“Whoever gets there first and scales efficiently will get access to a huge bucketful of business. There won’t be much profit margin. But there will be enormous volume.”
“All this puts pressure on the global photonics community to come up with different approaches. “
You're active on both sides of the Atlantic. How should European companies react to the news a few weeks back that over US$ 600 Million is to be invested in the new Photonics Manufacturing Institute? It is being established at the Research Foundation for The State University of New York in Rochester. Is it direct commercial competition or just a logical extension of capacity to kick-start global mass-production?
“There were three finalist bids to host this new institute. But the goal in each case was to generate interest from industry. My role was to help with the California submission. USC was the lead university, which also included proposals from UCLA, UC San Diego, UC Berkeley, Arizona State University and Ohio State University.”
“There were also comprehensive submissions from Florida and the State of New York. It was an open secret that our California plan was a horizontal open-source approach, and the other two were vertically integrated plans.”
“For some people, vertical integration has some core benefits. One is control. The more activities you carry out in the manufacturing and distribution process, the more control you have over the entire flow of goods until they reach the end customer. On the other hand, it requires your company and its leadership to have expertise in multiple distribution channel activities. The roles of designer, foundry and chip packager are distinct. A manufacturer wanting to distribute its goods directly to customers must not only have production strengths, but also the ability to market and efficiently distribute those goods. “
As the project is managed by the US Department of Defense for the next five years, does that have any effect?
“It has some effect, of course. I believe there are currently 7 of these institutes of manufacturing being set up around the country. The National Network for Manufacturing Innovation (NNMI) intends to build a manufacturing research infrastructure so U.S. industry and academia collaborate to solve industry-relevant problems. They want to create, showcase and deploy new capabilities and manufacturing processes.”
“The initiative is driven by the Obama Administration who clearly want to make the US more competitive. It’s about bringing back jobs for US workers in USA based companies or foreign companies with US employment facilities. The White House office of science and technology then hired the US Department of Defense to allocate the money”.
“Of course, the DOD has its own interests, asking what can be developed to help their particular interests. But, in the end, the US government doesn't really mind if the end user is a defense department or a US consumer. So we had to put a proposal together that caters for a bit of both. In particular, one that meets the low volume expectations from the DOD, while providing a platform for high volume in the data-centre segment. Export controls on technology may come up because of the Defense involvements, but it wasn't seen as the major impediment.”
“Our proposal had four main areas we thought needed development. The first was fibre optic transceivers for data centres. The second one was RF photonics, which has not yet been integrated yet. The third topic was Integrated photonics and sensing. That will benefit both the DOD as well as consumers. We chose "packaging" as the fourth application. We still feel that this is important to opening up this key technology to many other applications. You may be able to put a few devices on a chip, but how do you connect it all up and package it at a cost level that commercial customers will pay for? Defense is used to completely different specs and pricing levels.”
“As someone who is active on both sides of the Atlantic, it is clear that the investments in integrated photonics are starting to show results. If you look at where the investment has been made in the last 15-20 years, the answer is mostly Europe and little bit in Japan.
The Brainport Eindhoven region has certainly been a driving force in bringing teams and initiatives together for integrated photonics, both locally and across the continent.
So is the announcement in Rochester a good development from the European perspective? I think it is, because it reinforces the fact that photonics is here to stay, especially as an integrated platform.
“I believe the large customers are more concerned about the price than what technology is inside the box. So, although a few hundred million dollars sounds like a lot for the new US Photonics initiative, in fact it is simply an early stage foot in the door. They want to see similar pilot production lines develop like in the European Horizon 2020 program. It’s true that latest investment is probably an order of magnitude bigger than what has happened so far in Europe.”
“But to really scale up this industry, you need tens of billions of dollars. So, Europe can still maintain its leadership role. The open question is whether Europe can get the IBM's, Cisco’s and Intel’s of this world to put in the billions. I think it is a challenge worth accepting. Europe has taught itself to collaborate efficiently, it has direction, a clear set of goals and lots of small innovative companies.”
“If you think about the US, the winning teams will be using a vertically integrated model. The Europeans, like Jeppix and a few others, are all open collaborative, horizontal models. I see advantages to this approach. You can negotiate various pieces of the pie that don't cost you a fortune. It’s not cheap, but neither is it outrageously expensive to develop new chip designs. It means an expanding market for smaller companies like Smart photonics, Effect photonics, VLC photonics and companies like that.”
“That was also what drove the California photonics proposal, an open source approach using a silicon brokering business called MOSIS.”
“The US government has decided to fund a vertical model. There is still a lot we don't know, but I'm sure big silicon fabs will benefit. But what about the disruptive “little guys”? Everything is done in one place, you have to go to a particular foundry. It might be better for the DOD as a way of keeping trade secrets. But I think some US startups and SME's may well benefit by looking to collaborate with partners in Europe, using its open source model.”
What do you hope the audience will take away from your presentation on September 23rd?
“An Integrated Circuit is simply a bunch of CMOS transistors on silicon, everyone knows that. I believe, in the near future, we’re unlikely to displace the incumbent. But it is going to be symbiotic with a platform that has photonics in it. Basically, you are going to have electronics closely integrated with photonics. There are various approaches to how that is going to happen like silicon photonics, indium phosphide, gallium arsenide, optical polymers etc. But I'm sure we'll figure out the technologies that will be adequately scalable for the markets soon”.
“On the 23rd September, I will show the roadmaps we've been working on. They make the commercial goals and challenges ahead much clearer. Integration is going to be essential to get the costs down and keep customers happy!”
Michael Lebby is the opening keynote speaker at the first Photonics Integration Conference taking place in Eindhoven on September 23rd. He divides his time between a role as CEO of Ottawa-based OneChip Photonics, a family base in San Francisco, and being a professor and chair of optoelectronics at Glyndwr University in Wrexham, Wales. UK. They have a world-class optics program connected with the European Extremely Large Telescope Project being built in Northern Chile. Michael is helping to build an active integrated photonics program, as well as raise the profile of this Welsh University. Michael recently joined Lightwave Logic (LWLG) as a board director to help guide their optical polymer technology in photonics applications.