Inveneo’s offices, next to the Virgin Megastore on Market Street in San Francisco, have the feel of a small-town plumbing-supply company. Beat-up desks and tables in the three-room headquarters are surrounded by piles of boxes, electronic equipment scattered on the floor, and two stray hand trucks. Kristin Peterson, who runs the place, wears her blond hair in a pixie cut; her typical work attire is a green long-sleeve T-shirt, jeans rolled up to her calves, and running shoes. But in such humble trappings beats the heart of a revolution that threatens the tech industry’s superpowers. Inveneo and a handful of similar operations—such as One Laptop Per Child, which hopes to make $100 computers—are a nightmare for Microsoft, Intel, and Dell because such ventures could blow a massive hole in computer prices and win over the next wave of  new computer and internet users.

In the next decade, the most explosive growth in technology will come from “the next billion users”—the industry’s label for that huge slice of the global population that so far hasn’t been able to afford PCs and network connections. (About 1.2 billion of the world’s 6.6 billion people currently use the internet—generally speaking, the wealthiest 1.2 billion people.) Inveneo is a nonprofit working to bring computing and the internet to people who have never touched a mouse pad, in places such as a refugee camp in Uganda. Inveneo’s computers run a Linux-based operating system on a cheap chip made by Advanced Micro Devices intended for photocopiers. The Inveneo machines, including a router to connect to the internet and e-mail, cost $400 or less, depending on how many are ordered, and consume so little power they can run off a car battery. Why does that worry Bill Gates? Machines made by Inveneo and O.L.P.C. might allow the next billion first-time computer buyers to get online without ever using Microsoft Windows software, Intel chips, or Dell computers.

The implications of that could be dire for Microsoft: it could be cut off from the next decade’s hottest tech market. And you can be sure that Microsoft, Dell, and the rest have done the math. They are now scrambling to get into this game. “Seven years ago, when we first approached companies to ask for their help, they’d dismiss [our idea] and tell us to go down the hall to the grants department,” says Rey Ramsey, C.E.O. of One Economy, another non-profit working to deliver information technology to impoverished regions. “Around three years ago, there started to be a flip. Now they see there are legitimate business reasons to be in this space.”

The tech powers need to defend themselves, which helps explain the constant stream of “next billion” initiatives they are pitching in press conferences. Their programs are always sheathed in high-minded spin, typically including language about “closing the digital divide” and “getting poor villagers out of poverty by giving them information technology.” But behind the lofty goals is a ruthless analysis of the bottom line. Last year, Intel committed $1 billion to its World Ahead program; in April, Microsoft chairman Bill Gates traveled to Beijing to unveil his company’s Unlimited Potential operation. The primary aim of both projects is to bring technology to schools in developing nations. Dell and Lenovo both recently unveiled cheap computers for the Chinese. Only Hewlett-Packard seems to be missing from the dash to “do well by doing good,” as the business-chic aphorism goes.

There’s no reason to be entirely cynical about the motivations of these companies. “Long-term, are they jockeying for position? Of course,” says Inveneo’s Peterson. “Are they doing this only to get into these markets? No way. They are very committed to helping.”

Still, big organizations in a brutally competitive industry don’t simply decide to simultaneously launch high-profile public-service programs just because it’s What ­Jesus Would Do. Behind the curtain, the tech behemoths know they have to get their hooks—and software—into the developing world before some cheaper, good-enough rival technology gets there first.