There are striking similarities between MACOM Technologies Solutions and Infinera – not only regarding the most obvious characteristic that they have likely, more or less, reached rock-bottom on their valuations, but on the comparable paths that they took to get to this point. Each unwisely moved up the food chain from the optical chip level, which had immediate negative ramifications, as Infinera ignored the admonishment that it would rapidly result in a substantial decline in the size of the total long-haul system market, and MACOM quickly became what has been accurately described by some people in the industry, as an ineffective holding company for acquired parts of differing worth (not to mention MACOM entering deeply into the “hyperscale vortex” with Amazon). Each vendor was driven by over-the-top marketing/sales efforts, with the tendency to hype non-existent, economically-advantageous solutions, while giving inadequate attention to technology limitations – in the case of Infinera, the drawbacks with its PICs for new development efforts, such as in providing openness, and concerning MACOM, the sheer lack of coordination/communication between the piece-part operations. The two of them failed to take advantage of impressive acquisitions – the MACOM-AMCC merger had what could only be described as disastrous consequences, losing key technical talent from the PAM4 supplier early afterwards (we understand there are still difficulties with its CDR at 53 GBaud), and Infinera refused to make the higher-margin, metro products from Transmode a priority, partially because it would have been an admission that its PIC-based products were never ideal for mesh applications (explaining why these circuits excelled at long-distance transport, and initially DCI, which are often point-to-point apps). Additionally, they have been following similar strategies to try to get out of their predicaments in recent months.
While Infinera is at least pretending to go for its “Hail Mary” play by supposedly leapfrogging to 800G (with a relatively limited market size), with its PICs, while not mentioning the evident problems with them in just achieving 600G (we said, as early as November 2017, that “[o]ne way or another, we would not be shocked if ICE5 never becomes real in terms of commercial offerings.”), MACOM has similarly been promoting being part of the analog, “Open Eye” MSA in certain situations, over a DSP approach, although it should be noted that the use of analog in the past was proven not to be successful at much lower GbE rates. MACOM has also been working with GlobalFoundries on Silicon Photonics (SiPh), which clearly has its shortcomings, generally speaking, as a technology, but for the former, the ultimate target seems to be mainly for apps higher than 800GbE, which could be so way out in terms of timing, as to be an impractical consideration for getting out of its present quandary.
Moreover, while the positive effects of one of Infinera’ board members appears unmistakable, MACOM has chosen to make a CEO change from its own board. However, while a replacement at the top of MACOM was a must, there undoubtedly have to be major executive replacements at lower levels of the management structure. For example, even when it comes to SiPh, with engineers at different divisions not talking to each other, we have heard that the drivers are not currently being designed to be compatible with a Mach-Zehnder modulator.
Most importantly, our expectation is that MACOM and Infinera will be moving in more rational directions based on legacy considerations. With the latter, it will be increasingly taking advantage of the very long-term, operational ties of “Tellabs” with the incumbent ISPs, and the former can now focus principally on its RF chip roots, a sector that we viewed as its ace-in-the-hole ever since it delved so deeply into optical module components.
Although MACOM will be going against a heavyweight like Sumitomo, with its partnership arrangement with STMicroelctronics in the GaN space for 5G power amplifiers, it would not be terribly surprising if MACOM were able to obtain at least a moderate share of the space. Despite the latest reported changes in the international, political situation (at least, at the exact time of this writing), the cancellation of its joint venture with Goertek could still be vital to its widespread success.
In fact, although the reputations of the two companies addressed in this piece have been tarnished, the exposure of MACOM in optics is certainly far less than Infinera, as demonstrated by the former in already closing seven facilities around the world. Macom also has had the luxury of already repositioning itself as a pure chip play, with inherently higher margins, an option that is not available to Infinera. Moreover, the latter has the problem of a plethora of products to be sorted out.
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]