SHANGHAI, China — A small startup based here developing RF chips is looking to capture 50 percent of China's fast-growing market for PHS handsets, which currently stands at around 23 million users and is growing by about 2 million users a month.
Led by a core group of overseas Chinese, Comlent Technology Inc. is looking to displace Japanese manufacturers as the major supplier to Chinese handset makers such as market leader UT Starcom. The company joins a small but growing number of startups in the Asia-Pacific region that are tackling more complex RF design.
At the same time, Comlent is a tangible result of the Chinese government's policy to lure back overseas Chinese -- with corporate and personal incentives -- so that they can train local engineers and help seed a slow-to-grow IC design industry.
Despite being in the early stage in its semiconductor industry development, China already claims to have about 400 IC design houses. But most industry insiders say the number of serious players is well below 100.
Only a handful have annual revenue over $10 million. Most are doing low-end, copy-cat designs for white box goods and are not that adept at isolating new market opportunities.
Comlent CEO Kai Chen said his company is trying to differentiate itself by bringing true technology innovation to the table, and mixing it with China's lower cost structure. "We are not just using Chinese labor to make the chip 10 percent cheaper. We are revolutionizing the whole architecture," he said.
Comlent is looking to exploit a common complaint among PHS users in China — dropped calls. Current RF chips used in PHS handsets communicate with just one cell site at a time, Chen said. Comlent's chip, which will use direct-conversion (zero-IF) technology, will simultaneously talk with four cell sites.
Single-chip, direct-conversion technology, which puts the onus on the radio to bring the zero IF intelligence-only signal to the baseband, should give Comlent a cost advantage over a superheterodyne, two-chip method. Chen estimates he can reduce the handset bill of materials cost by as much as 20 percent.
Comlent said a major PHS handset maker in China has already paid its nonrecurring engineering costs and accepted engineering samples. The chip will be made on Jazz Semiconductor's 0.35 BiCMOS process and ship in volume next summer. Comlent also has a 2.4 GHz cordless phone transceiver being made on a 0.18 RF CMOS process at Taiwan Semiconductor Manufacturing Co.
So far, the company has no plans to immediately expand into more mainstream communications chips, where it could run into stiff competition from more established foreign companies, said Chen, who is a former IBM Microelectronics engineer.
Instead, the company will stick with niche markets. It is just starting R&D on 3G, and is considering a dual-band GPRS/TD-SCDMA transceiver. Most of the early work will be done by a team of senior engineers in southern California, who have come to Comlent from Texas Instruments, Broadcom, IBM, National Semiconductor and Conextant.
Shanghai will remain the base for IC layout, system integration, and fab process engineers, which are easier to find, Chen said. Typically, experienced RF designers have been hard to find here so he has hired a handful of inexperienced RF designers that are being apprenticed to a few senior designers.
Potential pitfalls
Even though Comlent already has a major customer in hand, the company still faces uncertainty. The PHS market — known as PAS in Japan — is arguably illegal in China because the Ministry for Information Industry, China's IT regulator, has not officially sanctioned the use of spectrum for PHS.
So far, like many things in China, the situation remains vague. The explosive growth of PHS this year (one telco provider went from 9.7 million lines in 2002 to 17 million by August this year) is due to big pushes from China's biggest fixed-line providers, China Telecom and China Netcom. Neither has licenses for wireless networks, and so they loosely describe PHS as a fixed-line, limited wireless solution.
"MII has turned a blind eye," said Ken Grant, an analyst at MFC Insight, a Beijing-based telecoms consultancy. The question is how much longer will it continue to do so. Comlent's Chen is hoping the market grows fast enough that it would make sense to keep PHS — also known as Little Smart or Xiaolingtong. Grant agrees. "If it got to big, there would be a problem unwinding it all," he said.
Other chip makers are also looking at the PHS market, which at the moment is too hot to ignore. With advances in chip design increasing quality of service for end users, specifically when it comes to fewer dropped calls, then the market could become even more attractive.
Chen, at least, is counting on that. His firm will take in about $5 million next year, pushing it into the top 10 of China's fabless IC design houses. "By 2006, we can do $60 million," he said.
