Sumitomo’s Datacom VCSELs: Saving Face

Avatar photoPublished on
Blog 1576436019 292

Despite being a data communications module producer, Sumitomo’s purchase of Emcore’s VCSEL business in 2012 was a mistake. The former did not really know what it was buying. (We understand there was even a lawsuit because Sumitomo felt it got sold a bill of goods.) For example, all of the high-quality epi growers had already left the American supplier. Yet, regardless of paying only a relatively small amount of money ($17 million) and despite the fact that there has been barely any association between the Japanese vendor and these types of lasers in the industry, there has been an insistence to avoid losing face on the deal. (It may even explain why employees who were let go after closing the operation in Albuquerque, New Mexico in 2014, were given generous, six-month severance packages). After the technology was moved to Japan, Sumitomo was willing to lose money on continuing development efforts. Every piece of manufacturing equipment in NM was either crushed or sold, partly because of issues involving transfers to Japan needing to be new rather than used, along with tax implications. Some excess equipment in Japan was repurposed, providing the reason why it has taken the company five years to make a VCSEL. fibeReality has also discussed that the leader of Sumitomo’s automotive cable harness organization is solely responsible for championing this cause. In the rest of this article, we will talk about the following five matters: 1) another example of somebody at the top of the ladder at Sumitomo, who has been responsible for maintaining a “happy story;” 2) the fact that the supplier is willing at times to go after niche opportunities; 3) the potential threat to its 850nm VCSEL business; 4) the main reason the firm will most likely avoid the 940nm space; and 5) its international trade policy.

While Sumitomo at the moment controls about 50% of the GaN amplifier market, our latest intelligence points to a high-level executive there initially becoming convinced that pseudo-bulk GaN crystal was wonderful, although it actually has numerous problems. Nevertheless, since making a strategic decision, the supplier has spent a good decade losing money, while chasing a theoretical goal.

In continuing to throw resources at the technology, it resembles the same mentality related to VCSELs. It will also be interesting to see what happens with Sumitomo’s joint development effort with II-VI related to GaN on SiC, as the Warren, NJ plant, at least in the past, had definite shortcomings.

Although historically, Sumitomo does not tend to be connected to high volume/low margin businesses, making the Emcore VCSEL purchase a head-scratcher from the beginning, there have been some examples of targeting specialty gear and being open to customization. Certainly, compared with the sensor space for VCSELs, focusing on the communications marketplace with these devices, definitely has to be considered a niche tactic.

Also, Sumitomo designed the first lightning cables — AOCs for the USB 3, which became a specialty item in that it never took off in a big way. Moreover, there has been a change in the future game plan with the cable harness business in making both copper and optical wiring versions.

The biggest concern for Sumitomo in getting involved in the VCSEL data communications space is the possibility of being decimated by six-inch VCSEL players, which are dealing with inadequate growth on the 940nm side. The company was hoping that Broadcom, which excels at 25G and higher capacity, and itself, would be able to dominate the sector. However, II-VI began to increase its share before the acquisition of Finisar (the latter having alienated customers on the datacom VCSEL transceiver side), and we have heard that Lumentum has initiated is own 850nm VCSEL marketing effort.

Perhaps one reason for Broadcom moving back to modules is that it can now dominate upstream with VCSELs. Sumitomo may be too far away in moving downstream, across the board, in order to get back into transceivers in a significant manner, and Lumentum, of course, will finish getting lower on the food chain. Apparently, II-VI is solely focused on 5G, relating to transceivers.

On Sumitomo’s strict datacom interest in VCSELs, the firm is still investigating the 3D sensing space, for long-term prospects in the automotive space. Nevertheless, in already feeding the 940nm supply chain in a very substantial way, by owning the world’s crystal substrate segment, it would not want to compete with its customers.

Concerning trade procedures, although Sumitomo has heavy sales to Huawei including for GaN amps well as for shortwave components, and there is even a personal relationship between the two CEOs, the former strives to stay under the radar with this association. In fact, customers tend to view Japan as more of a US kind of entity, while in comparison being nervous about CIG, as there is an increasing push for non-Chinese suppliers, given the current political situation. While Sumitomo will not necessarily state it publicly, it usually abides by the US policies, and so, there should continue to be an absence of restrictions with the company selling products, including VCSELs, to the States.    

For our fast-growing, totally separate, quick update blog, which is exclusively on fibeReality’s LinkedIn page, please follow us here.

As always, fibeReality does not recommend any securities, and this writer does not invest in any companies being analyzed by us.

[written by Mark Lutkowitz]