Q&A: For POET Technologies, the Pieces are Falling Into Place with Super Photonics
January 27th, 2021
For POET Technologies, the Pieces are Falling Into Place with Super Photonics
POET Technologies CEO Dr. Suresh Venkatesan says, “The burden of proof is always on the innovator, not the incumbent.” That understanding of its underdog status has driven much of the company’s decisions in recent years, Venkatesan said during a discussion focused on Super Photonics, POET’s recently launched joint venture with Xiamen Sanan Integrated Circuit. [http://www.sanan-ic.com/]
Dr. Venkatesan pointed to POET’s tactical deliberateness that has allowed it to present to the optoelectronics semiconductor industry a solution to overcome long-standing obstacles. The POET Optical Interposer™ is created to be a foundational platform for technology development and the optical engines that will be assembled and tested at Super Photonics are scheduled to be among the first products evolving from that platform.
Super Photonics is one of the pieces of a multinational operation that POET has steadfastly pulled together to serve what it anticipates will be a massive market. [https://poet-technologies.com/docs/presentations/2020%20POET%20Annual%20General%20Meeting_FINAL.pdf] The facilities include:
- Singapore: Optical interposer process and design, platform design, supply chain management
- Allentown, Pennsylvania: Optical engine design and product development, device design, supply chain and manufacturing engineering, product management and marketing, and customer technical support for Europe and North America
- Shenzhen, China: Optical engine design and validation, module reference design and applications engineering, customer technical support for China
- Xiamen, China: Optical engine assembly and testing through Super Photonics.
In a discussion with journalist Adrian Brijbassi, Dr. Venkatesan, POET President and General Manager Vivek Rajgarhia, and Vice-President of Product Marketing and Business Development Edward Cornejo discussed the state of progress at Super Photonics and the partnership with one of the world’s largest semiconductor manufacturers that is a crucial part of POET’s robust ecosystem.
Q: How does a giant company like Sanan Optoelectronics, from all the way across the world, hear about a small company like POET?
Vivek Rajgarhia: Like most things in the world, it starts through your network. A former colleague reached out after hearing about some of us joining POET, and said it looked pretty interesting. And at the same time Xiamen Sanan had been looking to expand its presence in the data center market. It already has a huge portion of the world’s LED market, and about seven years ago it saw growth opportunities in 5G and data centers. They created a subsidiary called Sanan IC, which is providing foundry services for GaAs [gallium arsenide] and GaN [gallium nitride] products, but also making power amplifiers that go into 5G base stations.
They have set up an InP [indium phosphide] supply chain and a world-class foundry for InP wafers that can support the growth in the data center market. They are well positioned in the 5G market, especially within the base station networks, but they still need to get to a level similar to their position in the LED market.
Sanan IC saw that POET has a platform that can take their InP-based devices and evolve them to make a deeper penetration into the cloud data center market and also 5G, because 5G is also using optical components. Once we showed them our technology we were able to build an agreement.
It was the first week of January 2020 when Suresh, Tom [Thomas Mika, POET’s CFO], and I visited Sanan IC in Xiamen. This was before the pandemic was known. We had the opportunity to see the scale and world-class manufacturing facilities and systems they had established. There were robotics and robots in use, and it was quite something to see. Suresh, who is a highly experienced foundry technologist, was really impressed with what they had.
So we saw there was an opportunity to set up a joint venture that could scale. We would be able to set up the integration and manufacturing of the optical engines based on our novel optical interposer platform and Sanan IC had the capability to invest. Sanan IC will also provide the indium-phosphide chips. Not all the devices will come from there but several critical devices, especially for 100G and 200G, will be supplied to Super Photonics, which is highly cost sensitive. And then we own our own factory jointly.
It was highly synergistic how all that fit together.
Q: I imagine the legal paperwork alone must have been immense to establish the partnership.
VR: Getting the term sheet and the LOI [letter of intent] took some time in order to determine the right structure for engaging and setting up the joint venture. It was less about negotiation and more about putting the right terms and structure that would provide for the longevity of this relationship as well as setting the foundation for growing the relationship from the start.
Q: I think POET has made it clear what Sanan IC does for you, in terms of providing mass production and industry validation. What else does POET do for Sanan beyond what you’ve just mentioned?
VR: POET brings the novel integration platform where we can integrate the devices that Sanan produces. Customers want an engine — it’s not a module that we are producing at Super Photonics, we don’t intend to make a module. The engine has a certain level of integration with the lasers and PDs [photodiodes] with the waveguides that the module customers can use. There is a big cost benefit and a big scaling benefit for our customers through our process.
Q: There’s the business practice of Knowing Your Customer’s Customer. How well do you know the needs of Sanan IC’s customers?
VR: On the optical side, we know them very well. I have been working with those customers, personally, for 25 years. I started working in the China market in 1996. I was one of the first people at Lucent Microelectronics, Optoelectronics products business, to start marketing its components there. More recently with MACOM, I started the optical business in China. So there is a lot of my network that exists there. Although there are many new companies that have sprung up and there is a next generation of talent in the workplaces, many of the leaders in the organizations are the same people I have dealt with for years.
Q: Suresh mentioned during the Needham Investors Conference Call on January 14 that a focus of 2021 is the build out of Super Photonics. What is the status of that build-out? Is a CEO in place?
VR: It’s going pretty well. The company Super Photonics is under the registration process. We have approval for the name and the final approval should come in a couple of weeks. However, we have already hired the candidate who will become the CEO and also the vice-president of R&D, and they have left their former companies and are with us. There is also our technical team that is being put in place and there is a CFO who is coming on board. With Sanan having a significant presence in the area, they are putting the support personnel into place.
We have identified the facility where we will start off and then Sanan is constructing a new building and they will lease part of it to Super Photonics. The needs of Super Photonics were kept in mind when the build-out of this new facility was conceived.
Q: All of that build out of the facility and the hiring of the personnel, does that impact the burn rate of POET Technologies?
VR: All the cash is from Sanan IC — all of the operations expenses as well as capital expenses.
So we design the product but there is a lot of work that needs to be done to take that product to mass production. What we call the NPI, or new product introduction, process and it includes the qualification phase and customer sampling phase. All of the costs for that work goes into the operational expense of the joint venture.
Q: On January 19, POET announced the opening of POET Optoelectronics Shenzhen Co. Ltd., and the appointment of the esteemed Dr. Jinyu Mo as Senior Vice President of Asia. [https://poet-technologies.com/news/2021-jan-19.html] Is the key purpose of this new center to help Super Photonics achieve its goals?
VR: So now that we have the technology platform done, it has to be turned into a product that can be validated at the system level, and when I say system level it’s how the customers would use it and how it would work once the electronics go around it. The team in Shenzhen will be completing the optical engine platform design at the system level. In addition, they will be working closely with the customers to make sure the application reference design and support is all there, and to ensure the applications of the products can be geared to the customers and the end customers.
And, yes, they will also be working closely with the joint venture. There is a high-speed train connection between Shenzhen and Xiamen that provides good access for our team at POET Optoelectronics to Super Photonics.
For our 100G, 200G, and 400G engines [for telecom and datacom applications] in the Greater China market, the work will go through the joint venture. For other markets, like North America, Europe, and elsewhere, especially for 400G and artificial intelligence, POET will be servicing customers in those jurisdictions.
Q: And a lot of that work for servicing global customers will be done at the Shenzhen facility, too?
There are different levels of design. One is the technology platform — the interposer, constructing and designing it, developing the processes of how we integrate different devices onto the interposer; most of that work is led by Suresh and the Singapore team.
Then it’s how you align the light to the external world of fiber, the optical design work. Allentown is a hub for that kind of engineering and our team there is focused on that effort.
Then you have to make it work at a system level: How the electronics get attached to it, how do you measure the performance to make it into a product that is validated for the end customer. A lot of that work is what the Shenzhen team will be doing for all of our customers.
Q: If a customer comes to you and says, ‘Ok, we’re convinced about what the optical engine can do. We want X number of units for our 100G or 200G transceivers,’ for example, can Super Photonics deliver?
We already are working with certain customers, and those customers realize that the actual production will come from Super Photonics. So it is a big confidence to those customers and our customers’ customers. They see there is a reliable, scalable supply chain, and that we are not a start-up company. We already have a foundation to attract the business. There is no risk that our customers see in terms of our ability to supply volumes.
Several high-speed devices are made at Sanan IC through their foundry. The interposer is made at SilTerra, our foundry partner in Malaysia. It makes the CMOS silicon wafer for us, based on our specs and our IP and know-how. Those wafers will be sold to Super Photonics as a component.
It is very important to emphasize the fact that POET will retain all of its IP and trade secrets related to the optical interposer platform.
Q: As you’re envisioning how things will unfold and thinking about where you're going to be over the course of this year, what does December 31 look like for Super Photonics? How much is it producing on a monthly basis?
VR: By the end of the year, we will be in mass production. At least with certain engines for 100G and 200G, and 400G probably will not be in production but getting to final samples that customers can design into their systems. In 2022, we will be going into mass production and revenue with 400G as well. And 2023 is when we expect triple-digit million-dollar revenues for Super Photonics.
Q: What is mass production defined as? If you’re doing 10,000 units a month would that be mass production?
VR: Yes, I would say 10,000 a month is mass production.
Edward Cornejo: Yes, for the kind of products we’re talking about, that’s a good number.
Q: And how many customers would those units be supplying?
VR: We will be starting off with a handful of customers. It’s not going to be 20 customers, but a few. There is a sales cycle that happens. We send samples to customers, they design the system to use our product, they do their qualifications, and then they go into production. And we align our processes to go along with those steps.
By the end of the year we would expect a handful of customers having us designed in and some using our product in production, and then in 2022 we expand and penetrate to secure other customers.
Q: At some point this year, the first mass produced optical engines will be completed by Super Photonics. What's that going to feel like?
Suresh Venkatesan: Getting a product out and running in production is a feat. We’ve gone through this tech dev to product deployment cycle multiple times in our respective careers. Honestly, what I’ve seen is people talk about the “a-ha” moment and the excitement of early innovations but a lot of value actually is in translating the research and development into mass production. I have seen many technologies just not make that transition. That would be a really big accomplishment once we have this rolling through and produced in high volumes.
EC: I’ve been through many cycles of new technologies evolving and there was a time in my career when I thought when 100G came out and it was mass produced that would be it. I would be done, I would be retired. I don’t believe that anymore. In fact I don’t even believe 1.6 terabits is the end game. Maybe it’s 6.4, maybe it’s 12.8. These things keep evolving and it’s changing.
The importance of what we’re doing at POET is it’s a hybrid integration platform. If you validate that through mass production, you can build anything out of that. You can expand everything and make all kinds of chiplets. And not just for data centers. You can go into edge computing, and that’s artificial intelligence, IoT [Internet of Things], cars, all kinds of stuff.
That’s what makes me excited. Once we validate that we can couple light — we can deal with reflections, we can deal with stresses, we can place the thermals and the electricals in the unit — then we can do whatever we want, and we’ll be doing it at a very low cost and at wafer scale.
Ever since I’ve been in optics, coupling light has been the issue. You solve that problem without having external lenses or isolators or additional parts, and that’s a huge deal. A lot of people have tried to do this, and we’re not doing it with just one device. We’re doing it with multiple active devices. It’s kind of like a little Lego busting, you can build whatever you want once it’s validated.
Q: You've been involved in developing groundbreaking technologies throughout your careers. Based on your experiences, how important is the optical engine to the industry?
VR: POET is really changing the way optoelectronics is being done. We are transforming the existing approach. So to see these optical engines being produced, 10,000 a month or 100,000, whatever it may be when we start, it’s not only a business success, I feel it is a success for the industry. It’s a milestone for the industry. We have been doing things in a particular way as an industry for a long time and now we are doing it differently, and it’s taking all of us to a higher level.
This approach that emerges with the optical interposer can proliferate into all photonics. It’s similar to traditional semiconductor packaging, which is a key. The missing piece of the puzzle in photonic integration has been with waveguides and how to use them to control the pathway of photons in a similar way that the flow of electrons is controlled in existing processes. We have achieved that control of photons and light with the optical interposer.
SV: That’s the nature of innovation especially in the U.S. and in Silicon Valley. Established companies will put a lot of investment into development and a certain way of doing things, and they will continue to bang their head at these problems with a lot of investment that they can afford. And then there is always going to be a small company that comes about and says, ‘Hey, let’s look at the problem in a slightly different way,’ and try to articulate solutions that might work, and then eventually that’s how technology evolves. The model now is more desegregated in where and how innovation occurs.
We are in that mold with our fabless model. We have seen with our experience collectively what issues exist in the industry and we have articulated potential solutions to that problem.
And, of course, there has to be peer review in this process. You may think you’ve got a great product, but that’s not one of those things you can decide for yourself. You need industry recognition.
We have been vetting this platform with many customers and potential partners, and there has always been interest and excitement. The question from them is can you actually do what you’re saying, because this sounds too good to be true.
The burden of proof is always on the innovator and not on the incumbent, especially when you’re trying to do something different than what people have done for many years. The burden of proof is always much, much higher, which is why it has taken us a fair bit of time to get to the point where we feel comfortable with the data we are generating and we feel comfortable with our ability to bring these solutions to market in the timeframes we’re talking about — getting to production by the end of the year.
The burden of proof will be on us to dislodge the incumbent, and that’s why it’s important for small companies like ourselves to have strategic partnerships.
Previous Articles:
Why POET Technologies Will Have a Positive Daily Impact On the World
[https://poet-technologies.com/blog/2020-dec-16.html]
Explaining POET’s Game-changing Optical Interposer Breakthrough
[https://poet-technologies.com/blog/2020-dec-15.html]
Note: Author Adrian Brijbassi has previously written about POET Technologies for the Globe & Mail and has since become an investor in the company.
< Back to Blog