UniPixel shares traded down 26 percent Friday on startlingly heavy volume of seven million shares. A number of different factors came together to drive the stock lower. I'm presenting what I've heard about each of these below, along with some of my own commentary. I would emphasize that much of this is based on rumors circulating amongst the buy and sell-side on Friday.
1) On Thursday, the US PTO published an Apple (AAPL) patent application related to a touch sensor design for large area touch screens. I can understand how this would cause some uncertainty, but it's hard for me to believe this would create a large decline in UniPixel's share price. The patent addresses the design of a grid pattern for a touch sensor. OEMs/display manufacturers sometimes design their own sensor patterns for use on their machines. They may patent these designs. There isn't much new in this development, obviously most OEM and display companies don't manufacture their own sensors - they request that their sensor supplier use their proprietary design. In other words, if Apple were to one day work with UniPixel, Apple might have UniPixel engineer UniBoss according to the design assembled in the patent application.
2) There was a story on Bloomberg that Goldberg Capital, a top five holder of UniPixel, had lightened their stake. Again, it's hard for me to believe this headline would cause panic selling. Regardless, the story was false. Goldberg didn't buy enough shares in the recent secondary offering to keep their percentage ownership in the company flat - I guess that's what the author intended to say. Further, I spoke with Goldberg Capital Friday; it was buying shares in the open market during the panic selling.
3) Craig Hallum made some comments about Atmel (ATML) beginning to ramp production. Atmel had mentioned that it would soon be qualified on its last quarterly call, so I don't read this part of the Hallum call as the kind of call that would move UniPixel.
Another part of Hallum's call had to do with the fact that Atmel said that its line would produce 1mn units/mo which would equate to around $100mn annually. This boils down to around $8.30/unit if you simply divide the units by the revenue, so some salespeople at Hallum may have been making the call that Atmel would be pricing at around $10/unit.
There are a couple things wrong with this analysis. The $8.30 figure would be correct if Atmel achieved 100% yields, and produced at capacity for the entire year. If you look at most analyst models for UniPixel, you'll see that they have the company producing at levels far below capacity (some at 20-30%) for the first couple years of production. In fact, in my model UniPixel ends the year at a bit over 1mn units/mo capacity but is generating revenue at a rate of just $20mn annually. Yet you would be mistaken if you assumed that I was using a price of $1.60/unit for UniBoss. Instead, I'm assuming that the company is producing at levels far below stated capacity. Atmel is likely doing the same in its prediction of revenue generation.
The other significant point is that Atmel is selling its XSense into smaller form factor devices. In these devices, ITO is more competitive from a price perspective, and XSense would likely be priced at less than $20 ft2. If UniBoss sells into these markets, it would face the same pricing environment. I believe that the company has even referenced this on a call. Most analysts have reflected reduced pricing in their UniPixel models as production volumes ramp significantly, and the company expands its product base into smaller devices.
4) I thought the most significant rumor that came to light Friday was the one regarding UniBoss's revenue ramp delays attributed to Microsoft Blue (MSFT). To put this development into context it's necessary to understand how it fits into the short thesis. Professional shorts (as opposed to the shorts who comment on SA) don't believe that UNXL will be able to produce commercial quantities of UniBoss. Part of their thesis holds that as UniBoss begins its scheduled production ramp we will start to hear about delays. These delays will always be attributed to someone other than UniPixel, but the ramp will be delayed over and over again.
The Microsoft Blue delay fits perfectly into this short thesis. Many of the larger longs understand this thesis, and they may have been shaken by the Blue delay as well. In my opinion, this rumor was the most important driver of the selling on Friday. That an item like this can drive the stock down more than 25 percent, speaks to the nerves of investors who understand what a critical juncture we're approaching on the road to commercializing UniBoss.
Some have speculated that a delay attributed to Microsoft Blue is sheer nonsense. This is just an excuse to cover for the fact that UniPixel is having problems producing UniBoss. I don't believe this is the case. As suggested in this article, the updated OS from Microsoft is going to have a large impact on the notebook market, particularly in the third quarter. Why launch a new model with an operating system that will be outdated as soon as it hits the shelves?
The Microsoft Blue delay is real. Could it also be hiding something more insidious inside UniPixel? Sure, but I think the possibility is remote. The company's body language has never been as strong as it is now concerning the manufacturing process around UniBoss. The new product is benefiting from the combined oversight and collaboration of some of the best process and technology engineering talent in the world. These engineers hail from companies such as Kodak (EKDKQ.PK), Dell (DELL), and Intel (INTC) (the latter two being my best guess). UniPixel management seems very excited about the pace at which production is developing.
We investors need to get used to the fact that UniPixel is just a component supplier within several large systems. Delays are commonplace in these systems because the systems are complex and involve so many players. A last minute tweak to some software might cause problems for a controller which needs to be redesigned. This might lead to changes in the touch sensor or to other parts of the system, and there is a cascading effect which introduces more delays. All component suppliers are used to these delays. This is why it's important to have a portfolio of customers and design wins - one might be advanced while another is delayed.
UniPixel is putting together a portfolio, but it's not there yet. This leaves the company vulnerable to specific model delays (and to system delays in this case). Though the delay is real, it will not be very meaningful to 2013 numbers for UniPixel. More importantly, it shouldn't impact 2014 numbers whatsoever.
With UniPixel shares down 25 percent on no real fundamental change, long-term investors (like Goldberg) were accumulating shares Friday, and they will return to the stock Monday morning. Long-term investors believe they are being given a gift - after all, UniPixel's ability to produce UniBoss has not changed one bit. That said, the shorts have made a point. The Microsoft hitch has highlighted the fragility and uncertainty in the UniPixel story - some of which cannot be controlled by UniPixel and/or Kodak. This may be the first of many bumps in the road - get used to it. Delays logically increase UniPixel's risk premium. Each side has something to which they can cling regarding Friday's sell-off.
My perspective hasn't changed meaningfully. I think that the probability that UniPixel and Kodak will successfully commercialize UniBoss is still high, but the timing of that commercialization has become less certain (warranting a small sell-off). Recent support for the commercialization argument has come from some competitive successes. These successes have validated metal mesh as a viable technology. The hardware industry is coming to understand that not only is metal mesh real, it is the logical answer to producing large format touch displays. This is removing obstacles to the adoption of the technology. Very high quality products embedded with metal mesh sensors are appearing on the shelves of your local Best Buy and in other places. The proof is here, metal mesh works and it produces high quality displays.
I think the shorts are mistaken to believe that ATML/CIT, Mirae Nanotech, or FujiFilms could produce a workable mesh and that Kodak/UniPixel cannot. XSense shows us that a small company can invent a ground-breaking technology in the field - later to be commercialized by a large partner (even if that partner has limited experience in commercial films). Mirae's success shows us that a smallish functional film producer can bring a somewhat inferior, 10 micron silver emulsion product to work in a high quality, large format display (such as Samsung's 23" AIO shown above). UniPixel has a more elegant solution than these firms, and I believe the obstacles to UniBoss production are fewer. This small technology outfit produced a breakthrough in fine-line conductor printing, and now the behemoth functional printing expert, Kodak, will commercialize it with UniPixel's help.
None of this suggests that UniPixel's UniBoss is a guaranteed success, but it would worry me if I were a short. Knowing that UniBoss claims a process that is simpler than that of its competitors, and materials that are often easier with which to work, I would be uncomfortable assuming that Kodak and UniPixel could not achieve success. UniPixel has demonstrated they can hit product specifications in less-than-ideal, small run situations. They've also told us that the product holds up well in all manner of tests on a stand-alone basis, leading me to believe that the chances of testing successfully when integrated into a display are very high. They could fail, but the stars have certainly aligned favorably.
It's going to take a few quarters to rid the production monkey from UniPixel's back. There will be more set-backs, but there will also be plenty more victories. My guess is that future delays or set-backs will not impact UniPixel shares to the extent we saw last Friday.