At Strategies in Light last week, your humble editor sat down with Eric Kim, President & CEO, Mike Krames, Chief Technology Officer of Soraa. Below is our conversation.
ER: Shuji shared some of the history of the company. How much VC have you raised so far?
Soraa: Over $100M since 2008.
ER: So with $100M from Kholsa and others, why start with the MR-16?
Soraa: The MR-16 is a great entry point. We have a very focused and purposeful strategy which we are implementing and the MR-16 fits perfect. This product is specified by lighting designers—which is important to us. It is used in hospitality, restaurants and light-critical merchandising areas.
Because of its very small size, it is a natural fit for GaN on GaN—our core competency. If we can successfully build the MR-16, we can build anything! It is one of the most challenging applications for LED and our GaN on GaN gives us an unfair advantage.
Also, it’s a huge market. There will be 550 million MR-16’s shipped this year. LED adoption is low despite the energy efficiency of LED because much of the LED MR-16s in the market are substantially below the performance and the light quality levels of halogen.
ER: You speak about GaN on GaN and I see it all over your literature and your booth. I have no idea what that means.
Soraa: When Gallium Nitride (GaN), the light-emitting material, is deposited on common substrates, like sapphire, the crystal lattice mismatch between the dissimilar materials results in a high level of defects in the GaN. In addition, the deposition process on sapphire requires complicated transition layers before the GaN active layers can be deposited, resulting in long growth cycles and wasteful use of materials and equipment. In addition, Sapphire is thermally insulating and electrically insulating—requiring a complicated chip architecture and a very complex wafer fabrication process. With GaN on GaN, which is short for Gallium Nitride on Gallium Nitride substrates, the wafer fabrication process is quite simple as there is a crystal lattice match at the atomic level; no transition layers are needed. In addition, the thermal expansion coefficients match and there is no bow in the wafer. Finally, GaN substrates are optically transparent and highly conductive. This allows a device architecture wherein the power (and the heat) levels are uniform throughout the device and providing for much higher power density operation than for conventional LEDs. There are no hotspots or thermal gradients created.
With GaN on GaN, we push performance to unprecedented levels. Do you ever see a traditional MR16 in 75 watt? No. Why? There is too much heat in that small 2 inch area. Because of our high power density operation, we see a path to make a 75-watt equivalent. You can’t do that with traditional halogen, or other LED technologies, including GaN on Silicon. GaN on GaN allows us many competitive advantages.
ER: Bridgelux and Toshiba are promoting GaN on silicon. Isn’t that better?
Soraa: We believe GaN on silicon is a dead end. There are many issues to solve especially removing the silicon. It requires a thin film process, which is complicated. The eptiaxy is extremely complex. As Shuji Nakamura noted in his talk, one cannot get to the highest light extraction efficiency in thin film LEDs. With its limited performance and complicated structure, the power density cannot be cranked up. The stresses are far worse with GaN on silicon. These are some fundamental reasons why GaN on silicon is severely limited in its comparative performance advantage and in its long-term reliability in the real world operating conditions. GaN on GaN has 1000 times less defects compared to this technology.
ER: So why do companies pursue GaN on silicon?
Soraa: Many of those companies have long histories of silicon (semiconductor) and have huge fabrication equipment already built. These old machines are sitting around with unused capacity. There is also a tendency for the semiconductor guys to think LEDs are easy to make. They don’t really understand the challenges of producing high quality LEDs required for the lighting market.
ER: Does this glut of equipment give them an advantage over Soraa?
Soraa: Not really. CMOS fabs do not include the equipment necessary for processing GaN. There is also a glut of Blu-Ray out there, which is perfect for GaN on GaN, since we can use the substrates no longer in demand by the Blu-Ray laser business.
And don’t lose sight of the fact that the reason our competitors like the GaN on Silicon is they believe it allows for a low cost product. Focusing on efficiency and low cost alone does not ensure a broad base of support. The proof is CFL which is relatively efficient and cost effective. Yet even with the huge government subsidies and rebates, penetration of CFL is only about 15%–and it’s been around since the 1980s. Quality of light is as, or more, important than efficiencies and cost.
ER: Shuji Nakamura told me a little about the early days. Tell me about your funding.
Soraa: We have raised over $100M in three rounds and have put together a very strong team of about 250 employees, mostly in the US. We have over 100 years in engineering experience in LEDs.
ER: Will there be a fourth round?
Soraa: We’re not going to get into that level of detail with you today, but I will say that we are in a very strong position in 2013 and we won’t be raising any money this year.
ER: What are your market channels?
Soraa: We focus on the commercial market. We pull the business through the specifiers and lighting designers and push it through electrical distributors such as Weidenbach Brown, Regency Lighting, Villa Lighting, and NEDCO Supply, to name a few.
ER: Will you sell your MR-16 in the DIY Channel?
Soraa: Eventually, yes. But today, the commercial market works very well as we have a compelling ROI and the level of performance and the light quality demanded by the users.
ER: What will the end user pay, such as a furniture store that purchases 100 lamps?
Soraa: Since we sell to distribution channels on volume based pricing, the distributors set the end user prices based on their business situations. When you check online today, you can see Soraa products being sold in the $40 to $60 range.
ER: Will the lamp work on existing MR16 low voltage transformers?
Soraa: Yes, we have done extensive testing with new and existing transformers and dimmers. And you can also view the interoperability of our lamps at the Lighting PlugFest booth that we sponsored here at the show.
ER: Dimmable?
Soraa: Yes.
ER: Where did the Soraa name come from?
Soraa: The Name Soraa means sky in Japanese.