Your browser is not supported. Please update it.

12 June 2006

Intel could quit cellphone chip business

Intel could exit the cellphone chip business following a 90-day review of its operations, which will be completed later this month. This would represent more than a quick fix to reduce its costs and boost its margins. It would also reflect the fundamental thinking that underpins Intel’s broad strategy ‘ that the cellphone and the PC will cease to be separate devices with different architectures and chipmakers, and that Intel’s future depends on taking a key position in that unified architecture. This will be a world of numerous formats for processing and communications products, embracing the current notebooks, cellphones, music players and games consoles as well as new devices; but with all those variations sharing basic principles of design, and all wireless connected to the internet and each other through a range of radios. In such a scenario, it will be only a short term setback for Intel to acknowledge failure to dent Texas Instruments in the conventional cellphone chip market. The life or death struggle with TI, Qualcomm and the rest of the cellular chip sector will be ensuring that the next generation is built upon building blocks in which Intel can excel ‘ IP networking, PC-style universal processors, and its most critical project of all, the digital multi-radio chip.

So if Intel does decide, as part of its current rationalization project, to sell off its communications unit, this would not affect its commitment to Wi-Fi and WiMAX ‘ the WLan chipset Centrino has been its major growth generator in the past two years and is the first success story for the Intel quest to make mobile and wireless communications an integral part of the PC architecture, rather than a separate technology with a different set of suppliers and assumptions, like cell radios. In January 2005, Intel underwent a major reorganization into five groups focused on total platforms rather than individual products ‘ the mobility group, digital enterprise, digital home, digital health plus channel. The first of these includes an integrated set of technologies underpinning mobile devices such as notebooks, but increasingly the XScale is looking peripheral in a strategy dominated by Wi-Fi and WiMAX.

This is why a sale of the main communications unit product line, the XScale cellphone processors, would make sense. With this range, Intel sought to enter a market in which it had no experience, where key alliances were already in place, and where its PC model was illogical. Intel has struggled with XScale ‘ although it no longer breaks out figures for the unit, and last year separated it from the more profitable Centrino, in 2004 it reported a loss of $791m on $5bn in revenue, and former CEO Craig Barrett admitted it had been a ‘disappointment’. Communications group chief Sean Maloney said Intel would ship 30m cellphone units in 2005 ‘ a figure not since confirmed ‘ which would be about 4% of the total, dwarfed by TI’s almost-50%.

Also in the firing line for sell-off is the IXP network processor and possibly the flash memory business, whose operating loss widened in Q1 to $104m from $32m a year earlier. Intel CEO Paul Otellini announced in April that the company had started a 90-day top-to-bottom review that it expects to result in $1bn in direct spending savings going forward. This followed a 38% first quarter drop in net income and 5% decline in revenue to $8.9bn, mainly under competitive pressure in the PC market from AMD ‘ a pressure that puts even more spotlight on Wi-Fi, an area where AMD does not shine.

“We will restructure, resize and repurpose Intel for where we think the business is going,” said Otellini. “Every part of Intel will be included in this. No stone will be left unturned, or unlooked at.” Digital Home and Digital Health are also lossmaking, but Intel expects them to grow quickly and deliver $1bn in revenue apiece by the end of fiscal 2006.

An exit from conventional cellphones would not detract from Intel’s pet project, the digital adaptive radio, and would actually increase the pressure on the Wi-Fi, WiMAX and UltraWideBand initiatives to succeed, since wireless will be the most critical competitive edge in next generation business and media client architectures. The digital adaptive radio would provide a single, low cost radio that would intelligently move between different radio links according to criteria such as signal strength or capacity. This is the real daring at the heart of Intel’s ‘handtop’ roadmap to set the agenda for the nature of the next generation client device. If it can shrink Centrino and either desktop Windows or desktop Linux into a cellphone-sized package with an adaptive radio to allow connections to many wireless options and a large disk drive, and can make that device run on battery for five hours or more, what is left for the architectures of Texas Instruments/Nokia and Qualcomm, especially when the specialized cellular networks are giving way to all-IP systems? In that context, XScale should be regarded within Intel as a limited-life product anyway, and its sacrifice unimportant in the medium term.