Back in May we published a commentary raising concerns about the Atom CPU inventory situation among Taiwan's notebook vendors, following reports that Intel was cracking down on chip dumping in the China market. We speculated that perhaps at least one major notebook vendor was sitting on a pile of Atom CPUs that were rapidly losing value and said that, if our analysis was correct, we could expect to see more evidence around the end of the second quarter. Today Sylvie Barak at the Inquirer has posted a report claiming that Asustek is in conflict with Intel over just such an issue. "Asus apparently became overconfident in 2008, ordering massive number of Atom chips and stockpiling Eee PCs, motherboards, VGA cards, DVD ROMs and more, thinking it was too big to ever be affected by the credit crunch. Turns out Asus was wrong and was left 'holding the baby' with excessive inventory of finished products, CPUs and chipsets since Q4 last year." "[Intel is] reportedly going as far as banning Asus from accepting any China-based white-box orders. A recent Intel dictat is said to forbid distributors from selling processors to customers outside their local market, especially on the shady Chinese grey market."

Looking back at the product offerings at the recently concluded Computex Taipei 2009, you couldn't help notice that companies highlighting mobile DTV solutions based on the DVB standard were focusing on free-to-air solutions (DVB-T) as opposed to solutions targeting the subscription-based DVB-H standard, despite the latter being the technology standard developed specifically for digital mobile DTV.While DVB-H (the H is for handheld) has several advantages over DVB-T for mobile TV solutions, such as time slicing to reduce power consumption, system makers indicated that issues still need to be overcome before DVB-H ecosystems can be successfully developed.The makers noted that DVB-H networks often need to be jump-started by mobile handset service providers, who are wary of the investment costs required. In addition to being expensive the process of developing a DVB-H network tends to be complicated, one chipset executive indicated. Infrastructure needs to be deployed, while content, middleware and handset design needs to be developed and managed. So far very few service providers have felt the incentive to drive the market and those who have entered the market are not seeing much success, the executive noted.In addition, system makers do not have the freedom to help drive the market. Makers could not simply release DVB-H products on the market, as software needed to adhere to a specific network, which was usually set up by the service provider. Too much control was in the hands of the service provider to make the DVB-H networks a success. In fact, according to Yannick Levy, CEO of mobile TV chipset provider DiBcom, the global market for mobile TV chipset solutions is likely to contract 20-30% in 2009.However, while service operators have cooled somewhat to mobile DTV, system and silicon providers continue seeing opportunities, especially in the unconstrained market of free-to-air digital TV. In addition to having the freedom to design products as they wish, makers cited a number of reasons why free-to-air solutions, especially those based on DVB-T solutions, are becoming more popular. In terms of market acceptance, an increasing number of markets are now offering digital TV broadcasts, providing users with more content. In terms of power consumption, silicon providers of DVB-T solutions have made great advancements in power-savings, with one Taiwan IC design house noting that an entire DVB-T solution now consumes only about 500mW. And in terms of pricing, falling prices for components and increased integration is providing attractive price points for vendors targeting developing markets, where mobile TV is the most popular, especially in automotive solutions.GPS IC design house SiRF recently launched its SiRFatlasIV solution that allows system makers to integrate PND and mobile TV functionality in a device priced in the US$100 range. SiRF founder Kanwar Chadha commented that the SiRF solution is targeting free-to-air devices because the most important thing for growth in mobile DTV is delivering an out-of-box experience. People do not want to sign up for, or have to pay a subscription fee to get mobile DTV, Chadha noted.In fact, Telegent Systems, a silicon tuner design house that focuses on free-to-air mobile TV solutions, believes the mobile TV market is not suffering a setback this year. Free-to-air mobile TV, received over both analog and digital terrestrial broadcast TV networks, is rapidly growing the overall size of the mobile TV market, said Weijie Yun, president and CEO of Telegent.At Computex, Telegent was showcasing its hybrid digital and analog free-to-air TV solution with partner China-based notebook maker TopStar and the company indicated that it shipped more than 20 million free-to-air mobile TV receivers in the period between mid-2007 and the end of 2008. This highlights the strong market reception that consumers have for the free-to-air business model, Yun explained.Actually, most mobile TV tuner IC design houses have recognized the importance of free-to-air and are no longer simply focusing on DVB-H to promote their solutions. Leading mobile chipset providers such as DiBcom and Siano Mobile Silicon have introduced integrated chipsets that are compatible with multiple technologies and offer support for both DVB-T and DVB-H, as well as other standards such as ISDB-T, T-DMB and CMMB.One other interesting mobile TV silicon entrant showcasing product at Computex was UK-based Elonics. The company demoed an integrated multi-band RF tuner implemented in CMOS which is designed to allow users to reconfigure the RF front end for different broadcast standards. The tuner chip supports frequencies and standards including DVB-T, DVB-H, ISDB-T, CMMB, T-TMB and MediaFLO, as well as supporting GPS L1 band (1575MHz) and FM radio (64-108MHz).Like other silicon tuner providers, the company targets portable device such as handsets, which have no choice but to utilize an integrated silicon tuner due to size and power constraints. However, Elonics also plans to target the traditional TV market, including the flat panel TV space. These segments are still mostly supported by traditional RF tuners that come in the form of tuner cans, which are big, expensive, high power devices that consist of multiple parts. Elonics CEO David Srodzinski indicated that innovative monolithic silicon integrated tuners are now approaching the quality of canned tuners and this represents a huge potential market opportunity for the company.Srodzinski noted that TAM (total available market) for traditional TVs is in the 200 million range, but the key market segment for silicon tuner designers will be flat panel TVs. Srodzinski stated that currently no flat-panel TV utilizes a silicon tuner. However, this will change over the next few years, as silicon tuner quality continues to improve, Srodzinski stated. Still, however big the market potential for silicon tuners is, the newcomers will still have to compete with more established silicon tuner designers such as Infineon and STMicroelectronics.

Sitting in the Digitimes office last week, GPS IC design house Sirf founder Kanwar Chadha took out an idea book that he had commissioned back in 1995 when Sirf was just getting started. "You have to remember, at that time there really was no established market for GPS devices," explained Chadha. "Therefore, I had to hire a designer to work with me on this book to draw up scenarios that we thought would be ideal for GPS."Page after page in the book reveals the prescience of Chadha and Sirf. There are drawings of PNDs, telematics systems, walkmans and watchmans – all with accompanying text describing how GPS can be integrated into such devices and how location-based servies will make our lives much easier and safer. Chadha then breaks into a story about going to Disneyland and being one of the first people to arrive. He parked in an empty parking lot early in the morning but when he returned at the end of the day, there was no way to tell where his car was among the ocean of 30,000 other cars in the parking lot. While the point of the story is that this is yet another example of how location-based technologies can step in to make your life easier, it also serves as a metaphor for the recent state of affairs for Sirf in the market.The company was the first to take GPS to the masses, and was a main driver in bringing costs of navigation devices from a US$3,000-4,000 range to affordable levels. However, while Sirf was among the first to enter the market and has played a leading role in the growth of GPS-enabled devices for many years, a flood of competitors has steadily chipped away at the company's market share and the company's stock price dipped below US$1 earlier this year. The parking lot (market) is now full of competing players and it can be argued that Sirf has been looking for its proverbial car for about the past year or so.One challenge the company has had to face is the spread of GPS radios to devices that are not primarily designed for location. Chadha describes these devices, such as digital cameras and handsets, as having location merely as an attribute. In these devices, GPS is but just one type of radio that makers are looking to integrate. Other types of radios such as Wi-Fi, Bluetooth and FM receivers are also being integrated into the devices, which means designs that can incorporate all the radios into one solution have an advantage. As a GPS specialist, Sirf found itself losing ground, especially in the handset space, to players such as Texas Instruments (TI) and Broadcom that can deliver complete solutions bundled together.Perhaps recognizing this, Sirf and CSR, a Bluetooth market leader, decided to combine forces by merging earlier this year. The paperwork on the deal has slowly been processing – Chadha noted that the SEC recently approved the proxy filing for the merger – and a shareholder meeting is scheduled on June 25 to receive final approval on the deal. The merger should close this month, Chadha stated.Chadha voiced confidence about the role a stronger Sirf can play in the market for these types of devices, based on the expertise of Sirf and CSR. For devices where GPS is only an attribute, the key to implementing GPS is to make it easy, stated Chadha. Combining radios such as Wi-Fi, Bluetooth in an intelligent design improves the experience for the user, and this is an area of strength for the company. GPS is very sensitive so innovation will also be required to provide optimal co-existence of all these radios. The companies that can innovate in that space will be winners, Chadha stated.In addition, one type of solution cannot satisfy every customers, he noted. If an attach rate is 100%, it may make sense to combine all the solutions. But if Bluetooth has an attach rate of 80% and GPS has an attach rate of 50%, then you will continue having discrete solutions, and Sirf and CSR both hold leading positions in their respective markets.Chadha highlighted the importance of technology leadership by pointing out that every industry goes through the same process. In the beginning there are only a few products so customers are willing to pay more. Then the product becomes a commodity and people are not willing to pay anything for it. Ultimately though, customers realize that cheapest is not the best. At that point people are willing to pay more to get more. One key development moving the GPS market to the next level is that mobile handset service providers are now learning how to monetize location-based services, and therefore they no longer want to compromise on their GPS solution. Sirf is still the platform of choice for GPS, Chadha argued, and the company is getting requests from clients to specifically include Sirf technology in their devices.While location-attribute devices are adding GPS as a feature, Chadha indicated that Sirf is also focusing on extending the reach of location-centric devices such as PNDs and automotive telematics solutions. These devices would not exist without location-based services, but Chadha stated that more functionality is being added to these devices, specifically in the form of entertainment.Last year the company launched its SiRFprima platform to satisfy demand for entertainment applicaitons on location-centric devices. The chip includes an ARM11-based multifunction processor, location engine, on-chip DSP and hardware-accelerated 3D graphics engine and was targeted for multimedia devices priced in the US$500 range. Sirf then filled out the product lineup this past month by launching its SiRFatlasIV solution that targets PND-centric devices in the US$100 range. Chadha stated that SiRFatlasIV will support applications that do not need much 3D visualization and feature a smaller screen size. Mobile DTV solutions are expected to drive sales initially with some navigation solutions expected as well.Chadha pointed out that mobile TV integrated into PND solutions is very popular in emerging markets, while adding that a PND that doubles for a mobile TV does not necessarily have to be viewed in-car. The most important thing for growth in this area is delivering an out-of-box experience, Chadha stated. People do not want to sign up for, or have to pay a subscription fee to get mobile DTV. In markets such as Korea where you can simply turn on a device and start watching TV, Sirf has had great success, Chadha indicated. And with the emergence of free-to-air low-power mobile DTV solutions, SiRFatlasIV-based devices will be extremely popular, he predicted.Are these developments enough to get Sirf back in the driver seat? Chadha believes that there is still a long way to go for all GPS players in tapping the potential of location-based services and technology so Sirf still has a lot of work to do. Chadha mentioned making location work everywhere, including indoors, which is becoming increasingly important as location-based services start to take off. These services will mostly target consumers who have little tolerance and are very binary in their decision making – it either works or it doesn't – so a 90% success rate is still not good enough, Chadha stated. Power is also an area that needs to be explored. Ultimately, location-centric devices need to become wearable, and how often do we need to charge our watches, Chadha stated. We are almost there for handsets and we can prove it in the next generation of products, but we are a long way from there, Chadha said, pointing to the picture of a watchman in his book.