CNET también está disponible en español.

Ir a español

Don't show this again

Christmas Gift Guide

The $100 PC: How do we get there?

Steve Ballmer says the world needs a really cheap computer, but Mike Ricciuti warns that cheap doesn't come easy.

Microsoft's chief executive, Steve Ballmer, is no stranger to controversy.

This is, after all, the man who once said "to heck with Janet Reno" during the early days of the company's antitrust saga, and who more recently has called Linux a "cancer."

Will Microsoft leave it to the hardware makers to give back some green?

So Ballmer's remark last month that a $100 computer would help alleviate software piracy and spread computing to developing nations seems tame in comparison.

"There needs to be the equivalent of a $100 computer, not just a $400 computer, if this stuff is really going to go down-market in some of these countries," Ballmer told attendees at a conference last month sponsored by Gartner.

But by suggesting that the industry needs to offer a cheaper PC, Ballmer puts himself and his company in an awkward spot: Will Microsoft take additional steps toward making low-cost PCs a reality by sacrificing some of its fat profit margins on Windows? Or will the software giant leave it to the hardware makers--already operating on paper-thin margins--to give back some green?

Let's take a look at where things stand. Microsoft currently charges PC makers anywhere from $50 to $75 per computer to license Windows, according to analysts' estimates. While Microsoft points out that the price hasn't changed much--in actual dollars--in years, hardware costs have been plunging. So the cost of Windows as a percentage of a PC's cost has actually grown significantly.

In recent months, Microsoft has introduced low-cost versions of Windows, called Windows XP Starter Edition, especially for developing countries such as India, Indonesia, Malaysia, Russia and Thailand. Word is that PC makers are paying roughly $36 per copy for the software, which can only be sold on new PCs.

Likewise, while top-shelf chips from Advanced Micro Devices and Intel can cost upward of $1,000 each, both companies are working on low-cost chips for developing countries. For instance, AMD's Sempron CPU can be had for as little as $39 when purchased in groups of 1,000.

AMD has even gone so far as to make available a blueprint for how to build a PC for as little as $185. Called the Personal Internet Communicator, the machine is geared toward families that make the equivalent of between $1,000 and $6,000 annually. AMD says that three companies in India and Latin America are already signed up to market versions of the machine. But that system wouldn't include a monitor and would run a variant of Microsoft's Windows CE, not the more capable Windows XP.

Similarly, Intel is investigating ways to make low-cost PCs available in Eastern Europe, India and other developing areas. The company, sources say, has been selling a special low-priced processor-and-motherboard combination as part of a project code-named Shelton. The ultimate goal is to produce a PC for as little as $199.

Halfway there?
Still, that brings us only halfway to a $100 PC, especially one that leaves room for any profit margins (or at least only halfway to a $100 Windows PC). If the goal truly is to give low-income people in developing countries access to computers, and--let's be honest--to expand into the next big market for computing gear, there are cheaper and easier ways to go than Microsoft's model.

If you want to spread low-cost access to computing, and ultimately protect intellectual property, maybe it's time to revisit Larry Ellison's concept of the network computer. You probably remember that the Oracle executive pushed the notion of small, diskless "appliances" that included a monitor, keyboard, network connection and not much else--especially no place to run pirated software.

The idea was that all of the smarts were pushed down to network computers from server computers running Oracle's database and communications software. Oracle was to make its profits by selling the server software to hosting companies, Internet service providers, governments and the like.

If the goal truly is to give low-income people in developing countries access to computers, there are cheaper and easier ways to go than Microsoft's model.

Ellison even founded a company, Network Computer, Inc., to manufacture and sell the devices. But after reinventing itself as a TV set-top box maker, that venture crashed and burned when it failed to get additional funding back in 2003.

Other companies--including Gateway, Sony and the former Compaq Computer--introduced cheap Net-surfing machines four to five years ago, and all ultimately exited the market as the cost, and profitability, of PC hardware plummeted.

Now, the modern version of the network computer concept comes from Sun Microsystems, which is pushing its Linux-based Java Desktop System as a low-cost way to provide computers to people in China and elsewhere. The company signed a deal last year with China Standard Software to provide JDS to millions of consumers. Sun is also aiming the program at India, Israel, Japan, South Korea, Vietnam and other countries.

Sun's ultimate goal is make JDS ubiquitous not just on PCs, but on cell phones, set-top boxes, game devices, cars--you name it. As with Oracle's original network computer concept, Sun expects the big money to be in server-based software and hardware, and in the network infrastructure needed by network operators to make the whole scheme click.

I know what you're thinking: Those who don't learn from history are doomed to repeat it. That may be true in Sun's case. But Sun's concept has one advantage for end users: price. Take a peek at Sun President Jonathan Schwartz's Web log. Never mind the $100 PC--Sun's president seems willing to give away JDS and make his money on service providers.

Clearly, it seems we're closer than ever to ultra-low-cost computing for the masses. Here's why: Enterprise technology sales are stuck in the slow lane. Microsoft has tempered future growth estimates. And Sun and other companies need fertile new markets for their gear.

Analysts expect 1 billion PC users worldwide by 2010, up from the current estimate of roughly 660 million users. The bulk of those new users will be in developing nations.

Devising a low-cost PC isn't an exercise in altruism. At stake is an opportunity to gain a foothold in what could be the biggest technology market opportunity this century. Ballmer has thrown down the gauntlet. So who's going to take up the challenge?