By James Kisner, Senior Vice President, Equity Division Loop Capital Markets
It’s that time of year again! The entire optical communications industry will converge upon one location at the 2018 Optical Fiber Communication (OFC) conference. OFC is returning next week to San Diego, California for the first time since 2010. We thought it would be interesting to compare a few facts about our world at that time (March 2010) compared to today:
- Amazon’s market capitalization was $57 billion versus $744 billion today. Alphabet (which used to be called Google!) rose in value from $136 billion to $767 billion, while Apple rose from $202 billion to $888 billion. (Source: Factset.)
- Facebook’s user base in February 2010 was 400 million. As of 4Q 2017, that number had risen to 2.2 billion (Source: Facebook.)
- US smartphone penetration was around 20% (per comScore).
- The first iPad hadn’t been launched. (The first iPad wasn’t released until April 3, 2010.)
- Verizon’s most recent (2009) annual capex was $17.05 billion. In 2017, it was $17.3 billion. (Some things never change…at least not much!)
- Finisar’s stock price was $15.10 vs. $20.40 today. Notably the shares would peak at around $43 less than one year later in February 2011, driven by robust WSS sales, chiefly to Huawei. Over the same time period, the NASDAQ composite index increased 210% from 2,374 to 7,372.
- As of 4Q of 2009, Netflix boasted more than 12 million subscribers and that more than 48% of users had streamed more than 15 minutes of content in 4Q. Netflix’s total 2009 revenue was $1.67 billion. At the end of 2017, Netflix disclosed over 117 million streaming memberships streaming over 140 million hours of TV shows and movies per day. Netflix’s 2017 revenue totaled $11.7 billion.
- “Keeping up with the Kardashians” had already been on the air for more than 2 years.
With that trip down memory lane, it’s exciting to think about what changes the future may hold. Below we highlight a few of the key topics upon which we believe industry participants and investors alike will be focused.
All Eyes on China.
It’s fair to say that the demand for optics from China-based OEMs like Huawei and ZTE remains a hot topic for optical component vendors in particular. After robust demand in China in 2016 driven by the buildout of 100G long haul networks, 2017 proved to be a disappointing year and the outlook given by public companies for 1Q of 2018 was relatively dismal. With the benefit of hindsight, it has become clear that 2016 purchases were significantly in excess of 2016 end demand. This can be seen most acutely in the reported financials of companies like Oclaro and Acacia. In the case of Oclaro, revenue shipped to China declined 11% in 2017 after growing 74% in 2016. Acacia’s sales into the region similarly declined 24% in 2017 after growing 127% in 2016. A number of optical component vendors have publicly stated an expectation that demand for optical components in China will rebound in the second half of 2017 as provincial deployments gain steam and early trials of 5G begin to trickle out. Longer term, market research vendor Lightcounting is forecasting for optical transceiver demand in China to more than double from $1.6 billion in 2017 to $3.5 billion in 2023. A potential wrinkle in this positive outlook for domestic optical component vendors is China’s recently published 5-year plan from China’s Ministry of Industry and Information Technology (MIIT). A goal of this plan is to have three China-based optical component vendors in the top 10 (by revenue globally) by 2020, up from one or two today, and one of those companies in the top 3. Nonetheless, China will clearly remain an increasingly important end market for optical components over the coming years.
Will 400G Datacom Be Less Competitive than 100G?
While most of the action in the datacom market right now is driven by adoption of 100G client optics – namely CWDM4 and PSM4 – the industry is already looking ahead to adoption of 400G data center optics. Oclaro has already publicly stated it expects to ship PAM-4 enabled 400G QSFP56-DD transceivers in 2018 and quoted industry forecasts suggesting a projected market size for this form factor of $280 million in 2019 and $720 million in 2020. Of course, the OSFP form factor, with its larger footprint and power envelope, is also vying for a chunk of the 400G datacom market. A key question is whether the competitive dynamics in 400G datacom will be as brutal as 100G appears to be. Another lingering question centers on business models. Fabrinet, for example, has discussed engagements with multiple customers to effectively execute an “end run” on transceiver manufacturers by procuring parts/engines and assembling them. Amazon has been frequently mentioned in industry news sources as conducting experiments in this area, and we believe Google’s engagement with privately held Innolight bears some resemblance to a contract manufacturing model. We expect news flow around 400G datacom to be heavy at OFC 2018.
Dynamic Data Center Interconnect Market.
Data center interconnect (DCI) continues to become an increasingly important application as networks continue to move to a cloud model. In a presentation given at ECOC in 2017, Ovum projected the DCI market to exceed $4 billion in 2018. Hyperscale companies like Google, Microsoft, Facebook and Amazon hold massive sway in the adoption of DCI solutions, and power efficiency, programmability, and cost per bit remain critically important attributes for this customer group. Coherent and direct detect technologies are competing as two different ways to address DCI. Microsoft’s announcement at OFC in 2016 that it would use Inphi’s ColorZ PAM-4 based DCI solution made significant waves across the industry. We’d note that privately held Ranovus also recently announced a 200G direct detect solution leveraging its quantum dot laser and silicon photonics technology in a CFP2 form factor that it claims is lower power and one-third the cost of coherent and half the cost of PAM4 direct detect solutions. Of course, Acacia is also targeting this market with its 1.2 terabits/second AC1200 module that is sampling in the first half of 2018; Oclaro has similarly talked about a DCO module under development that is also capable of that level of data throughput. Of course, a number of systems-level companies have been addressing this space with purpose-built DCI boxes for years. We believe there will continue to be a lot of opportunity for optical vendors to share in the DCI market but we expect the space to continue to be highly competitive and dynamic.
400G/600G on Horizon in Long Haul.
While burgeoning growth of investments and dynamic competitive environment in data centers and metro networks is exciting, there are also notable developments afoot in long haul applications. For example, 400G technology is now ramping commercially with platforms like Ciena’s WaveLogic Ai. In a February 2018 blog, Ciena estimated that 400G adoption is ramping at 3x the speed of 100G adoption, and the pace of adoption of long haul technologies has been accelerating over time. We’d also note that Nokia, in advance of the show, published a paper on its Photonic Service Engine 3 (PSE-3) chipset and claims the chip will provide adjustable wavelength capacity from 100G to 600G. The PSE-3 chip will be sampled in the second half of 2018 and is expected to be commercially available in early 2019. Nokia also claims that the PSE-3, which implements “probabilistic constellation shaping,” pushes optical performance up against the Shannon limit, the maximum possible information transfer rate dictated by the laws of physics as modern science understands them today. While we are only now beginning to see the “green shoots” of 400G and beyond adoption, we are excited to see how the industry grapples with the limits set by physics as data growth continues apace.
Figure 1: Pace of Adoption of New Optical Speeds
Sources: Ciena, Dell'Oro
3D Sensing Market Has Officially Arrived. With the announcement at the recent Mobile World Congress in Barcelona that Samsung’s new Galaxy S9 and S9 Plus will contain facial recognition technology, we believe the jury is “in” that facial recognition technology adoption will continue to ramp in the next couple of years, particularly in the consumer electronics space. While the focus of OFC is clearly not on consumer electronics applications, we expect 3D sensing to be a prominent topic at the show, particularly amongst investors.
Don’t Forget 5G and Cable. The percolating ecosystem for 5G mobile networks will bring with it new architectures and new demands on optics for fronthaul and backhaul applications. Similarly, the collective efforts by cable operators to pull fiber closer to end users potentially creates new opportunities for optical vendors.
And with that we look forward to seeing you in San Diego!
Posted: 8 March 2018 by
James Kisner, Senior Vice President, Equity Division Loop Capital Markets
| with 0 comments