The Next PC Revolution will be Televised
By Om Malik

(Business 2.0) – About two miles inland from Chennai's long white-sand beaches--part of a stretch of Indian coastline hammered by last year's tsunami--the downtown sidewalks buzz with life during the morning rush. Vendors flog coconut water and newspapers while rope-thin men in ankle-length lungis jostle for space with office workers in polo shirts and slacks. Three floors up in one of the city's numerous office towers, past a row of cubicles and half a dozen programmers, 38-year-old Rajesh Jain points to a table that holds, he'll tell you repeatedly, personal computing's next big thing.

The silver-and-black box is tiny, actually. At 8 inches by 6, it's the size of a fat paperback. But plug it into a computer monitor or a television, and Jain's boast begins to make some sense. A tap on a keyboard brings up an array of generic-looking applications you'd find on any PC--a Web browser, e-mail, and word-processing and spreadsheet programs. An MP3 player blasts "Dus Bahane," the latest hit song from Bollywood. All that software is open-source and free. Better yet, it runs on a distant back-office server instead of the little box, which itself sips less electricity than a child's night-light, contains no hard drive, and doesn't make a sound. But that's not what gives the Nova NetPC--made by Novatium, Jain's 10-month-old startup--its industry-altering potential. It's the machine's expected price tag: $100, including a monitor, keyboard, and mouse.

The $100 PC has long been considered the hurdle to clear in order to reach technology's biggest pot of gold--affordable computing for the masses in countries like Brazil, China, India, and Russia. Make no mistake, this isn't just altruism. A cheap PC is a great business opportunity for anyone who can build a 10 percent profit margin into each device, as Jain's company is trying to do. That's why chipmaking goliath Intel is working on cheaper processors targeted overseas, why Microsoft has begun selling a $20 stripped-down version of its Windows operating system, why giants like Advanced Micro Devices and Google have partnered with maverick MIT professor Nicholas Negroponte to develop a $100 laptop. And long before them, Oracle and Sun Microsystems chiefs Larry Ellison and Scott McNealy tried, and failed, to market so-called network PCs.

So what gives tiny Novatium an edge over such high-profile competition? Most of those companies have focused on making traditional desktop PCs or laptops cheaper by using older, slower chips and skimping on memory and hard-drive storage. Novatium, on the other hand, has created a state-of-the-art network computer that mimics a traditional desktop machine at a fraction of the cost--and that will soon be made to run on any television, anywhere.

That's not to suggest that Novatium has a clear path into millions of living rooms around the world. The Nova won't hit the Indian market until October at the earliest, and the company has yet to sign up a broadband partner, without which the business plan--to bundle the machine with high-speed Internet service--is DOA. For Novatium, though, the only path to the mass market starts with a product that, while inexpensive, cannot be considered cheap. "I don't know of any company that has succeeded by telling customers, 'I'm offering a cheap product to you because you can't afford the other products on the market,'" says Novatium CEO Alok Singh. "We have to deliver an experience beyond the traditional PC."

For nearly three decades, the personal computer has been more than just the dynamo behind Moore's Law: It helped the United States keep its edge as the technology leader of the world. But the PC has hardly been a guarantor of tech dominance. The United States led the world in mobile-phone technology in the 1980s, only to watch European and Asian companies run off with the business. Why? The Europeans settled early on the standards and a business model--bundling affordable phone service and handsets--that ignited the mass market.

Novatium sees a similar opportunity lurking today. Just as millions of Indians skipped land-based telephones altogether and went straight to wireless when it became affordable, Jain and Novatium's other two founders--Ashok Jhunjhunwala, a renowned engineering professor at the Indian Institute of Technology, and Ray Stata, chairman of U.S.-based Analog Devices, a $2.6 billion chipmaker--are betting that they'll skip desktop PCs and go straight to network computers.

While PC sales growth has slowed in nearly every major developed nation, computer sales in India have jumped 44 percent in the last year, hitting 3.4 million units in 2004. But that's just a week's worth of the total global output, and PC household penetration in India, a country of 1 billion people, stands at a miserable 0.9 percent. Still, if you slash the going rock-bottom price by two-thirds, the thinking goes, millions of people--with no allegiance to Microsoft or Intel--will join the PC revolution. "It's obvious we're borrowing from the mobile world," Jain says. "Computing must become more like a utility, where you pay for monthly usage."

Pushing computer usage deeper into the impoverished masses, some think, wouldn't just enrich Novatium's founders. "Network PCs are the only way to bring people from the periphery into the mainstream," says Ashok Kumar, a PC industry analyst at investment banking firm Raymond James Financial.

The herculean feat Novatium faced from the start was getting the unit cost far enough under $100 to reach a single household, much less clear a profit. The cheapest PCs today, which sell for $300, incorporate a $75 processor, $25 worth of communications and graphics chips, a $75 hard drive, and a $25 fan. So they've already eaten $200 before the computer case and Windows OS are added.

Network computers currently sold in the United States--like Sun Microsystems's Sun Ray machines, used by banks and drug companies--cost about the same as standard PCs and require a controlled corporate environment to keep them running and updated. That's impossible in places like India, where heat, dust, and power outages are accepted facts of life. "The technology for the $100 PC has to come from within the markets that are likely to adopt them," Stata says.

With $2 million in seed money from Stata and Jain--a software engineer who sold the Internet portal he started, IndiaWorld, for $115 million--the founders hired a handful of engineers and set up shop in Chennai to pursue what Jain calls "an impossible dream." Novatium's device would have to be as easy to use as a TV, consume very little power, and still function like a state-of-the-art PC.

The first thing to go was the hard drive; in its place, engineers suggested, a simple USB port would suffice. Users could store data over the network or on ever cheaper USB memory sticks and external hard drives. Engineers also found that a $10 16MB flash memory chip could replace far more expensive RAM.

Finding a microprocessor, the engine of any computer, wasn't nearly as simple. Stata knew that a high-performance power hog like Intel's latest Celeron, at $75 a chip, was out of the question. Cheaper and slower chips were on the market, but would they be available in the quantities required?

Stata saw the answer to the problem cradled in his hand--his cell phone. It ran on a simple, stripped-down chip called a digital signal processor, already a standard in the wireless industry. If limited to just a handful of tasks and a tiny operating system--well suited for Novatium's planned device--a DSP, Stata knew, would work as well as anything made by Intel. The chips also ran on just 5 watts of power, compared with 25 for a Celeron. Not only would that make it possible to keep a Novatium device running on backup batteries--a common requirement in outage-plagued Indian homes--but it also meant doing away with the expensive cooling fan. Best of all, though, was the cost: about $10.

Next came another daunting technical snag: creating software that would mimic Windows apps from afar. Jain had previously launched a startup that leased software as a service on the Net. Novatium would have to operate much the same way, except that the service--pushing a handful of desktop applications to the Novatium box--would be included in the price. And the apps would have to run glitch-free.

The challenge was appealing enough to 34-year-old Vinod Kumar, a top software engineer at IBM in India, that in February he dumped his 10-year career at Big Blue--and took a big pay cut, he says--to work for Jain and Novatium. Once the hardware specs were ready, Kumar and half a dozen other engineers set to work writing Novatium's ultralean code, which could handle everything from Web browsing to streaming video and webcam sessions.

By the end of February, the company had nailed down the chips it needed, and the software applications were running fine--individually. Switch from one app to another, though, and the Novatium box ground to a near halt. The problem was a sluggishness that sets in when there's a lot of activity on the display--in technical terms, a drop in the screen refresh rate. The rate sank to 10 frames per second, compared with 30 on a typical PC. Jain and Kumar knew they needed to bring it up or they were doomed.

It turned out to be the problem no one could crack, even though Kumar and other engineers worked around the clock for weeks. With investors grumbling phrases like "do or die," newly hired CEO Singh called an all-hands meeting. Without enough chairs for all 40 staffers, they sat on the carpet in front of the demo machine. "Let's try to understand what the problem is," Singh recalls saying. "I don't care if we solve it."

Within a few minutes, his engineers began to chip in ideas. "Nobody had looked at it end-to-end before to understand the problem," Singh says. With a set of hypotheses about what was bogging down the machine, the crew went back to the drawing board to try to prove their theories; every evening they compared notes. Soon they came up with a standardized test to identify critical bottlenecks. Not long after that, solutions began to emerge. Today, Singh says they've achieved a refresh rate of 25 frames per second--good enough to take to market.

Of course, actually getting the Nova to market will be the hardest part of the trip, and there are plenty of reasons for Novatium's investors to remain nervous. Foremost is the lack of a major Indian broadband partner or ISP whose service will be bundled with each NetPC. Without one, Novatium has no means of reaching the computerless millions en masse.

Then there are the rivals. Wyse Technology, a fallen angel from the client-server era, is offering its terminals as PC replacements to businesses and Internet cafes in China and India. Intel is building a low-cost processor targeted at the same countries.

Still, Novatium has grabbed the pole position for now in the race to build a $100 PC. Jain says the company is deep in talks with several telecom partners and expects a deal soon. For the time being, Kumar and the rest of the coders are working to make their NetPC display correctly on any television set that accepts RCA input jacks. That's key to the plan, because Novatium envisions the TV--a central fixture in the homes of many of the world's poor--eventually becoming the system's monitor, dropping the price even further. Compared with some of the technical hurdles already overcome, Jain says, the coding problems aren't particularly daunting and will be resolved in plenty of time for Novatium to hit its target of September for rolling out its first production model. Besides, Jain points out, nobody ever said triggering the next PC revolution would be easy.