Carving out a niche in the RF market

In every walk of life the big players go after the big bucks. This means that in the GaAs industry, the likes of RFMD and Skyworks concentrate their efforts on winning sales to the leading handset makers. Although these massive chipmakers will also go after smaller markets, these secondary targets will never be a top priority for them, and completely satisfying customers operating in these areas is not a chief concern. Consequently, these smaller markets offer an opportunity for start-ups, who have virtually no chance of success in the cutthroat handset business. Competition is far less fierce in these lower revenue markets, which are served by fewer firms, and by fully focusing on a niche a start-up can give a level of customer care that would be difficult to match in a big outfit that has to fight wars on many fronts. One start-up that’s identified lucrative niche markets within the GaAs sector and is now winning significant sales is BeRex. This fabless firm, which was founded in 2004 in Korea as BeRex Corporation, initially targeted a mobile infrastructure market with GaAs HBTs. And a few years’ later – after the leaders of the company identified a second opportunity, this time for GaAs pHEMT products in various military and commercial applications – it set-up a subsidiary, BeRex Inc., in San Jose, CA. The driving force behind both these successful ventures is the company’s CEO Nahm-Wook Lee. After graduating in economics, he worked at Samsung’s offices for many years, where he rose through the ranks to become president of Samsung Information Systems in the US. He then embarked on a new career, working as a consultant to local start-up companies. During that time a friend of his suggested that the infrastructure market offered a great opportunity to makers of GaAs chips. Back then, wireless communication was migrating from CDMA to W-CDMA. “Where there is change, there is opportunity,” reasons Lee. “I carried out a market survey, looked at the business opportunity, and asked a friend: Could you find me a semiconductor engineer who could design a product?” In late 2003, his friend introduced Lee to Alex Yoo, an electrical engineer with a PhD from Oregon State University. BeRex pHEMT chips are primarily used in military and commercial microwave amplifiers This duo teamed up with Young Moon Kim, who has a friendship with Lee that dates back to the 1960s, and the three of them started developing a gain block amplifier for the wireless infrastructure market. This selffunded venture initially made rapid progress – but then the team came across a stumbling block. When they started talking to potential customers prior to production of their device, they discovered that many of these firms were starting to focus their efforts towards Korea. Although Lee’s early years were spent in Korea – he went to Seoul National University – by then he had put down roots in California. Yoo, too, had a family in Silicon Valley, so relocating would involve some hardship. But moving to Korea made a lot of sense from a business perspective, so the founders decided to head to Seoul, forming BeRex Corporation in 2004. A year later the company started delivering its first product, and thanks to rapidly increasing sales, it reached profitability in 2007. Since then, the company has always been in the black, with revenue growing every year. What’s more, it has no debt. The company’s success can be put down to a variety of factors: It launched a strong product in a receptive market at the right time; its rivals were complacent and failing to focus on the needs of the customer; and it adopted a fabless approach, working with local foundries. Arguably, it is the latter factor that has contributed most to the success. On paper, outsourcing growth is more expensive than carrying it out in-house (with the benefits greatest in high-volume markets, such as the production of HBT power amplifiers for mobile phones). But in practice, in the low-volume markets that BeRex serves, the fab is a not an asset – it’s a burden. Not everyone appreciated that back in 2004, but more and more people are coming round to that way of thinking. An outsourcing model One of the biggest benefits of the fabless approach, according to David Snook, Sales and marketing Manager for BeRex Inc., is that it allows a company to tap into the capacity of a foundry, without having to incur its maintenance costs. What’s more, he says that a fabless firm does not have to worry about ensuring a sufficient wafer throughput to maintain product quality and reliability. “To do that, you have to have a lot of product coming out of the fab.” In Snook’s opinion, BeRex’s fabless approach also allows it to focus on putting together and maintaining a strong team that excels in designing chips, ensuring their quality, and fostering great customer relationships. In contrast, if a fab is on-site, its running can take precedent over everything, while keeping its technology competitive can require further investment. A big concern facing any start-up operating in a niche market is that its orders are too small for fabs to be interested in its business. “Each fab has its own minimum order requirements,” says Snook, who reveals that this could be several wafers, or it could be a dozen of them. “These are not seriously limiting factors, but they are certainly factors that we take into account as we make business decisions.”  Although individual orders may be low, these mount up. And this means that fabs are willing to deal with smaller, loyal customers, if both sides establish a good working relationship. “Are you honest? Do they like to deal with you? Qualities like that don’t appear on any spreadsheet, but they are very critical to that relationship,” says Snook, “Our relationships with our fabs are very, very good – they know who they are dealing with personally. They are dealing with the same people all the time.” This strong relationship helps BeRex’s to offer rapid turnaround times. While some firms quote six weeks, eight weeks, or even 16 weeks, BeRex promises to ship a production product within just two weeks. “I don’t think we’ve broken that promise to our customers in almost eight years of operation,” says Snook, who reveals that the money saved from fabless operations is re-deployed on engineering resources and building up an inventory. Back in the US Today BeRex generates most of its income from sales of infrastructure products, with revenue from devices built for microwave applications not coming far behind. Lee became aware the microwave market through a business friend in the US, who suggested that he should take a look at the RF and microwave amplifier market. “When I looked at the potential of that market, I thought we could do better than any of the current vendors.” To sell products to the US military, BeRex Corporation founded an independent company in Silicon Valley, BeRex Inc.. By taking this step, the organization avoids issues related to imports, exports and the sharing of information. “We need to keep a firewall between an offshore company and a US company,” explains Snook. The pHEMTs that BeRex Inc. has been producing,which can be used in military and commercial microwave applications, operate at up to 40 GHz and have a gate width of 0.2 mm to 2.4 mm, enabling output powers of 0.25 W to 3 W. Transistors with a smaller gate deliver less power, but produce more gain “We can operate as a subsidiary, as long as we have that firewall – but we can’t have that firewall and be the same company.” This west-coast outfit focuses on supplying the military and commercial markets with pHEMTs, which are designed in the US and built in native fabs. In addition, it acts as a distributor for products made by BeRex Corporation, such as GaAs HBTs, SiGe gain blocks, switches, dividers and amplifiers. The pHEMTs that BeRex Inc. has been producing, which can be used in military and commercial microwave applications, operate at up to 40 GHz and have a gate width of 0.2 mm to 2.4 mm, enabling output powers of 0.25 W to 3 W. Transistors with a smaller gate deliver less power, but produce more gain, so they are often used in the first stages of amplifiers employing up to five stages. BeRex Inc. started business in 2008, the nadir of the current economic storm. However, it has not struggled in this climate, due to significant sales to the military. “The military never disappoints,” says Snook, “and seems to fill the void in those years.” Although the last few months he has seen a slight ‘softening’ of this market, commercial markets are on the way up, particularly in the US. “We’ve not seen the same level of expansion in Europe, but I expect it to be there.” The US outfit has several products in the pipeline, including some GaN devices and a line of GaAs MESFET chips. The latter products have been developed for microwave customers in search of a level of linearity that is not possible with a pHEMT process. BeRex, Inc. is headquartered in San Jose,CA, at the heart of the Silicon Valley “Some people need the better OIP3 that the MESFET provides over the pHEMT process,” says Snook. “We can select the process. We can go out and pick the best. Not many people are doing MESFETs anymore, but this niche market still needs it.” BeRex GaAs HBT parts, such as the BG18C, target wireless infrastructure like base stations and repeaters Launching MESFETs forms part of the strategy for BeRex Inc.: To be a ‘one-stop-shop’ for high-quality pHEMT, MESFET and HBT bare-die for military and commercial applications. If it executes on this front – while it’s parent company continues to enjoy success in the infrastructure markets in Korea, and more recently China – it will provide further proof that it is possible for start-ups carve out a profitable niche in the GaAs chip industry. Nahm-Wook Lee the founder and president of BeRex, Inc. and BeRex Corporation, understands the power of the fabless semiconductor business model