Member News - 18 December, 2012

Industry Member News

18 December, 2012

Industry News

North American Passive Optics Estimated at $3 Billion

The passive optics segment of the photonics industry has annual revenues of about $14.2 billion, with about $3 billion going to North American companies, according to OIDA. Half of the total goes to about 20 companies, and the rest to hundreds of others. OIDA reported these findings this fall. More details will appear in the upcoming installment of its market report. 

OIDA chose to be inclusive with this number. Many in the North American industry concentrate on “precision,” or “high-precision” optics—products with a surface precision of sub-wavelengths down to Angstroms. The OIDA estimate includes that, but also much lower precision lenses, such as consumables like contact lenses, a great business if there ever was one. 

Many optics companies fabricate and coat materials, but much of the value in the industry comes from assembling more complex optical systems. For example, a lens for a photolithography station can sell for over $1 mllion each, and ASML sells a lot of stations in a good year. Even mobile phone camera modules are complex in their own way, and the overall volume in that business is huge. 

An interesting question is why there seem to be so many acquisitions in the optics segment, yet there seems to be as many companies as ever. Founders of mom-and-pop companies are cashing in, yet new companies perpetually form. Where’s it going? Is this the end game, or is the big change still to come?

Every mom-and-pop company seems to have its PhDs that know the business, making it hard to scale past a few million dollars in sales. Companies can operate in this semi-custom mode up to about $50 million in sales. Beyond that, a more globalized system has to be put in place. Yet, a lot of products will never expand beyond a few million in sales. The best that an acquiring company may be able to do is to keep the pieces in place as it diversifies. Its competitive advantage comes from juggling the complexity better than the competition.

A video of the OIDA presentation from this fall is available to OIDA members and more details will be documented in an upcoming OIDA market report, coming in 2013. For trouble accessing the October presentation, contact Brooke Hirsch.

Government News

OIDA & SPIE working together on data collection for the National Photonics Initiative

We reported in the October newsletter about the new National Research Council report from the Harnessing Light committee, called Optics and Photonics: Essential Technologies for Our Nation. The report called for a “national photonics initiative”, or NPI, to interpret and carry on the spirit of the report. With most of the recommendations in the report asking something from government, this means that the NPI would need to articulate and advocate for photonics to the necessary governing bodies.

Five societies have joined in an ad hoc NPI committee to start this initiative. The societies are: OSA, SPIE, IEEE Photonics Society, the American Physical Society (APS), and the Laser Institute of America (LIA). OSA, SPIE, and the IEEE Photonics Society have also committed funds to support certain efforts of the committee. Tom Baer, Executive Director of the Stanford Photonics Research Center, is chair of the committee.

Among other things, the NRC report and the NPI committee stressed the importance of obtaining key economic data regarding the U.S. photonics industry. The committee assigned the task of defining and collecting the data to a subcommittee consisting of Tom Hausken (OIDA/OSA), Steve Anderson (SPIE), and Dr. David Mowery (UC Berkeley economist, and a member of the Harnessing Light committee). The subcommittee will propose and oversee a contractor to conduct the actual work.

This data collection effort will be the first of its kind in the U.S., and will not be completed quickly. For more information, please contact Tom Hausken or Laura Kolton.

OIDA Upcoming Events

Why Scale-Out Data Centers are Hot—In More Ways Than One

Data centers are a hot topic, but do you know why? Data centers were called out in the new NRC report, Optics and Photonics. They were featured in a series this fall in the New York Times. And OIDA is dedicating another workshop to the future needs of “scale-out” data centers. In components, the industry is heading toward a Red Brick Wall—but how many in our industry know it? 

Start with the boundary conditions on your basic warehouse-scale data center. The data center operators—Google, Facebook, Apple—can’t expand data centers indefinitely, because it would drive up the cost, footprint, and power consumption. The largest data centers already each consume 50 to 100 MW—that’s megawatts—of power. That’s why they locate them near places like the Columbia River Gorge, to save on the transmission cost of electricity. 

Next imagine that the traffic into an operator’s data center increases as more people use applications more. This might mean greater need for storage and switching equipment, and more traffic inside the data center. Since the overall warehouse has limitations on growth, the inter- and intra-rack interconnects have to continue to improve, at the same or lower cost.

Doing more at lower cost—what’s new about that? Every photonic company hears that. Every day.

The problem is that modern data centers don’t “scale up” in the way that traditional data centers do. They “scale out.” Most of the traffic is “East-West”, which means that as the traffic into the data centers increases, the traffic from server to server inside ramps up dramatically. Every server is essentially connected to every other server, and that doesn’t scale well. The interconnects expand rapidly, and therefore so does the cost and power dissipation, which as we know, has a top limit. 

Therefore, it means rethinking the architectures. It may mean that an entire row of racks has to be designed as a unit. It may mean dumbing down the robustness of products, to lower the cost. Can the industry agree to new standards? Can it solve the scaling problem without them?

As if that’s not enough, this workshop will consider the market opportunity for the interconnect components. With the big data center operators making sweet profit margins—as much as 25% net profit, or even greater—there has to be something in it for the component suppliers, and their investors. OIDA has commissioned Ovum to take a crack at this, and tell us if it’s a big enough opportunity to put money into.

OIDA is co-hosting the workshop with the U.S. National Science Foundation, and NSF’s Center for Integrated Access Networks (CIAN), which is led by the University of Arizona. It will be held on March 17 at the same time as OFC 2013, in Anaheim, CA. For more information, contact Brooke Hirsch.

Happy Holidays to you and your loved ones! Thank you for your support of OIDA this past year and we look forward to seeing you in the New Year!