Google Ventures Leads $20M Round for Transphorm to Battle “Hidden Tax” in Power Conversion

2/23/11Follow @wroush

There’s a hidden tax when you pay your electricity bill, says Umesh Mishra, a professor in the college of engineering at the University of California, Santa Barbara. It’s the energy lost when power is converted from the alternating current coming out of a wall socket to the direct current needed by computers, TVs, and other appliances, or from one voltage or frequency to another. By some measures, 11 to 12 percent of all electricity generated in the U.S. dissipates into the air as heat thanks to inefficient power conversion.

If we could get that power back, it would contribute more to the economy than solar, wind, and every other form alternative energy put together. It would make electric vehicles more economical, improve the output of rooftop solar modules, cap energy waste in buildings, and reduce the huge energy bills paid by Internet companies like Google that run vast data centers. (These companies are, in a sense, taxed doubly: they’re forced to buy even more electricity to run the cooling systems that carry away the heat leaked by their thousands of servers.)

The silicon-based transistors inside most power converter modules hit a physical limit years ago—there’s simply no way to make them any more efficient. But there’s a company in Goleta, CA, outside Santa Barbara, that’s spent the last three and a half years perfecting a next-generation power conversion technology based on gallium nitride, a semiconductor commonly used in light-emitted diodes (LEDs). Called Transphorm, the company was co-founded by Mishra and his former student Primit Parikh, and it has been working behind a thick, almost NSA-worthy wall of stealth—if you’d Googled it before today, you’d have found speculation that it’s working on “hydrogen generating technology,” among other misinformation.

But the company decloaked in a big way at a press conference today in Mountain View, CA, where Google Ventures, the search giant’s venture investing wing, revealed that it was the lead investor in a $20.2 million Series C financing round completed last year for Transphorm. Foundation Capital also participated in the Series C round, joining previous investors Kleiner Perkins Caufield & Byers and Lux Capital. The company has raised about $38 million in total.

Transphorm’s financing round was actually disclosed in an SEC filing in May 2010, but the documents escaped the notice of journalists; they show that Google Ventures partner Wesley Chan, Kleiner partner Randy Komisar, and Foundation partner Richard Redelfs are all board directors at the company.

It’s the first publicly disclosed energy investment for Google Ventures—but it’s not hard to understand why the fund would be interested, despite the green energy industry’s reputation for being capital-intensive and slow to yield returns. “One thing that attracted us to Transphorm is that we have an intimate undersanding of this problem, with a global network of data centers,” Bill Maris, managing partner at Google Ventures, said today. “Having an understanding of the problem made it a natural fit.”

“Google Ventures is highly tuned to the need for improved energy efficiency, especially when it applies to servers,” Mishra said in an advance interview with Xconomy on Tuesday. “When we made our presentation to Google Ventures, there was an appreciation of the scope and importance of the problem.”

Transphorm will sell its custom-designed power modules directly to power equipment manfucturers, who are expected to embed them in everything from consumer electronics devices to industrial motor drives. If the products win widespread adoption, it wouldn’t just benefit data center operators, says Mishra. “Imagine taking the whole West Coast off the grid—the impact would be huge. Transphorm’s impact could be that big, and saving this kind of energy is a huge business opportunity,” he said at the press conference. “It’s good for the planet and good for business.”

“It’s a real pleasure to finally share what this venture has been able to accomplish,” Komisar said at the event. Transphorm “has met or exceeded expectations at every juncture, which is really unusual.” Kleiner Perkins invested in the company, Komisar said, in part because of the scale of the problem—he said the energy savings from better power converters could be three times greater than that from other much-heralded technologies such as digital lighting. “The opportunity is to take 300 coal-fired plants off-grid, effectively, with more efficient power conversion modules,” said Komisar.

In power converter modules based on gallium nitride rather than silicon, waste heat is reduced by up to 90 percent, according to Peter Hébert, a board observer at Transphorm and a managing partner at Lux Capital, which invested in Transphorm’s Series A round. “This is absolutely transformative,” Hébert says. “One of the biggest and most important applications would be HVAC, building systems, elevators—buildings consume something like 40 percent of all electricity produced. Also, power supplies for servers and storage in data centers. And other applications ranging from inverters for electric vehicles to solar panels and wind. Every point where you are converting electricity.”

According to Mishra, the company is already supplying test converter modules to customers in the areas of computer servers, photovoltaic inverters, and motor drives for building systems, and expects to have products for sale in the next 12 to 16 months. “The last [market] to bear fruit will be hybrid cars and electric vehicles, because the design cycle in the automotive market is so long,” he says. The company plans to demonstrate its technologies two weeks from now at the Applied Power Electronics Conference in Forth Worth, TX, the power electronics industry’s biggest annual event.

If Transphorm can help makers of power supplies shave even a few percentage points off efficiency losses, it will be the second huge win in Mishra’s career. He’s already famous as one of the pioneer researchers in applications of gallium nitride for the LED lighting industry; Nitres, the Santa Barbara company he started in 1995 to commercialize that work, was acquired by Durham, NC-based LED manufacturer Cree in 2000 for $300 million, and is now operated as Cree’s Santa Barbara Technology Center.

Transphorm, like Nitres, has a core staff of researchers who all came from Mishra’s lab at UCSB. “We have a good blend of fresh graduates and people who have gone on to get industry experience, and they form the nucleus of this team,” Mishra says. “Arguably—no, I will just withdraw that—it is the best team in the world working on this problem.”

To understand the problem itself requires a bit of a detour into material science and electronics. The basic function of a power converter is to turn DC power into AC, AC into DC, or reduce the 110 volts of energy that come out of a standard U.S. wall socket to something appliances can handle. Decades ago, most converters were actually designed to shed excess energy in the form of heat. Later, switched-mode power supplies were developed that converted power much more efficiently by temporarily storing it in transistors, inductors, capacitors, and other electronic elements.

In today’s switched-mode power supplies, the transistors are made of silicon, which is easy to work with but has a built-in limitation: its narrow bandgap, meaning, loosely, the amount of energy required to get electrons moving. Explains Mishra, “The bandgap reflects the amount of voltage you can place across a certain thickness of silicon, and you can only sustain a certain amount of voltage because beyond that the small bandgap of silicon causes the material to break down”—in a literal puff of smoke. Gallium nitride has a much larger bandgap, three times that of silicon, which has an exponential effect on the amount of voltage transistors made of the material can hold. “The most conservative estimate is 10 times, and if you risk smoking and cracking you can go up to 200, but everyone agrees the benefits are an order of magnitude or more,” says Mishra.

Conveniently, gallium nitride is also a better conductor than silicon when a transistor is switched on, losing less power to resistance—and since it’s so much better at holding a voltage, less material is needed, so transistors can be thinner. (It’s gallium nitride’s wider bandgap, and its stability under high frequencies and high thermal loads, that also make it ideal for applications like LED lighting and the violet laser diodes in Blu-ray players.)

Ingenious power electronics manufacturers have taken silicon-based converters to the material’s absolute limits, Mishra says, but “you can’t change the bandgap—that’s a material constant.” But Mishra says it wasn’t until 2007 or so that he started to become aware of the scale of the efficiency crisis in power conversion. “We are wasting hundreds of terawatt-hours in energy in power conversion—the losses we are tolerating today are more than all of the electrical energy generated by renewables. The fact of the matter is, the need exists and therefore the market exists.”

But to satisfy that market, Transphorm had a lot of work to do. For one thing, gallium nitride isn’t available in nice, cheap wafers the way silicon is. The material has to be grown on a foreign substrate—and that always raises problems with strength and stability. Secondly, while gallium nitride performed great under constant electrical current in the lab, it wasn’t so hot at handling switched current—which was a problem, since it’s only by switching very fast that today’s power converters can handle high voltages. “There was what is called in field a problem of dynamics loss or dynamic resistance, when you switched the devices,” says Mishra. “So that had to be tackled.”

Finally, switching gallium nitride transistors very quickly at high amperage created a lot of electromagnetic noise—far more than FCC regulations would allow. “You can’t afford to have noise generated that actually interferes with everything else around you, so we made sure that the circuits and packages we developed were such that no additional electromagnetic interference was generated,” says Mishra.

In other words, building practical high-efficiency power converters wasn’t just a matter of replacing silicon with gallium nitride. Transphorm had to find engineers who could put gallium nitride power switches into circuits, ad who could put the circuits into modules that manufacturers can in turn build into their hardware. “To attack the problem you need a symphony orchestra, not a one-string band,” Mishra says. “You have to innovate in all aspects of this value chain, and our team has experts in all of these domains.”

As important as the efficiency gains from its converters could be, the 75-employee startup and its investors are equally excited about the material’s size advantages. Makers of solar modules, and hybrid and electric vehicles are all looking for components that will make their systems smaller and lighter. “This is about higher efficiency, smaller size, and lower EMI [electromagnetic interference],” Komisar said at today’s press conference. “This is not just a better-faster-cheaper enterprise. It is a brave new world enterprise.”

It’s hard to gauge the actual scale of the market opportunity Transphorm is pursuing. But in the data center market alone, the power savings from improved efficiency could vastly outweigh the cost of new gallium-nitride-based power converters, which Mishra says will be higher than traditional components. Electricity is the single largest cost at most data centers, exceeding the expense of servers themselves; by some estimates, burning fossil fuels to generate the power that data centers suck up puts more carbon dioxide into the atmosphere than the nation’s entire fleet of aircraft.

Colleagues say that Mishra is widely respected not just as an engineering whiz when it comes to gallium nitride, but as a nice guy and an expert entrepreneur. “Frankly, for us, when we first invested, it was based on the people,” says Lux’s Hébert. “They have made us look like geniuses for that strategy. Umesh and Primit are not only technically superior, but really as individuals they believed they could create a great company in this area. We have always been believers in energy efficiency, but a lot of what’s been accomplished around power conversion is a testament to them as individuals and their ability to recruit a fantastic team to demonstrate the technology and prove it out from a business perspective.”

Xconomy goes the extra mile to bring you in-depth startup coverage. Compare this story to:

Energy Efficiency Startup Transphorm Emerges From Stealth, Raises $20 Million From Google Ventures (Tech Crunch)
Google’s investment in Transphorm could be good news for electric cars (VentureBeat)
Transphorm Unveils Efficient Power Module, $38M from Kleiner, Google (GigaOm)
Transphorm Nabs $20 Million From Google Ventures, Kleiner (AllThingsDigital)
Google Ventures Backs Transphorm’s Efficient Power Conversion Technology (Fast Company)

Wade Roush is a contributing editor at Xconomy. Follow @wroush

By posting a comment, you agree to our terms and conditions.