Lately the market for shares in UniPixel, Inc. (UNXL) has resembled a battlefield. The myriad impassioned articles and comments here on SA, on other boards, and in the stock's startlingly high short interest demonstrate the adversarial tension. Coupled with the binary nature of the UniPixel story, this tension is bound to lead to an all-out, take-no-prisoners conflict where one side is thoroughly routed.
Longs have been on the offensive this year, winning a few early skirmishes and sending the shares up over 100% YTD. Yet these minor clashes seem to have only emboldened short-sellers who have responded by entrenching themselves further each week, increasing their borrow to 4 million shares, which amounts to almost half of the 8.3 million share float. Clearly, the real battle is yet to come, and it could be spectacular.
This battle will be joined as UNXL makes, or fails to make, its next partnership announcement in the coming weeks, and it will draw to a close as the company scales-up production over the next three to twelve months.
If shorts are correct and UNXL cannot execute, the stock will quickly trade to hat-sizes or below, obliterating long gains. If longs prevail, they can look forward to an extraordinary ride. Feeding off a frenzied short squeeze, optimistic longs will reasonably set their sights on a $200 to $265 share price in twelve to eighteen months, as manufacturing partner(s) fire up their printing presses and UniBoss quickly establishes itself as "the" touch sensor in large format displays. UniPixel's promise of somewhat extraordinary share price appreciation is dictated not only by the fact that UniBoss could quickly become the standard in making touch ubiquitous in large displays, but because the business has already been won (assuming the ecosystem partner is signed).
For shorts the impact of this scenario may verge on cataclysmic. Shorts must borrow to enter into their trade, and they're subject to unlimited losses. A big win for UNXL can quickly translate to margin calls and monstrous losses for this cadre, who would be forced to outbid bullish longs in a thinly traded security in order to buy-in their position en masse. A short squeeze, a small float of shares, and a massive book of business mean that if UniPixel can successfully demonstrate manufacturing scale, the appreciation seen in the last few months may very well look trivial next to what we see in the coming year.
Short-sellers' ability to reinforce their position and keep UNXL down has slowed to a trickle recently, casting an ominous shadow ahead of the looming fight. UniPixel shares are now exceedingly difficult to borrow. UNXL has consistently made the Reg SHO Threshold List for the last few weeks, meaning that the levels of "failure to deliver" have exceeded certain thresholds. UNXL is being shorted even though it can't be borrowed - this is also known as "naked shorting." Charles Schwab (SCHW) has been regularly sending out system-wide emails to clients holding UNXL, asking them to lend their shares (and offering a paltry 5% rate). The stock loan group at my prime broker is unable to locate shares for borrow, they see daily buy-in pressure, and their last locate carried a negative rebate in the high 30 percent range, which would place the borrower on the hook for almost 40% per year in interest payments.
Who Has The Most Fight?
Having established what's at stake in the impending battle, we need to understand which combatants hold the strategic advantage. While the outcome cannot be predicted with a high degree of certainty, we can get some ideas about probabilities by understanding each side's strategic positioning. Below, I present some of the most important arguments, and show why the shorts seem to be at a strategic disadvantage at this point.
Because my fund is a shareholder of UNXL, I will face a chorus of criticism from the short community for offering a "one-sided" assessment. However, I have spent a substantial amount of time researching UniPixel's short story. My knowledge of the short case is sound, and the presentation of it is objective. If the shorts have more substantive material to offer, we would all like to hear from them.
In The Trenches With Shorty
Most of us have read the short stories previously published on SA. These range anywhere from ridiculous, deceptive, ill-informed, to just weird. Some of the articles carry the seed of a short thesis, but jump to outlandish conclusions based on flimsy analysis and little substance. There's no need to waste a lot of time refuting these pieces. A number of authors on SA have already covered that ground. Needless to say, UniBoss exists, UniPixel does produce it, the market opportunity is large, finished ITO sensors are expensive, and Dr. Bob Petcavich works for the firm full time. These are facts that you can establish for yourself with minimal due diligence.
Instead, we'll look at a more informed and realistic short perspective. The most cogent thesis I've heard came from someone who was close to UniPixel a few years ago. This contact had a personal grievance with management and is involved in a business that could compete indirectly with the firm, so he may not have been disinterested or objective. Regardless, he was knowledgeable about the company's history and the industry, and a lot of what he said made sense.
This contact acknowledged that he had spoken with a couple of shorts. The fact that his arguments were showing up almost verbatim in various short posts on SA and elsewhere indicated that this contact had given shorts a good deal of ammunition and conviction. Some shorts had even taken his arguments one step further, liberally sensationalizing them to buttress the short case.
The gist of his thesis is that history will repeat itself at UniPixel because the main actors have not changed. In his view, UniPixel was unsuccessful in launching both Time Multiplexed Optical Shutter (TMOS) and Fingerprint Resistant Film (FPR) products in the past due to the fact that they weren't able to commercially produce the products and/or they didn't understand the market opportunity. He also noted that the company has engaged in partnerships in the past that went nowhere, citing the likes of Samsung (SSNLF.PK) and Avery Dennison (AVY).
Some shorts have misconstrued one element of his argument. My contact's claim that nothing came of previous agreements was manipulated to suggest that the deal announced in December is not real. The fact that UniPixel was not allowed to disclose the customer's name only added fuel to fire in this conspiracy theory. Some shorts have warped his commentary even further to claim that the whole enterprise of UniPixel is a fraud.
Of course all of this is a far cry from this contact's idea that because the company won't be able to produce at a commercial scale, nothing will come from the agreement. He actually states that he thinks the UniBoss technology is interesting (and he notes that it probably makes UniPixel an appealing takeover candidate) - not that he thinks the company is a fraud.
There's no need to spend much time addressing the fraud argument. Texas Instruments (TXN), N-Trig, and a large PC OEM (rumored to be Dell (DELL)) have each spent several months and a significant amount of money making themselves comfortable with UniBoss and the UniBoss production process. They don't think UniPixel is a fraud. Ditto for several controller companies with whom I've spoken. Further, we have been able to triangulate through a controller company, who is working with both UNXL and the PC OEM, the existence of the relationship. There is no doubt that UniBoss is for real, and so is this agreement.
It's an indisputable fact that UniPixel is not producing TMOS or FPR film at this point. It's also absolutely true that the commercial scale-up of Diamond Guard has been delayed. The question is how should we interpret these facts? While history doesn't always repeat itself, it often rhymes. Accordingly, we should really be asking ourselves what that history is and what kind of rhyme we can expect.
Time Multiplexed Optical Shutter technology was developed by UniPixel as a novel, quasi-MEMS display technology for flat panel televisions. Without going into the gory technical details that are well above my level of understanding, the technology promised the opportunity to make flat panel displays better and less expensive than traditional LCD technology, while still using the LCD manufacturing infrastructure.
TMOS faced an uphill battle. The technology was complicated. It involved moving parts, and it integrated with a number of discrete television components, making it even more complex. Moreover, if you look at the existing FPD technologies (LCD and OLED), each was put through development with the help of multi-billion dollar investments. UniPixel intended to disrupt this ecosystem with technology that had been developed for just a few million dollars.
Considering the substantial obstacles, TMOS did pretty well. UniPixel signed Samsung to a joint development agreement and was able to develop working prototypes. The company ran into problems, however, working with its partner. Samsung is notorious for a "do whatever it takes" attitude toward acquiring technological know-how. The company supplied UniPixel with faulty backplanes, slowing TMOS's progress while soaking up as much knowledge about the product as possible. Meanwhile, technological advances allowed pricing for LCD televisions to fall rapidly, reducing the appetite of panel manufacturers to switch to TMOS. In the end, UniPixel's failure with TMOS wasn't that they weren't able to commercialize the technology; it was that they became overly reliant on a single partner, and they didn't assess the market dynamics correctly.
Fingerprint Resistant Film was a somewhat different story. It was a less involved technology than TMOS, and it was launched commercially. The issue with FPR was that UniPixel didn't sell much of it. UniPixel's distribution agreement with Targus went south. Targus moved to change the terms of the deal shortly after signing with UniPixel, because one or two competitors had become very aggressive with pricing. UniPixel was unwilling to sell to Targus at the newly proposed level, which would have implied a less-than-fair royalty for the use of the technology. Once again, the problem wasn't in commercialization of the technology (although some shorts claim that yields were suboptimal), the real issue was a faulty market assessment.
The shorts are correct in that UniPixel's management has made errors in the past. The gravest mistake was the overly optimistic market opportunity assumptions. Strategic decision making about partners was also faulty. It seems, however, that the lack of progression from development to commercialization was due in large part to these errors rather than an inability to commercialize faulty technology.
Will History Repeat?
Is management likely to repeat the errors of the past and misjudge the opportunity for UniBoss? Our short contact feels that this is precisely the case. He believes that the touch sensor market is fiercely competitive, and that many investors underestimate the level of competition. He thinks that UniBoss prices will fall precipitously because: 1) ITO sensor prices are plummeting; and 2) there are 6-10 other metal mesh technologies, many of them Asian, that will outcompete UniPixel.
There isn't much doubt that the touch sensor market is highly competitive. Not only will UniBoss compete with ITO and other metal mesh technologies, but also with a mind-boggling array of other technologies and materials. Nonetheless, we'll focus on ITO and metal mesh because these are the technologies against which UniBoss will compete initially.
ITO sensors will likely dictate the pricing environment for at least the next six to twelve months. ITO is the dominant incumbent technology, and metal mesh manufacturers seem content for now to price their product at a 10 to 50 percent discount to win business away from this technology.
For the last several years, ITO touch sensor prices have declined by 15-20 percent per annum. While this is a fairly steep decline curve, it's not atypical for technology hardware. Is there evidence that this curve is steepening? No. In fact, this article (paid subscription),which was forwarded by Mike Malouf of Craig-Hallum, claims the opposite:
The sources said Lenovo is seeking the technology since ITO thin-film type touch screen materials are in tight demand largely due to high allocations to companies such as Samsung Electronics, and because large-size thin-film type ITO material pricing has been on the rise.
It makes sense that ITO sensor prices would be stabilizing or even increasing at present. Windows 8 is causing a bit of angst for OEMs and ODMs who are worried about a lack of touch component availability as larger displays begin integrating the new operating system and incorporating touch functionality. Apple (AAPL) and Samsung are rumored to be putting competitors on guard as they secure their respective supply lines, and command a large portion of production. This tight supply viewpoint was corroborated by two analysts who attended CES a few weeks ago and spoke with multiple OEMs. Not only are hardware manufacturers worried about securing adequate touch sensor supply, but they also fear for a lack of chemically-strengthened glass (e.g. Gorilla Glass, potentially creating an opportunity for Diamond Guard).
An ITO sensor buyer explained to me his frustration that there's not enough current or planned, quality ITO film capacity in the industry. This would obviously augur for higher ITO sensor prices. He noted, too, that buyers are generally willing to pay up to two times more for high quality, highly reliable ITO film that comes from manufacturers such as Nitto Denko. So when we hear about very low priced ITO film rolls, we should consider their quality and conductivity before assuming that they would be used in a high-end consumer device where they would compete with UniBoss.
The margin structure of major ITO film manufacturers certainly doesn't support a rapid decline in pricing. Take a look at recent financial results from Teijin Kasei, Tyobo, Ushine Photonics, and Innovation and Infinity. It's no wonder that manufacturers aren't racing to build-out ITO film capacity - many aren't making any money selling what they're currently producing. Dominant supplier Nitto Denko is earning healthy returns, but is said to price its film at a 100% premium to the lesser-known brands. Further, since much of the equipment used to manufacture ITO film can be changed over to an alternative use, it's difficult to imagine that there wouldn't be some sort of self-correcting, supply-reduction response if prices were to fall dramatically.
The last point on ITO pricing is that UniBoss's sweet spot is large format (12" diagonal and up) displays. Because ITO begins to run into physical limitations at this size, the price-per-diagonal-inch for the sensors accelerates as the display size increases. Were ITO sensor prices to decline, larger format ITO sensor prices would likely show some price stickiness. Prices for the workarounds that enable these larger ITO sensors to function are unlikely to decline as fast as ITO-related costs.
Regardless of the fact that the ITO film and sensor markets may have found some stability recently, in the name of conservatism we ought to assume they resume their annual decline rate of 15-20% before long. The ITO supply chain is constantly figuring out ways to lower costs, and those savings are generally passed-through in this competitive market. The important take-away is that this scenario still allows UniBoss a very compelling opportunity. We'll discuss this point later.
What about the idea that other metal mesh companies will hurt UniBoss pricing? This doesn't seem to square with the facts either. In understanding the nature of this argument, it's helpful to get a feel for how UniBoss measures up competitively. We know that management of UNXL frequently cites the fact that UniBoss is the only product that uses an additive production process, making it the least costly solution to produce. Yet we're not going to rely on management to make the case. Instead, let's look at the issue from the perspective of hardware manufacturers and controller companies.
A contact at a large controller manufacturer told me that his firm had evaluated six to eight metal mesh companies in the course of their due diligence. This company is now one of UniPixel's key partners. The due diligence included work on the various film products, and it also extended to the respective production processes. Their assessment was that UniPixel would likely be one of the leaders in the space. They were impressed not only with the technology and the manufacturing process, but also with the management team and its ability to get things done.
In addition, UniPixel's announced UniBoss partners (TXN, N-Trig, & the PC OEM) completed extensive due-diligence on UniBoss, the UniBoss production process, and on other metal mesh technologies. Potential partners (think ecosystem & manufacturing partners) are doing the same work -- auditing UniBoss, the product's production process and its ability to scale, its ability to integrate with the display supply chain, its ability to weather extreme temperatures, its ability to withstand changes in air moisture, its durability when the sensor is bent repeatedly, its ability to retain electrical and aesthetic properties over time and in stressful environments, and more.
These partners and potential partners know each of these technologies at a level that you and I cannot. They see the warts and they see the opportunity, and they take into account current pricing, but they look much further down the road as well. They are making strategic decisions with long-lasting ramifications, and they want to partner with the process that scales best and will cost least in the long run. Based on this research, these partners (and hopefully a couple of potential partners too) felt that UniBoss compared very well with other metal mesh technologies, and they placed their chips accordingly.
UniBoss's strong competitive positioning vis a vis other metal mesh technologies can be seen in the terms the company secured in its first large commercial win last December. UNXL management sat across the table from a global PC OEM and said "no" when asked for equity. They gave the PC OEM a "limited exclusive license" to use UniBoss on a limited group of notebook designs. UniPixel management noted that the deal structure was not "normal," but it was exactly what UNXL had wanted.
A $50 million market cap company extracted $15 million in cash and a large volume commitment from a global PC OEM, and it did so from a position of strength. This says a lot about the value of UniPixel's intellectual property. Often young technology companies bend over backward to attract that first large customer, and we investors cringe at the terms. This was hardly the case for UniPixel. That a global PC OEM, well-versed with respect to the competing metal mesh technologies, was willing to agree to UNXL's terms tells us a great deal about how UniBoss stacks up next to the competition. My guess is that the deal currently being negotiated with the ecosystem partner will reflect the same.
Why would metal mesh companies attempt to compete on price at this stage of the game? There's an enormous opportunity in front of these manufacturers, and none of them have the scale yet to contemplate meeting five percent of this opportunity. Why would they beat each other up and ruin what could be a margin bonanza when there's plenty of business to go around? Why would they attack UniPixel when the UniBoss production process makes it the low-cost provider?
If you need further proof, talk to Atmel (ATML) or look at the company's marketing material for XSense. This product is being sold based on its unique functionality vs. ITO, not on its price competitiveness. The market opportunity for metal mesh is predicated on the idea that the technology can replace ITO with better functionality at a slightly lower price.
Atmel and Fujifilm seem to be the best positioned metal mesh manufacturers to challenge ITO dominance outside of UniPixel. As a UniPixel investor, I want ATML and Fujifilm to win business and help legitimize the technology. Unfortunately, they each seem to be running into difficulties at present.
This release from Carclo cites Atmel's difficulties in its "managed downstream supply chain associated with the differing manufacturing requirements between ITO and XSense touch sensors." While this article states that XSense's issues are "largely resolved," at least one industry participant believes that Asus has temporarily shelved the project. ATML would not comment as they are in a quiet period.
Fujifilm is rumored to be having difficulty with electrical migration in its silver emulsion films, and it may have stopped shipping to Lenovo. Electromechanical migration problems are endemic to silver where ions actually jump the rails from one circuit to the next, creating the potential for short circuits.
Regardless, it seems unlikely that UniBoss will see extreme pricing pressure from either ITO or metal mesh competitors in the near future. An important aspect of the short thesis seems to be a remote threat in the near-to-medium term.
Another factor to consider is that UniBoss has a very healthy margin profile. UniPixel's management team has repeatedly stressed the fact that the largest COGS line item for UniBoss is the PET substrate at $0.20/ft2 (<$0.10/ft2 at scale). With a very simple, continuous manufacturing process, UniBoss doesn't carry burdensome direct or indirect manufacturing costs either. My guess is that the product could have margins in excess of eighty percent when sold at a $20/ ft2 price point. What this says is that management has given itself an enormous cushion, so that if they misjudge the market in a meaningful way, they will still have a very healthy business. You might even say that they learned their lesson.
The notion that UniBoss cannot be commercialized is a core premise for our contact and the entire short community. We can debate the merits of the argument, but all the debating will be made irrelevant over the next few months. UniPixel will begin limited production of UniBoss during the second quarter, and full volume production will start toward the end of the second quarter or beginning of the third. The company may press release the first purchase order; this will give us an indication that UniBoss has passed all of its production testing hurdles. If production is delayed, the market will take a "shoot first, ask questions later" approach and send the stock reeling.
This said, the printing and plating lines are producing UniBoss right now. Several investors (myself included) have seen the commercial process at work. We've also seen the testing and measuring of the product that happens during and after production, and we can speak directly to the conductivity of the circuits, the width of the UniBoss lines, and their depth. But none of us are customers. We can make an educated guess that the product will be acceptable to the PC OEM, that yields are reasonably high, and that the supply chain will not have difficulty integrating UniBoss with its existing manufacturing processes. But until we see the revenue on the P&L, we won't know for certain.
Management and the Board of Directors at UniPixel are highly accomplished and intelligent people. What the shorts cannot seem to fathom is that these folks have learned from historical errors. With UniBoss they didn't give away the crown jewels to Dell, and tie themselves exclusively to the company as they did with TMOS and Samsung. If something goes wrong with one of their partners, they can rely on others.
Also, they don't need prices for ITO to stay elevated to make UniBoss a raging success. They can and will lower prices in the future, and they will still garner extraordinary margins. However, because of UniBoss's unique technical advantages price degradation may be gradual. The technology and strategy behind UniBoss is a world apart from past projects. Contracts have been signed, thorough due-diligence has been undertaken, history is unlikely to repeat or even rhyme.
Carclo Reinforces The Shorts?
One final point on the short case. This is not something that our contact brought up, but something that has been referenced on SA often. The Carclo claim against UniPixel is an annoyance for UNXL management, and in the unlikely event that Carclo wins its case, it would still be nothing more than an annoyance. UniPixel would pay a small royalty and move on.
It's highly unlikely that the claim will go anywhere, however. Read the claim - it rests on the idea that UniBoss is relying on technology used in the TMOS production process. Carclo/Xennia believe that UNXL is using that technology because they use it to produce XSense for Atmel. The fact that Atmel uses Carclo's technology means that its inkjet prints an expensive seed layer that uses palladium in relatively fat line widths. These lines require a subtractive, multi-step process to reduce to five microns, which is why XSense cannot compete with UniBoss in terms of cost.
UniBoss starts with an entirely different micro-contact printing process that doesn't use inkjet printers, palladium, or the surface energy principles that are core to the technology behind TMOS, XSense, and Carclo's claim. This is why UNXL has filed new patents exclusively around UniBoss - the technology is different from that which the company and Carclo developed around TMOS. It is this differentiation that gives UniBoss a competitive cost advantage over XSense.
You don't need to take my word on the subject. All of UniPixel's current and prospective partners could have easily exited their relationship with UniPixel had they believed that Carclo's claims were valid. Engineers from these partners have climbed all over UniBoss's production equipment and understand quite well how the production process works. They are convinced that UniBoss doesn't use a single inkjet printer head, and that UniBoss's ink is significantly different from Carclo's claim-- that's good enough for me.
Call to Battle
The ecosystem partner announcement will likely mark first blood in this battle. Perhaps from the short perspective it would be the lack of an ecosystem deal that would accomplish the same. Regardless, it's difficult to overstate the importance of this deal for UNXL.
Although the company will not name the partner in the coming press release, they have given us a list of candidates from which to choose (GOOG, MSFT, QCOM, or INTC). While we will likely be able to dig-in with channel checks post-announcement and discern who it is, the "who" is mostly irrelevant. Any of these partners would bring 25-100 million units annually to UniPixel and would firmly establish UniBoss as an early leader in the race to replace ITO. With this deal in the rear view mirror, the question from investors will shift from "How much can they sell?" to "How much can they produce?"
The answer to that question will come in the form of another offensive strike for UNXL shareholders. At some time in the coming months (according to the last conference call) UNXL will announce a manufacturing joint venture with a well-established film or sensor manufacturer. This deal would have the effect of making UniPixel a true IP-type company. Since UniPixel would need little in the way of operating structure, cash coming from the JV would largely fall to operating profit. It would allow UniBoss to rapidly achieve impressive scale.
My guess is that the ecosystem partner is playing a role in UNXL's decision to do this deal. The ecosystem partner may want UniBoss capacity to scale at a pace that UniPixel could never hit due to its limited human capital. The ecosystem partner may also feel that its ecosystem is apt to be more comfortable buying UniBoss from a large, well-known company rather than from UniPixel.
The first quarter conference call in May will provide more ammunition for the battle. We will likely hear that the company has received and recognized the first $5million installment from the PC OEM. We will also get an update on manufacturing progress, and potentially hear about the first purchase order. We may also hear clues as to whether the second quarter will see the first recognition of UniBoss revenue.
The short thesis calls for delays and excuses from management on the above milestones. Accordingly, we will know fairly soon whether the shorts are onto something or not. If they are, then the sell-off will be rapid and dramatic.
It all comes down to this: "Can management execute on its plan?" If it cannot, shorts will be rewarded with near 100 percent gains. If it executes, UniPixel will become the kind of homerun of which most investors dream. There's not a lot of room for compromise. There will be blood.
In 12-18 months the company and its JV partner could be producing upward of three million units per month. Allowing for some price degradation and allocating economics to the JV partner, this ought to conservatively produce somewhere in the neighborhood of $350 million in gross profit to UNXL. As an IP company with a streamlined operating structure, UniPixel would capture the vast majority of that profit after taking account for taxes.
How would we value $200 million in earnings growing at 30% per annum? It wouldn't surprise me to see the market put a 15-20x multiple on it. This would imply a $3 billion - $4 billion market cap, or approximately $200 - $265 per share assuming 15 million shares outstanding.
Does that back-of-the-envelope analysis seem outlandish and irresponsible? If so, you would have likely faulted me even more had I suggested last November that the stock would increase in value almost six-fold in the coming months without a penny of revenue. Sure, this scenario may turn out to be overly optimistic, but the valuation only requires the company to do what it has represented under terms of its contracts, assuming my guess on the ecosystem deal is correct. It doesn't even anticipate the fact that the company is likely to win further business, nor does it allow for any contribution from Diamond Guard.
Perhaps the greatest source of comfort I take as a UNXL shareholder stems from the knowledge that some of the best engineers in the world spent more than six months giving UniBoss and the UniBoss production process a complete physical, and then made financial and strategic investments in UniPixel. That these engineers regard the technology as worthy of these commitments tells me one thing. It says that these people believe the product can be produced, and it can be produced inexpensively.
Not only do these engineers know a lot more than we investors do about touch hardware from a technical perspective, they also get incredible access to UniBoss's production equipment and process, the know-how behind the process, and the product itself.
These companies are risking more than just their time and money in partnering with UniPixel, however. Product lines are being bet and supply chains altered based on UniBoss - not something these companies take lightly. These partners would not command their supply chain to work with UniPixel, risk missing a design cycle for several product lines, and risk angering their existing ITO supply chain if they weren't very comfortable with UniBoss and the economic advantage its use can impart.
Having Dell engineers (and those at either MSFT, QCOM, INTC, or GOOG) taking the same stance as the longs puts us at a decided advantage over the shorts. Even so, the outcome is not certain. None of us has perfect information. Only time will tell who wins this war.