As we pointed out in a recent blog post, Finisar’s potential for diversifying its VCSEL business into the smart phone business should not be characterized as “tremendous.” While the impact will certainly be noticeable, there should be ample competition from suppliers, such as Lumentum Holdings, Broadcom, II-VI, and perhaps one or more other players to cut down on its revenue potential. Assuming the relatively high level of competition, we estimate that additional sales per quarter for Finisar would be optimistically no more than $60 million, and we project only a few more points on total gross margin because of the large quantity of other products sold by the firm. One can easily bet that a company like Apple will want to maintain a good number of vendors in order to avoid any disruptions in the supply chain, and to keep pricing of these components as low as possible. We also believe that the advantages resulting from this new capability will not by itself tend to excite people enough to lead to an unexpected jump in sales of these new high-end versions of phones. In addition, while the impact of a larger number of VCSEL arrays in autonomous automobiles may be realized in 2020, other considerations will almost inevitably get in the way.

With VCSEL arrays for LiDAR (3D imaging), a pulse of light is sent, and then the time it takes to come back is determined to calculate distance. The iPhone and other such devices will be convenient to use as a measuring tool for all sorts of things. Still, the ability to calculate dimensions just by sending pictures over the Internet in order to buy shoes or get customized suits, or inserting a couch into the photo of a living room to see if it fits the style or layout, would almost assuredly not be the kind of functionality that will result in long lines at the Apple Store – and neither would facial recognition in order to avoid the use of passwords.

When it comes to self-driving cars, we are just not optimistic about regulatory communities throughout the world just rubber stamping such innovations without years of investigation and testing. Insurance companies and other parties will potentially hold up these developments as well. These kinds of changes to the bulk of automobiles tend to be more incremental in nature. Moreover, it is quite conceivable that we are in the middle of a car bubble, at least in the US, which is about to burst, based on large amounts of credit that have been handed out for these vehicles.

Getting back to Finisar, for whatever reason, the Street tends to treat the company more unfairly that its competitors. We first pointed this out when it came to the raw deal it received when it was justifiably negative about the prospects for silicon photonics, especially before the meaning of the term became a moving target. It is quite possible that the investment analysts misinterpret straight-shooting for disingenuousness, which is hard to fathom.

In the most recent quarter, Lumentum was not punished for less than stellar results as well as near-term guidance because of this 3D-sensing opportunity. It is viewed as a leader in the space despite having relatively limited experience on the consumer electronics side, and it appears to have been struggling with VCSELs for high-speed communications. The safer bet is probably with the player that knows these lasers backwards and forwards, and has successfully produced them with big yields for many years.

Please find details on our Clash of Metro 100G Optical Vendors with Shifting Network Paradigm and Clash of Optical Component Vendors & Technologies in Data Center Networks here. For our daily updates, please follow us here.

March 17, 2017 at 5:24 pm

Mark,

The reality for those VCSELs for autonomous cars and cell phones is much worse for finisar and lumentum than even your note suggests.

This market is already largely owned by another player. And the price per unit is in the range of a few pennies. While the volumes are huge, finisar is too late to catch up. They will have to invest too much to get into the market and will never make a profit from it.

Other parties already own this space.

Reply

1. Mark Lutkowitz says:

March 18, 2017 at 7:23 am

John,

I suspect that you may be speaking about one or more European companies, which have specialized in the consumer space. In fact, there has been some speculation that Avago used one of these firms for the computer mouse business. However, forgive me if I do not totally buy that it may be too late for Finisar. Regardless, your insights make the stakes even higher for Finisar in that the Rump Session at OFC will be an attempt to make lemonade out of a lemon — and perhaps the silicon photonics discussion (which, of course, is personal for the vendor) is as much of a diversion as copper cables.

Reply

Paul McWilliams says:

March 18, 2017 at 10:29 am

There are not many opportunities to differentiate VCSELs with better performance, which means yield and production efficiency will be the driving success factors. FNSR evaluated going with six inch wafers, and even took a look at Anadigics prior to its acquisition by II-VI for about $78M. II-VI kept the 6″ fab, which will target VCSEL production, but subsequently sold the RF assets for $45M plus a $5M earn-out.

While I am not about to shortchange the fab talent at Anadigics, there are two risks in running VCSEL on six inch wafers. First, there is a inherent challenge in the epi layer. Second there is a challenge in handling the thin and brittle wafers. Both of these are yield risks that FNSR decided overrode the potential leverage of using six inch wafers, and resulted in FNSR choosing to go with 4″ instead (FNSR was running its VCSELs on three inch wafers).

I’ve known FNSR’s founder, Jerry Rawls for years. I think he is as straight a shooter as they come, and more likely to provide a conservative outlook than over-hype situations. I also don’t believe he would invest as heavily as he has unless he had solid commitments in hand from one or more substantial customers, and have reason to believe that the vast majority if not all flagship smartphones will move to VCSEL by the end of 2018. If correct, this will swamp the aggregate VCSEL production capacity that exists today.

Beyond the obvious aggregate volume that the smartphone industry will drive, it will also significantly lower costs for VCSEL modules, and that will open many new markets for the technology beyond self-driving cars. Within these markets I suspect well see requirements for arrays that run from about 100 to around 300 VCSELs per die.