There was Infinera with the unflinching belief of the cost benefits of PICs (which in our opinion, was never proven), and being oblivious to a traditional long-haul player forcing Transmode to take a backseat in its metro program. Then there was Acacia Communications, although a highly proficient technology company, and the beneficiary of the positioning of its IPO during a perfect storm, the valuation for just a DSP player became absurd. Later it was Applied Optoelectronics, which pretty much had the 40G market to itself for intra-data centers transceivers along with a very low-cost operation, while there was inadequate appreciation for both the low-entry barriers at 100G as well as the other drawbacks of being in the hyperscale space. Now Lumentum Holdings remains the latest darling on the Street based solely on apparently being the market leader in 3D-sensing, while not questioning the great difficulty with achieving success, especially with the use of six-inch VCSEL wafers.
To the best of our knowledge, the production of wafers at that size has never been done previously at such a large scale. Philips, which is far out on the learning curve on this technology is now targeting the smartphone business with four-inch gear. So, Lumenum will inexorably pay a steep price in just being a pioneer.
Going back to the summer, we understand there were major technical problems at Apple over Lumentum’s VCSEL devices. We also picked up intelligence that Apple had gone one or two levels deeper, and was all over Lumentum’s suppliers, including for its epis, which are produced by IQE. We have discussed in the past about Apple being extremely strict, probably even more so than other manufacturers of consumer products, about its component vendors having stable processes and consistent yields.
We do not think that it is unfair to characterize Lumentum as getting carried away like a kid in the candy store about its 3D-sensing opportunity, including angering Apple in initially raving about its expected margins. Most critically, even if it turns out to be the case that any yield difficulties are not directly tied to the VCSELs themselves, Apple will not discriminate among its suppliers, and will tell Lumentum the yield is “your problem” – demanding even lower pricing for its lasers.
The double standard on the Street is remarkable. If Finisar were the recognized leader in the 3D-sensing space, we have little doubt that its valuation would have been adversely effected by the numerous stories over difficulty in getting facial recognition to work on the iPhone X as well as the delays in general of the production of the next-generation smartphones. Lumentum has been fairly invulnerable to such news. We also think Finisar received some extra punishment for being at least little more upfront about production delays in this space than Lumentum.
Also, while on paper, six-inch wafers are inherently more profitable because of the higher production in the number of dies, Finisar will not get much credit by financial analysts for being more prudent in sticking with four-inch wafers early in the game. Yet, Apple will take note on which supplier has shown the best judgement and the most experience with VCSELs.
While as always, fibeReality does not recommend securities, using history as a guide, just as happened with other sacred cows in the optical industry, we would not be surprised if Lumentum is destined for a dramatic adjustment.
[written by Mark Lutkowitz]