Who is a possible buyer?  Probably not Juniper…

In 2016 Juniper purchased silicon photonics company Aurrion in order to develop their 100 & 400G devices.  With this, Juniper hoped to be able to have a differentiated competitive advantage over others (ie Cisco).  Juniper’s touted advantage described in this Lightwave article from March of this year sounds very familiar:

 “While optical transceivers based on silicon photonics are plentiful, Juniper Networks believes its approach provides significant differentiation. The design strategy enables integration of photonic elements, including lasers and detectors, into the silicon process flow on the silicon wafer. The chip assembly design includes an optical loopback switch that connects the transmitter to the receiver. The switch enables full circuit testing during the manufacturing process, thus increasing the company’s ability to identify known good die. (The design wrinkle also will enable the performance of optical transceiver diagnostics in the field, Jones-Williams pointed out.)

 As illustrated in the photo above, the silicon photonics dies can then be flip-chip-attached alongside ASICs and other electronics on the same ball grid array substrate to create a transceiver. The resulting package can then be combined with DC-DC voltage converters on a PCB, Juniper Networks stated in a blog that describes the approach. A snap-on remateable connector enables fiber connection. Overall, the approach enables power and related savings – 20% over traditional approaches, Jones-Williams estimated.”

 It could be Aurrion is/was on the same path as POET and Juniper is close to realizing this differentiation.  A competitor may then look to counter that by working with POET to develop their own differentiated solution.  Or, a much more unlikely scenario, Aurrion flopped and Juniper is our Tier 1….

And yesterday, Juniper Networks, Inc. (NYSE:JNPR) Wells Fargo TMT Summit Broker Conference Call December 3, 2019 1:15 PM ET

An excerpt from this call, IMO relevant to the discussion, details anticipated timelines of 400G rollout for Juniper.  Essentially aligning with what Suresh has indicated with OI device development meshing with industrywide 400G rollout.  For Juniper, prototype/qualification testing 1H 2020, volume production 2H 2020.  Seems we’re on an identical trajectory, a race to the finish for sure!

Aaron Rakers

Okay. Shifting gears, 400. You brought it up earlier, that's kind of obviously insertion opportunity for you to get really into the hyperscale cloud customer base. So just walk us through what you see as you set up right now? How do we think about the availability of optics playing into the equation? And maybe I'll just -- the third part of that question would be, when do you think you have a good line of sight in terms of the POCs turning into true kind of footprint for deployments?

Kenneth Miller

Yes. No, great question. So the products are in market today. We have products, as many of our competitors do as well. There'll be more products coming over the next several months. So the products are there available, being tested. We feel confident about our systems, our software. Not only are we eliminated the disadvantage that we had at the 100-gig cycle, we believe we have some competitive advantage with our silicon. We have the most modern operating system now in the market with Junos OS Evolved. We feel we've done a really good job from product set. Now we just need to execute and do well in the proof of concepts, which we feel pretty confident about.

As far as timing is concerned, I mean the decisions are going to be made over the next several quarters. And the rollout, I expect to kind of start second half and be more meaningful in 2021. I don't have much at all baked into my 2020 financial model. I do think we should start to see some late in the year. It's largely dependent on the timing of 400-gig OpEx to your point, and that we expect to be kind of a second half story, and that will carry the system. The systems will be ready. The systems are ready. It's a matter of the OpEx cycle catching up, and we believe that's going to be a second half story.

Aaron Rakers

And do you think that the -- if the second half volume deployment start that means you've got design in visibility already starting in the first half? I mean, those POCs take typically for what length of time before you, say, the design in actually...

Kenneth Miller

So I think that the qualification cycles will obviously be done before the deployment cycles. We feel good about the ability to qualify. I still think the game as far as the wallet share and how much you could expect is going to play out beyond qualification. I expect there to be a multi-vendor qualification process of which we hope to be a participant and a few of these large customers. How much share we're going to win, I think won't be determined upfront. We'll both -- both vendors will get opportunities to prove themselves. Our goal is to obviously take some share upfront and then expand that share over time as we improve our capabilities.

Aaron Rakers

Okay. And how much is silicon photonics fit into the 400 story for Juniper? And where are we today on that?

Kenneth Miller

Yes. No, it's an exciting part of our opportunity. So we feel that the 400-gig silicon photonics, which still has some R&D to go. It's not a slam dunk, but we feel more and more confident every day, every quarter that we're making progress to the milestones we laid out to be ready for the cycle. So we're talking about second half next year. We feel that we'll have our product available really when the 400-gig cycle takes off, which is perfect timing. So again, it's not something I'm factoring in because we're not shipping product today, but it is a big opportunity that we're very focused on internally.

Aaron Rakers

And remind -- maybe take a step back, remind us again why silicon photonics is so important? What it does from the cost...

Kenneth Miller

Yes. So we think we could change the economics of the optical interconnect. And effectively, we could reduce the price. We could either reduce the price to maintain healthy margin or we could keep current price and have really nice margin. And what's important to note is the system cost of a switch or a router. It's about 30% of the total CapEx. The other 70% are actually the optical interconnects. So to really change the economics of networking, we felt it was important to go after where the big CapEx cost is and that's in these optical interconnects, and Juniper made that bet a few years ago when we acquired a company. We've been developing it. I feel it's a huge opportunity for us to either package the optics with our systems and win more system sales and/or sell optics direct to some large customers that are looking for a lower cost optic.

https://seekingalpha.com/article/4310350-juniper-networks-inc-jnpr-presents-wells-fargo-tmt-summit-broker-conference-call-transcript