In Search Of A PC For The People

The race is on to serve the next billion in emerging markets. Whose low-cost model will win?

In Silicon Valley, "the next billion" is shorthand for the vast potential market in the developing world, where few people have access to PCs. For Mark J. Beckford, it's almost an obsession. The 39-year-old Intel Corp. (INTC ) general manager has spent most of the past decade trying to create computers for that next billion. From his office in Shanghai, he oversees a team of about two dozen engineers in China, Brazil, Egypt, and India, all designing PCs suited to the needs and wallets of customers in emerging markets. "The next billion isn't going to come by pushing the same things," he says. "It requires new levels of affordability, access methods, ease of use, connectivity, and power."

These days lots of companies are trying to serve that same billion. Intel's biggest rival, Advanced Micro Devices Inc. (AMD ), is working on various possibilities, including a PC in India that sells for about $200. Taiwan's VIA Technologies Inc., the world's No. 3 designer of microprocessors, has launched a business group focusing on low-cost computers for emerging markets. Microsoft Corp. (MSFT ) in late May introduced a low-cost solution. And Nicholas Negroponte, a professor at Massachusetts Institute of Technology, is leading a group called One Laptop Per Child, which aims to produce a machine for $140 or so by yearend, and as little as $50 by 2010. "I think of digital access for kids as a human right," Negroponte said in an e-mail interview.

But as they target emerging markets, the world's info-tech powers are grappling with a host of difficult issues. How do you design a PC that's affordable to almost anyone? What traditional features do you ditch to do that? Will users in developing countries be satisfied with these computers, or will they resent the idea of getting dumbed-down machines? Do they want global brands, or are they happy to accept no-name alternatives?


Just about everybody agrees selling computers in the developing world is more complex than simply adding Mandarin, Hindi, or Arabic software to existing PCs. In many places, PCs must withstand desert heat and sand -- and cope with frequent electricity outages. That requires more rugged designs that are also energy-efficient. "In emerging markets we see a huge appetite for PCs," says Richard Brown, a VIA vice-president who heads the new PC-1 business unit, set up early this year to sell computers in developing countries. "But they need to be smaller, cooler, quieter, and greener."

Microsoft, meanwhile, has come up not with computers, but with an innovative way to finance them. On May 22, the software giant announced FlexGo, software that keeps machines from working until users type in a number from a pre-paid card. The idea is that consumers in developing countries, who might not be able to shell out $500 or more in one go, can afford perhaps half the cost of a PC up front, then pay for the actual hours they use the machine via the cards. After a certain number of hours -- and payments -- the computer becomes the property of the consumer. In the developing world, computers are "simply out of reach for people who would like a PC in the home," says Will Poole, senior vice-president of Microsoft's Market Expansion Group. "What do we do to change the equation?"

Most companies are seeking to solve that equation through innovative design. For instance, Intel in March launched its "Community PC," targeted at Indian villages where those who can't afford computers of their own share a common machine. The Community PC, which costs about $550, has a filter to keep out dust, can run on a car battery when blackouts occur, and is equipped with a one-button "recover" feature in case of crashes. Also in March, Intel introduced a $250 to $350 miniaturized desktop computer (with the clunky name of Low Cost Full Featured PC); Mexican officials have ordered 400,000 for delivery by November. And the company in May revealed a prototype of a new notebook called the ClassMate PC that carries a price tag below $400.

Designing machines for developing countries can be taxing for engineers accustomed to building PCs from off-the-shelf, commodity parts. "Before, it was just cut and paste, follow the guidelines," says Kent Geeng, president of iDot Computers Inc., which designs VIA's low-cost PCs. To reach the target price of $230, his engineers had to start from scratch. "A cost-effective machine is much more difficult than doing the high end," says Geeng. "It's like building a Nissan to drive at the same speed as a Porsche."

That's just what bothers some critics, who don't believe people in poor countries need to compute at Autobahn speeds. Stephen Dukker, who founded low-cost computing pioneer eMachines Inc. in 1998, now runs Seoul-based Ncomputing Ltd. His goal is to sell not PCs but "thin clients," diskless machines that work only if connected to a server. Since a single PC can run 10 of these devices, schools and libraries in impoverished areas could get computers in front of many more people far more cheaply than by buying actual PCs for everyone. Dukker's device -- little more than a chip surrounded by plastic -- costs just $70 or so. "Why are we trying to force people to use a computing technology that's not appropriate for them?" says Dukker.

PC alternatives, though, don't have a great track record. Oracle Corp. (ORCL ) and Sun Microsystems Inc. (SUNW ) tried to push thin clients in the late 1990s, without much success. Negroponte has attracted a lot of attention, but skeptics say that children -- and their parents -- may not really want a machine that cheap. "You have a low-cost device, but it's very limited," says Beckford's colleague, Intel Vice-President Bill M. Siu.

The real demand, says Microsoft's Poole, is for "a solid, midrange PC." With so many companies targeting emerging markets, there's a lot riding on the right answer.

By Bruce Einhorn, with Jay Greene in Seattle and Peter Burrows in San Mateo, Calif.

    Before it's here, it's on the Bloomberg Terminal.