SAN JOSE, Calif. Advanced Micro Devices is changing both its products and business model in its plans to address low cost PCs for developing countries. The moves are part of a new direction in the wake of cancelling earlier this month its Personal Internet Communicator (PIC), a low cost PC in a sealed case aimed broadly at developing countries.
AMD plans to roll out to systems makers in the next two months multiple low cost PC reference designs based on its Geode processor and tailored to specific markets in developing countries. The company decided it will not sell to service providers complete systems like PIC which it said was too generic to serve diverse market sectors.
"We are going back to our core competency of selling chips. We will make reference designs and work with OEMs and ODMs, but we will not do an end product," said Billy Edwards, chief innovation officer at AMD.
The new reference designs will "scale up, down and sideways" from the PIC which was based on the low-end GX500 version of the Geode. Some of the designs will remove features to provide higher reliability and lower cost than the PIC which sold to services providers for as little as $200.
The reference designs ultimately will include processors ranging from the Geode LX700 to the NX1500 and operating systems including Windows Starter to XP Home and Professional, Edwards said.
"There are a series of new devices we are working on. There is something to be said for understanding market needs and tailoring systems better for them," Edwards said.
All of the systems aim to provide a lower tier of products than those based on AMD's Sempron x86. They will target a range of markets from simple systems that let farmers check market prices to more fully featured models for small businesses that cannot afford traditional computers.
"We are focused on going below the traditional model," Edwards said.
AMD still believes there will be a significant market in low-end PCs for developing countries, although that market is taking off slower than anticipated.
"We thought we could ramp [the PIC] faster, but our timing expectations were off. These markets are very different from what this industry has been used to for the last decade," said Edwards.
Nevertheless some 83 percent of the world's population has no access to the Internet. A broad swath of people with annual incomes from $6,000-$8000 could afford something a notch below a traditional PC, according to AMD.
"You're talking about lots of people and that's the ultimate driver. The question is how fast we can do it," said Edwards.
Non-traditional PCs sold by services providers for Web access on a subscription basis will be the first approach likely to have success. "There are a couple of business models here that can work, but we need to build them up," said Edwards.
Further in the future, pay-as-you-go models such as Microsoft's FlexGo hold promise, but require local banks to set up new small loan services. "There's a whole infrastructure of banking to make this work, and banks don't change as fast as our industry," said Edwards.
"Eventually we'd like to merge our [reference design] efforts with [Microsoft's FlexGo], but they are not merged in the next wave of systems," said Edwards.
Still farther in the future is the prospect of low-end PCs subsidized by ads automatically presented to users. "Some people can't afford a computer, but they will look at ads to get Net access, and there are local advertisers interested in getting to these people," said Edwards.
Edwards defended the PIC for being reliable and easy to use, but admitted the generic design had a relatively high cost. He said PIC helped inspire work on other initiatives such as One Laptop Per Child (OLPC), which tales a specific focus on education.
The OLPC initiative manufactured its first batch of about 1,000 beta systems in November for testing by users in Argentina, Brazil, Libya, Nigeria and Thailand. The group hopes to make and ship millions of the systems once the design becomes final probably in 2007.