of 1.3 Bn smartphones for $10 mems|cam module where TSRA gets a 2% royalty and (b) 30 mm run-rate for DigitalOptics patents, the value of this business is around $5.0 – 5.5 per share. Clearly the valuation is all over the place

Catalysts:

-             April 7, 2014 trial against ASE, ChipMOS, STMicro, Qualcomm, ATI, and Powertech with potential settlement before that.

-             2H13 or 2014 Micron license renewal.

-             Favorable outcome or payment from Amkor case.

-             DRAM market recovery (which it appears should already be underway)

Risk

-             Micron may decide to litigate than to renew, another concern is that it is aggressively pursuing the TSV solution along with IBM and Intel. As this technology matures into 2015/16, they may attempt to ditch Tessera altogether. I do see this as a serious risk, but from speaking to contacts it appears that Micron’s royalties to Tessera relate to NOR Flash and other logic related devices that have nothing to do with DRAM.

-             Tessera could still lose key customers: but I believe the IP business estimate is conservative enough that this bear case has already been accounted for.

-             DigitalOptics ramp and cost-cut may take longer than expected – especially when VCM manufacturers cut cost in order to compete.

-             Mems|cam does have 2 competitors named Polight (funded by Texas Instrument) and Pelican (funded by Nokia), yet their products have the reputation of being fragile and expensive with a lower yield. There is little foresight as to how the competitive landscape may shape out, but Tessera does have a head-start in getting Taiwanese OEM and Oppo on-board.

Appendix 1: Background on Tessera, 3DIC, and TSV

Tessera’s core IP portfolio encompasses semiconductor packaging solutions that include (1) Ball grid array solutions (BGA), (2) solder bump / flip chip, (3) chip stacking, and (4) other advanced packaging solutions. At the risk of oversimplification, a semi-package is a protective package that (a) encases the integrated circuit within while (b) allowing it to quickly transfer electric signals to the circuit board. Most of these technologies are used on DRAM:

-             BGA had the industry start using metal balls instead of pins to transfer electric signals (much larger throughput and now industry standard)

-             Flip chips allow a much small package as the solution fuses the metal balls directly to the IC then “shrink wraps” it (instead of linking the IC to the balls with copper wires). This solution improves throughput once more and shrinks the size further.

-             Chip-stacking is simply putting one chip on top of another and linked either via solder bumps, substrates, wires, or TSV (Through Silicon Via, will get into later)

Yole Development has a very informative chart that depicts the innovation within this industry: this advancement in “fitting more powerful, smaller chips into small packages” comes naturally alongside the demand for smaller, more powerful devices (especially mobile). Unfortunately, semiconductor density has not only becoming harder to compress, the cost of doing so booms exponentially that only the largest DRAM firms can afford to scale production / R&D (Given the limit of lithography wavelength).

Source: Yole Development (left), De Dios and Associates  (right)

As a result, instead of trying to fit more transistors into 1 chip, the industry goes vertical and tries to stack chips another efficiently (3DIC). As it stands today, 3DIC solutions such as SiP (System in package), PoP (Package-on-package), and PiP (Package-in-package) are absolute necessities as the DRAM industry migrates to LPDDR2/3 and DDR3/4. Tessera still fares quite well in this regard given its (1) offering of entire packaging solution and (2) established solutions for chip-stacking, as proven by Samsung and Hynix’s license renewal until 2018.

However, I believe that assigning material value to TSRA’s IP licensing business beyond 2018 for the base case can be a dangerous proposition: not only does TSRA’s IP portfolio face meaningful expiry before this time, the current SiP, PoP and PiP solutions are likely only transitional solutions to a more advanced packaging solution called TSV (Trough Silicon Via). While the technology is not expected to reach mainstream capacity until 2015-2017, early sources have already indicated that TSV is superior in terms of # of connections (~5-10x vs PoP), capacitance per connection (~6x), average connection length thus lower latency (~200x), and relative power consumption (~6x) with only issues in cooling, testing, and yield rate. I am certainly no expert on this subject, but if IBM, Samsung, Micron, TSMC, Hynix, STATS ChipPac, Intel, and Amkor are all sinking massive R&D into this blossoming technology, believing Tessera can win the battle requires a certain degree of faith beyond my conservative (and ignorant) comfort threshold.