Freescale should get ready for change. I visited the Austin-based chip maker yesterday to talk about wireless and networking chips as well as broad trends in the industry, and walked away realizing that the firm needs to split itself up in order to survive.
The company has some very cool technology — especially around its multicore processors for embedded systems such as printers, storage arrays and routers — and a huge base of users for its Power architecture. But it has too many areas of focus. In the next two years, it’s unlikely that the company will have the same combination of businesses it has today.
Specialization is key in the chip-making industry because it allows a company to allocate its R&D more effectively, optimize manufacturing processes and generally improve profits. Freescale, which makes chips for automobiles, RFID systems, cell phones, base stations, networking equipment and industrial applications, designs both high-volume chips at advanced process nodes and low-volume chips that require a lot of manufacturing tweaks.
It’s likely that Freescale’s private equity owners will divide the company along the lines the firm established late last year: networking and multimedia; microcontrollers; cellular; and RF, sensors and analog. Each of the divisions made more than $1 billion in 2007 and could be combined with similar divisions at other firms such as Infineon, Broadcom, STMicroelectronics or even Intersil. Earlier this year, Freescale got a new CEO (from Intersil) with M&A experience, so change is certainly in the air.

broadcom
freescale
infineon
Technology-News
stmicroelectronics
semiconductors
intersil
Intel and STMicroelectronics have managed to produce a breakthrough in a new type of memory technology that could replace flash. Members of their joint venture Numonyx, which is trying to develop memory chips reliant on phase-change memory (PRAM) to store information, will present a paper today demonstrating how they’ve used it to double the amount of information they can store. It’s good news, but it still means it will cost twice as much to store the same amount of data on PRAM chips compared with the competing flash technology.
Phase-change memory relies on changes in temperature to store data; it competes with NOR flash, which is used to store the programs that run your VCR, set-top box and cell phone. The other type of flash, NAND, is used to store larger amounts of memory, such as the songs on an iPod.
Like all chips, memory follows Moore’s Law, which requires the space between transistors to get smaller. Kind of like doing the limbo, flash memory may not be able to go any lower. By allowing storage capacity to continue to expand while making chips ever smaller (and more cost-effective to produce), PRAM could provide the next breakthrough. However, although the breakthrough from Intel and STMicro puts PRAM at the same cost as NOR flash, it only stores half the data. So another breakthrough is still needed.

intel
pram
intc
Technology-News
stmicroelectronics
semiconductors