At the EPiCentre of the GaN Revolution

Author:
Ally Winning, European Editor, PSD

Date
02/18/2026

 PDF
Alex Lidow from EPC talks to PSD about the company’s latest attempts to bring the benefits of GaN to low-voltage applications

EPC

Alex Lidow, CEO of EPC

­  

When wide-bandgap semiconductors entered the market, the areas in which they operated were pretty well defined. SiC would challenge silicon at voltages above 600V, and GaN would compete with silicon in applications from around 100V to 600V. Silicon would remain dominant at voltages under 100V. Over the last few years, all of those assumptions have changed quite dramatically. We now see GaN products designed to compete with SiC and silicon at higher voltages, while SiC products have been developed to leverage their superior ruggedness at lower voltages. Perhaps the most interesting area of all is under 100V, where GaN has entered the market, and according to Alex Lidow, CEO of EPC, has the potential to replace silicon completely. EPC has already gone well beyond the supposed 100V threshold, and recently, the company has made several major announcements about how it is pushing those boundaries further.

 

The first announcement was a licensing agreement with Renesas, which will see the Japanese multinational gain access to EPC’s low-voltage eGaN technology and its established supply-chain ecosystem. The new deal is intended to assist in the adoption of high-performance GaN solutions across a broad range of markets. EPC and Renesas will collaborate over the next year to establish internal wafer fabrication capabilities for these products. In addition, Renesas will second-source several of EPC’s GaN devices, enhancing supply-chain resilience for customers.

 

Lidow expands by saying, “EPC is a relatively small company. We have just a handful of salespeople and applications engineers around the world. Renesas has hundreds of them, and the company has exceptional ICs that work well with our devices. That makes Renesas very well positioned to expand the global reach of our technology and to expand its ecosystem. The company has tremendous technical capabilities, and that’s a large benefit to us. Customers often ask for a second source, and to know that they have multiple sources and a common footprint is a large convenience. The agreement also opens up new markets - Japan is an area that we haven’t even started to penetrate, the same with China. Renesas has huge resources in terms of promotion and sales in those regions.”

 

The second recent major piece of news from EPC is that the first of its seventh-generation (Gen 7) eGaN family has entered volume production. The EPC2366 is a 40V device that extends EPC’s reach into even lower voltages. Also in the works are 25V and 15V transistors from the same family, which are currently sampling. Lidow claims that the Gen 7 platform can deliver up to 3 times better performance than equivalent silicon MOSFETs, simultaneously cutting conduction and switching losses while improving thermal performance. The new device has a typical RDS(on) of 0.84 mΩ and an RDS(on) × QG figure of merit (FoM) < 12.6mΩ *nC. It supports drain-to-source voltages up to 40V and transient voltages up to 48V, with continuous drain currents up to 88 A and pulsed currents of 360A. It comes in a compact 3.3 × 2.6mm PQFN package, that has a thermal resistance from the junction to the case of 0.6°C/W.”

 

On that thermal efficiency figure, Lidow explains its importance by saying, “that thermal efficiency is achieved by having the back surface of the semiconductor exposed at the top of the device. This is not possible for MOSFETs, which are cooled by a heatsink that's attached through solder to the back surface of the device, adding thermal resistance. Our devices are extraordinary thermally efficient for their size, and that makes a big difference when you're concentrating power into such a tiny device. The first thing is to not generate heat at all and GaN is much more efficient. After that, get any heat away from the device as quickly as possible, and that comes from the junction case resistance, which is much lower than than the equivalent double sided QFN package used by competing MOSFETs.”

 

Talking about the new platform, Lidow said, “Gen 7 is a significant improvement over Gen 6. It is crossing the 40V red line that MOSFETs have had for a long time. Rival companies say silicon is best at 40V and below, but, our 40V device offers significant performance improvements over the best of the silicon 40V devices out there, and that's just too compelling for engineers. There are three indicators for evaluating the performance of a device. One of them is the die size needed to achieve a certain on-resistance, and our die size is two and a half times smaller for the same on-resistance than the best silicon transistor. Input capacitance is an indication of how fast the device can switch, and of the high frequency losses in a hard switched circuit - our input capacitance is a third of that same rival. The third measure is output capacitance, which shows how well the device works in a resonance circuit, and the ratio of output capacitance between our GaN and the best performing MOSFET is five to one.”

 

The applications targetted by the 40V EPC2366 are the highest performance areas of the market, such as those found in data centres and computers and servers. Lidow estimates that the market for those applications is around twice as large as the market served by the company’s current 100V solutions. Rival devices for those same use cases are not simple silicon designs. They use complex processes for fabrication, which makes them more expensive. The cost and complexity of GaN designs were reasons that some critics claimed that GaN was unsuitable for use at lower voltages, but according to Lidow, that is no longer the case. He said, “the best low voltage silicon technology is very advanced, using techniques like deep-trench buried emitters. It's not cheap to make. In terms of cost comparisons, we have more die in a wafer, and our wafers aren't as complex. So cost is no longer a factor. As for design complexity, when you're running at 10MHz, parasitic inductances do matter a lot, but we're finding that new generations of engineers remember all the lessons that they have learned in college. They know circuit design. That means that we now have a larger and larger body of very skilled GaN designers out there.”

 

https://epc-co.com/epc/

RELATED