Profile - Gehan Amaratunga

Cambridge Semiconductor: how to spinout from university

By Nitin Dahad

If you've developed a great new technology in a university research laboratory, one professor who has successfully taken a research concept (USO1) and converted it into an externally funded, pre-product company (USO4) is Professor Gehan Amaratunga. Having successfully obtained funding and showed proof of concept, he has some very useful words of advice. Professor Amaratunga, 1966 Professor of engineering and head of electronics, power and energy conversion at the University of Cambridge, has this to say:

• Make sure you have a really good idea and understand the technology space very well
• Have a clear, quantifiable (meaning measurable, not theoretical) understanding of the advantages of the technology you are offering
• Remember that founders are not necessarily the best people to convert the ideas into a good commercial project

Gehan Amaratunga

This might sound like a truism, but one may be surprised by how many university spinouts fail to follow some very simple steps. Of course there is more complexity behind this simple advice. Gehan Amaratunga co-founded Cambridge Semiconductor, a fabless semiconductor company, with his colleague Dr Florin Udrea to commercialise what they considered to be breakthrough technology developed at Cambridge University. The company was spun out of Cambridge University in August 2000 with seed investment from the Cambridge University Challenge Fund (For non-UK readers, this is a special UK government-backed vc fund, specifically aimed at university spinouts. External fund managers manage the Challenge funds and investment decisions are made on a purely commercial basis).

Cambridge Semiconductor's technology was developed over the course of 15 years of research. It uses standard semiconductor manufacturing processes to integrate ultra-high-speed power switches and complex control circuits in a single chip. The company is developing a new generation of power integrated circuits with switching speeds claimed to be far in excess of products currently on the market.

"Our main motivation was to see a technology released to the market," says Amaratunga, who enjoys the academic world and will let people with more commercial, industrial experience assist in taking the project forward.

A background in research

After graduating in electrical engineering from Cardiff in 1975, Gehan Amaratunga obtained his Ph.D. from Cambridge in 1983, and subsequently held academic and research positions at Southampton University, the University of Liverpool and Stanford University in California. He and his co-founder have a track record of successful collaborations with industry partners and Amaratunga has also published over 300 papers and is cited as an inventor on 20 patents.

The Cambridge Semiconductor (CamSemi for short) technology evolved from research in 1986 in integrating high voltage devices with standard cmos. "It was product-led research and innovation, and we gained a lot of knowledge over the years. We first demonstrated a 600V power device with a 5V cmos capability in 1992," claims Amaratunga. "We worked with about seven major semiconductor companies, some of whom funded our research. So prior to founding our company, we had experience of transferring technology - one company, Dynex Semiconductor, has products which came directly from Dr. Udrea's research. This whole process gave us an important appreciation of what it is that would make an important step-change in the industry."

"Ten years ago, we would have had two options, if we wanted more funding, one was to spinout a company or develop a grant proposal and approach a semiconductor company [for funding]. But now we've become so much more aware of the possibilities of setting up a company to deliver a technology that would make a difference in the market, so we decided to form a company to commercialise the potential of our research."

Amaratunga believes this is the result of cultural change that has taken place in the UK. "I became aware of this when I took a sabbatical at Stanford in 1989, and saw what John Hennessy did with MIPS. (Professor John Hennessy initiated the MIPS reduced instruction set computing, or risc, project at Stanford in 1981, and played a key role in transferring this technology to industry in 1984-85, when he co-founded MIPS Computer Systems.)

"By 1999, the situation around Cambridge had changed, with an entire infrastructure of local talent, advisors and local investors." Amaratunga is careful not to stretch the Stanford analogy too much: "The real semiconductor strength is still not there in Cambridge from the fabrication point of view. But if we look at biotech startups, then Cambridge is ahead of Stanford in that respect."

Have the idea and a vision? Then DIY!

Amaratunga was clearly inspired by the Stanford experience. "When you have a really good idea, how do you take that forward? If you have the vision and have access to the capital, then why not do it yourself?" he says. It was this that triggered the formation of CamSemi in August 2000 by the two founders to explore power integrated semiconductors.

"It was quite timely, as the university had just set up its seed fund, called the Challenge fund," says Amaratunga. "The Challenge fund carried out its due diligence, and decided we were a good spinout opportunity. So we received a [Chilli S3] funding round of £250K in January 2001 to demonstrate the proof of concept. We had 12 months to do it, and we used a research foundry at the University of Southampton."

"We started presenting to venture capitalists in the final part of 2001, and the termsheet was agreed in February 2002 with 3i (they were convinced that we had an outstanding opportunity). By December 2002, we closed a [Chilli R1] funding round of £3.75M, but the road wasn't easy, as a matter of fact, it was very hard - not so much because of the mechanics of the process, but the uncertainty as to whether it would actually close, because of market conditions. The company also ran out of money early, but the university stepped in with a further convertible loan in September 2002, which allowed us to keep going, come to closure on the funding round, and deliver proof of concept."

Amaratunga believes one of the most important people is their interim CEO, John Lee. Through Odyssey Ventures, a Cambridge-based seed capital fund, Lee invests in early-stage high-technology companies where he can play the classic role of an angel, in investing, mentoring, building and developing the business. He is also a director of some of his investee companies, a director of the Cambridge Quantum Fund, sits on the Cambridge University Challenge Fund/Venture Capital Committee and is involved in the Prince of Wales Trust.

"Lee was not just a passive investor - but a real angel rather than just a private investor. He worked on a contingency basis until final fund closure, to get the deal done. The VC community had greater confidence because of his participation in the management."

Challenges along the way

The biggest challenge for Amaratunga has been the issue that has faced many startups in the last couple of years - to convincing the financial community that they should part with their money in yet another technology company, particularly in the current market downturn. "This was good for us because it played to our strengths. In many ways, in power electronics, the semiconductor area has been the Cinderella of the industry - but energy and power conversion products represent a major opportunity in the market, since almost every product requires some power integrated semiconductor."

"There is a huge requirement nowadays for using energy more efficiently and intelligently, particularly around the home in domestic consumer applications. Most consumer products involving electronics such as washing machines, home computers or televisions can be made much more efficient," explains Amaratunga. "We have invented several new methods for gaining significant performance advantages from the silicon chips used for power management and radio frequency generation. This will enable us to reduce the energy consumption of electrical appliances in a cost effective manner, simply by making them draw power optimally at all times."

CamSemi's Chilli R1 funding round was a syndication of various players. They were able to convince 3i, who committed £1.7M, Scottish Equity Partners and TTP Ventures, who invested £1.3M and £650K respectively, and also Odyssey Ventures and Cambridge University, bringing in a total of £3.75M in December 2002. "Despite the fact that we had to go through hoops to get the money, the vc community has been very responsible in their approach [given the market climate]," adds Amaratunga.

He believes the fact that the technology was coming from Cambridge added some cachet and weight to their company as an investment opportunity. "The good thing for us was that we were backed by Cambridge University; this provided a certain 'ballast' to our position. The university played a role in proactively helping us close the deal."

CamSemi aims to start sampling products in Q4 2004, deliver revenue at the end of 2005, and reach profitability in 2006/2007. Amaratunga says that the venture capital community recognises that the company may require a total round of £50M over this timeframe before they hit cash positive. "But they have come in with a long-term view." As for the Professor, he will stay where his heart is - in research, and as an advisor to other people who want to transfer their technology to industry, and he has been busy incubating some more spinouts already.

Comments on this story? Send an e-mail to Editor@TheChilli.com

© Chilli Publishing Ltd 2003

14JUNE2003