Business Forum: Silicon photonics startup finds path to success - an interview with John Bowers

Laser Focus World editor-in-chief Conard Holton asked me to share the experience in incubating Aurrion, which led to an acquisition by Juniper Networks. Given I was an investor and a director, I felt it more appropriate for me to interview Professor John E. Bowers, who was Chairman of the Board and co-founded the company with his student Alex Fang. John holds the Fred Kavli Chair in Nanotechnology at UC Santa Barbara. He is the recipient of numerous prestigious awards, and his bio can be found at http://optoelectronics.ece.ucsb.edu/profile/john-bowers.

Milton Chang: What was the initial business "idea" when you and Alex founded the company?

John Bowers: We believed that silicon photonics would revolutionize optoelectronics and that integrating the laser with the photonic integrated circuit (PIC) was essential to realizing the high-volume and low-cost capability of silicon photonics. We didn't know which product to focus on, so rather than choose one direction, we initially conducted contract research for other companies to demonstrate silicon PICs, and, consequently, we were profitable for the first few years. We worked in the medical, data storage, military, and datacom markets before deciding to focus on data centers.

MC: Please say a few words about what gets you the "high-volume, low-cost capability" for the benefit of some of our readers.

JB: The electronics industry has developed a very high-volume, very low-cost (per element or per area) manufacturing capability that includes fabrication, packaging, and testing. Our goal was to fit PIC fabrication into that existing framework with minimal changes. Now, it is relatively easy to transition photonics to high volume. Most foundries would like to run 50-wafer lots of 200- or 300-mm-diameter wafers, which translates into millions of die per lot, depending on the die area. In our case, we wanted to integrate lasers and amplifiers on the die, which allows the ability to self-test with standard automated test stations and to package with modern packaging, all of which require 200- or 300-mm-diameter wafers.

MC: Where did you get the initial seed capital?

JB: We invested the initial seed capital, and Alex ran a very lean company that benefited from several government grants, research funding mostly from DARPA and DOD, and development funding from subcontracts from aerospace companies. Once we focused on getting a particular product out, we needed serious funding, and approached you and a few other successful angel investors and strategic corporate investors. We felt it important to get investors who understood the pitfalls and could help navigate around the common mistakes that many startups make.

MC: How important were grants to Aurrion?

JB: The benefits of DARPA grants were huge. They required large improvements over existing technology and were the perfect match to demonstrate a PIC capability well beyond anything available anywhere. One example was a DARPA EPHI (Electronic Photonic Heterogeneous Integration) contract that demonstrated widely tunable, very-narrow-linewidth lasers integrated with modulators and photodetectors. The modulators and photodetectors had to be better than the state of the art, as well as integrated with the laser, which required developing a bonding process that allows three to five epi-materials integrated into every die. Fortunately, we had great collaborators like Joe Campbell at UVA, who transferred his high-speed, high-power photodetector process onto silicon.

MC: What were the hurdles initially?

JB: Hiring great people is the most important task and probably the most difficult. The early hires brought good ideas, great experience, and key insights into good and bad product directions that shaped the company. Building a solid Board of Directors was also key to formulating the company's direction and avoiding novice mistakes.

MC: Can you highlight a few critical decisions?

JB: Greg Fish and Alex Fang focused on improving the laser performance, yield, and reliability, which led to a new structure that performs better than what Alex had developed at UCSB. They spent the time to patent their innovations and were continually working to improve the performance.

MC: To give our readers a feel for the level of effort on IP, how many patents were filed and granted?

JB: Aurrion (which was recently acquired by Juniper Networks) holds over 70 patents that cover everything from the laser design to packaging and automated self-testing, and even optical networks.

MC: What do you feel is your competitive advantage when there are silicon photonics startups all around the world?

JB: I believe Aurrion has one of the best ways to integrate lasers with silicon PICs. For products that require only a single laser, that advantage is significant, but for WDM applications, that is a real differentiator. I believe this technology is the ideal approach for mid-board optics, or transceivers within the MCM or even on the electronic chip itself.

MC: I actually think it is your technical support and the team's ability to address performance holistically, from integrating the laser on the chip to packaging and the ability to meet milestones, and also the supply-chain management ability.

JB: Yes, we are only just beginning to see the cost and performance advantages. Automated packaging and automated testing are going to be huge differentiators. If the laser is integrated, then complete self-tests and diagnostics are available on the wafer and throughout the packaging process. A lot of key designers made many innovations, and captured them in a solid, detailed photonic design kit (PDK) that let us innovate rapidly and cost-effectively.

MC: Given a high percentage of Aurrion employees were your former students, how do you train your students to be so good at covering both science and production?

JB: We certainly benefited from hiring brilliant students from many professors at UCSB, Stanford, and MIT, and blending them with experienced engineers. We consciously tried to hire a range of skills and disciplines from the best companies. Alex is very good at gauging talent and at quickly understanding what motivates a person.

MC: I applaud your activities in promoting entrepreneurship at UCSB. How would you advise students who are interested, but do not spend serious time learning about business?

JB: I tend to discourage students from starting a company. It is good to gain experience first. Further, ideas are plentiful and it is easy to start a company, but it is harder to find products that are really needed in the marketplace and have some competitive advantage over other solutions. Once you hire people and once you take people's money, then you have a serious responsibility to see the project through and to make the venture successful. So, be honest with yourself up front and reject most possibilities, and then choose one and run fast and hard.

MC: Can you summarize, in retrospect, what made Aurrion successful?

JB: It is all about getting great people, from the CEO and management team to the employees to the Board of Directors. A great CEO can motivate them to work together to rapidly accomplish what may have seemed impossible.