Few companies have been exposed to as much pressure in 2007 as Qualcomm, with a string of patent-related legal battles and licensing disputes putting its IPR revenue stream under threat, and now Verizon appearing to sign the death knell for the CDMA giant’s next generation network for its CDMA base. The chip company has responded in the past weeks with a blizzard of announcements designed to emphasize the strength of its current business and the wealth of new directions in which it will move to compensate for the likely decline of CDMA product and patent revenues. Underlying all this is the shift beyond handsets, focused on the Snapdragon architecture, and a heightened focus on growth areas that are tangential to the core cellphone chip business, such as MEMs and software.
At a recent investor presentation hosted by Lehman Brothers, Qualcomm COO Sanjay Jha emphasized that, in the face of dramatic changes both in the chip giant’s own business and the mobile market as a whole, there would be a period of increased investment in pre-revenue projects that would not bear significant fruit for two or more years ‘ investment that would be more than justified by the ability to enter emerging markets at an early and profitable stage, and that would be supported by Qualcomm’s current strong financial performance (even with the contribution from Nokia’s disputed royalties discounted, the company has seen six successive quarters of record revenues, and is forecast to see a 24% year-on-year increase in revenues for 2007). According to Jha, the percentage of R&D going on pre-revenue projects will rise from 27% in 2006 to 41% next year.
The areas in which most of this will be spent are highly indicative of the strategy Qualcomm will be pursuing. The top four are OFDMA, MEMs display technologies, wireless Lans and the Snapdragon architecture, and services platforms.
The first of these has been the most controversial in Qualcomm’s recent history. The company strongly defended its CDMA technology against the shift towards OFDM-based systems for pre-4G networks, claiming CDMA remains more spectrally efficient and strong in performance for fully mobile connections. As the OFDMA tide became unstoppable, Qualcomm has made reluctant concessions, though one of these ‘ the Ultra Mobile Broadband 4G extension to the CDMA2000 roadmap ‘ now looks in doubt, with Verizon and Sprint rejecting it for their next generation networks and the CDMA equipment makers defocusing on it. If Qualcomm is forced to retreat from having its own OFDMA option in the market, it is likely to concentrate its OFDM efforts on stealing a march in LTE and fast Wi-Fi standards (and even, we believe, its much hated WiMAX), and on adapting its technologies – including the Flarion Flash-OFDM and former
TeleCIS 802.16e assets – for more specific systems such as the MediaFLO mobile television platform.
MEMS (microelectromechanical systems) are another key departure, and Qualcomm has a whole division, QMT, devoted to this technology area. MEMS combine silicon-based microelectronics with micromachining, adding ‘arms and eyes’ to the microelectronic brain (as the MEMS Forum describes it) to create tiny systems that bring together
mechanical elements, sensors, actuators and electronics on a common silicon substrate. MEMS have been used extensively in applications that rely on sensors for automation, such as automotive, and are now being built into cellphones and other mobile devices. For instance, the Apple iPhone and Nintendo Wii use them for motion detection, which will soon be adapted for new functions such as answering a call just by shaking the handset, and replacing menus with small movements. It is expected that, this year, 400m MEMS chip units will be shipped, about 5% of the total foundry market, and analysts predict the market will be worth $20bn by 2016, 70% of this in semiconductors. Qualcomm’s initial focus has been on MEMS displays, which are as bright as an LCD, but require no backlight and consume little power.
This approach will also become important in some aspects of another key Qualcomm initiative, in machine-to-machine communications, especially where combined with location awareness. Tracking and logistics, security, military applications and device management are particularly important. A year ago, Qualcomm announced plans to acquire nPhase, whose products are designed to enable enterprises to manage and monitor widely dispersed, fixed machine assets. Qualcomm first entered the M2M market in 1991 with the introduction its first wide area, wireless machine-to-machine application, OmniTRACS, designed for transportation fleet management. Analyst firm iDate predicts the M2M market will total roughly $280bn worldwide by 2010.
Combining several key Qualcomm directions ‘ M2M, location and GPS (the company is the world’s largest GPS chipset provider), and enterprise software and services ‘ this month saw the company launching MobileView, software for remote collection of mobile
and network performance data using GPS. T-Mobile Germany is one of the first triallists. “Remotely collecting performance data from opt-in users is the next trend in handset testing and network optimization,” said Tia Cassett, senior director of business development at Qualcomm CDMA Technologies.
Two of Qualcomm’s most important silicon platforms for its diversification programme will be Snapdragon and Gobi. The former is designed to take Qualcomm beyond handsets and into other wireless and multimedia enabled computing and consumer products. The first two chipsets shipped recently, with commercial devices expected in mid-2008. The platform incorporates a 1GHz control processor called Scorpion, based on ARM’s Cortex instruction set and supported by an array of 600MHz DSP and accelerator cores for baseband and video processing. An early adopter is Taiwan’s HTC, which is creating a Windows Mobile handset based on Snapdragon. Gobi is a software defined radio chipset to support handsets that can switch intelligently between different networks (see Wireless Watch October 29 2007).
With a common modem architecture for both LTE and UMB, Qualcomm claims to be protecting itself against the risk of the latter being stillborn, as well as streamlining R&D costs, and as well as expanding into OFDM and new vertical markets, it is also investing in new equipment formats, such as femtocells, and in new process technologies, taping out its first 45nm chip at one of its main foundry partners, TSMC, in August, and working on a 40nm follow-up.
This is all about convincing the investors and customers that Qualcomm is mitigating its considerable risks by spreading its bets across a wider range of technologies and by ensuring it is at the front of the pack in a selection of high growth markets where its engineering, process and integration skills can play to best advantage. Shares rose after the Lehman presentations, and there is a warming of sentiment towards the company after some recent positive decisions in the courts, but the moment Qualcomm sees any drop-off in its recently stellar financial performance, the impact of Verizon’s announcement will start to be felt in earnest. Last month, the power of short term investor behaviour, rather than long term strategic thinking, was seen when Qualcomm disappointed the markets with a conservative outlook for fiscal 2008. Shares fell 7% on fears that Qualcomm would lose market share to Broadcom next year, and on signs that, because of a change in the way Qualcomm reports revenue from license fees, its royalty business appeared to be less profitable than had been expected. Qualcomm said revenue and earnings would be below analyst estimates for its fiscal first quarter and the year ahead and predicted 2008 earnings per share (excluding its investment arm and other items), of $2.03 to $2.09, below analysts’ average forecast of $2.19 per share.
In the fourth quarter of fiscal 2007, Qualcomm’s profit rose to $1.13bn, or 67 cents a share, from $614m, or 36 cents a share, a year earlier. Revenue rose to $2.31bn from $2 bn. The company also forecast 2008 revenue of $9.5bn-$9.9bn, below average analyst expectations of $9.97bn.