September 26, 2011 — Fabless wireless semiconductor companies are putting out new chips and breakthrough technologies, adding "growth momentum" to the semiconductor industry value chain as a whole, according to a CyberMedia Research study, "The Fabless Wireless Communications Industry 2011: Analysis of Key Players."

The rise in fabless semiconductor companies coincides with so many wireless technologies emerging on the ground, noted Anirban Banerjee, associate vice president, Research and Advisory Services, CyberMedia Research, who asserts that semiconductor companies "need to be flexible" in technology platforms.

Fabless companies rarely outsource R&D, spending over half of their OPEX on R&D. In 2010, average R&D spend was 50.95% of the OPEX of fabless companies globally. Fabless chip companies typically run a headquarters with R&D and administration. The sales, marketing, and customer support offices are set up as needed, explained Faisal Kawoosa, lead analyst, Semicon Practice, CyberMedia Research. For wireless fabless companies, 56% of offices are in China, India, Taiwan, Singapore, Hong Kong, and other Asian regions. This keeps fabless companies in touch with original device manufacturer (ODM) and original equipment manufacturer (OEM) customers, as well as large populations of end users.

TSMC is the leading manufacturing partner to fabless companies, with 40% of fabless chip companies using TSMC as their sole or lead foundry partner. Amkor is the primary assembly and test (packaging) partner — 28% of companies use Amkor.

Fabless chip companies typically focus on one or a few technologies, approaching wireless infrastructure as well as terminals. The difference is in the breadth of offerings, wider in the case of terminal equipment and products due to the variety of user preferences across geographical and cultural areas. In 2010, not many products were announced by fabless companies developing solutions in the wireless communications space. This is because major operators in the wireless communications market were still hedging their bets — LTE versus WiMax — and fabless companies weren’t yet in the right business cycle stage. Most fabless companies globally have been
operational for a period ranging between 5-8 years; this is the time frame that is required to develop and sell a single silicon product.

Figure 1. Popular architectures used in chip design for wireless products. Source: CyberMedia Research, 2011

ASIC is the leading chip architecture in terms of product variety in the wireless communication space. Fabless companies are adopting system on chip (SoC) designs to increase performance in smaller form factors. ODM and OEM pressure for fast time-to-market is another driver for SoC adoption, the report shows.

Figure 2. CMR semiconductor market ? outlook for wireless fabless chip companies. Note: Size of the circle does not signify the relative size of the market. Source: CyberMedia Research, 2011

Today, PAN (Personal Area Networks) and BAN (Body Area Networks) are going beyond
Bluetooth and will become a critical piece of the network once Machine-to-Machine data
transfer volumes pick up in homes and offices of the future.

"In the arena when a device is connected through high speed 3G/4G networks, there will be a pressing need to have an equally efficient personal-area network (PAN) technology supporting faster throughput/transfer rates for sharing rich media content in homes, offices and other environments between a user’s many devices, or among a small but closed user group. This will be the next opportunity area for fabless wireless chip makers," concluded Kawoosa. At this point, no one standard is emerging for the industry. Multiple companies will try products, and the market will likely have shake outs in various areas.

WAN technology will continue to be dominated by WiFi although there will be evolution to 802.11n standards to support faster transfer rates with higher throughputs. Since many chip companies are offering solutions in this space with WiFi being the default standard, it is unlikely that more players will get attracted in developing WAN solutions. According to Faisal, "Investments involved in R&D, marketing et al for developing chip solutions for air interface technologies are high and commitments are typically in the form of long term contracts between device manufacturers and chip vendors. The level of competition in this domain is therefore expected to be moderate. " "MAN wireless technologies like 2G, 2.5G, 3G and 4G (LTE and WiMAX) will be the most promising opportunity areas for chip companies. Any wireless device requires an air interface to connect to a network, hence these technologies are expected to throw up the maximum opportunities for fabless wireless chip companies going forward," Faisal concluded.

A part of CyberMedia, South Asia