Microsoft has set itself a daunting challenge. On November 8, the Redmond-based software group plans to launch its Xbox games machine, laying down the gauntlet to Sony and Nintendo, the Japanese duopolists which completely dominate the global games console market.

The market entry for the company, best known for its Windows operating system and Office suite of personal productivity tools, is fraught with risk. The Americans have tried to challenge Japanese hegemony before.

In the mid-1990s, 3DO, led by a former senior executive from Electronic Arts, the world's leading games software maker, took on the Japanese in the console market and failed dismally. A similar launch by Philips, the rather better capitalised Netherlands-based group, suffered an equally embarrassing fate.

A botched launch would be hugely embarrassing for the Redmond-based group and an expensive fiasco even for a company with a cash balance the size of Microsoft's ($31bn and counting). The group has committed $500m to marketing the machine over an 18 months period. Analysts believe it will also lose between $100 and $150 on each machine it sells. Merrill Lynch estimates Microsoft could lose as much as $900m building a customer base this financial year.

The challenge is daunting. Manufacturing is difficult enough for established vendors: Sony botched the launch of its PlayStation 2 console because of shortages of the machine; Nintendo has had to delay the US introduction of its GameCube to be sure of enough machines; and Sega's launch of the Dreamcast was so disastrous it was eventually forced out of the hardware market.

Even if Microsoft has enough Xboxes at launch, it will achieve critical mass only if it can persuade software developers to create a sufficient number of games that are good enough to persuade people to buy its machine.

But if the risks are considerable, so too are the rewards. Once a customer base is established, the cashflows from software sales can be colossal. At one stage, the PlayStation division at Sony was generating more than a third of the entire Japanese electronic group's operating profit.

The strategy formulated by Robbie Bach, head of the Xbox division, has looked textbook, identifying the needs of games designers, players, manufacturers and distributors.

Mr Bach has promised to launch the Xbox in North America on November 8 with between 600,000 and 800,000 consoles. By the end of January 2002, he has promised he will have shipped between 1m and 1.5m machines.

Mr Bach is well aware of the dangers of not meeting the production targets. Unlike Sony, which manufactures its own critical chips and assembles the PlayStation 2 in its own plant, Microsoft has out-sourced chip supply and manufacturing. Flextronics, based in Singapore, will assemble the Xbox in two plants, one in Mexico and one in Hungary, to mitigate problems on a particular production line.

To reduce the risk, he has also held back the European launch until the spring of 2002, missing critical winter holiday sales. In May, he said that if everything had gone right, Microsoft could have supplied all three of the largest markets - North America, Japan and Europe. But he felt it was best to play safe. Some software vendors expressed surprise that Mr Bach preferred to launch in Japan, where competition will be fierce against domestic rivals, rather than in Europe, a larger market.

In the event, the company has now delayed the Japanese launch until February. It did not fully explain the reasons for the decision, raising concerns about manufacturing. Microsoft said it wanted to concentrate on the North American market.

But Mr Bach admits that supplying the top-performing hardware will not be enough. "The key is software, not only at launch but through the pipeline going forward," says Mr Bach.

"We will sell every console we can make at launch. The problem is four to five months after launch. Will there be Xbox games out there exciting people that makes them want to buy the machine?" asks Seamus Blackley, Xbox advanced technology director.

To achieve that compelling pipeline of games, Microsoft has put a huge effort into getting software developers to commit to its platform. Sega's Dreamcast was hugely handicapped by the refusal of Electronic Arts, the largest independent publisher, to develop games for its machine.

Microsoft claims that 200 companies around the globe are developing Xbox games. These include Electronic Arts and Activision, the two top US publishers, Eidos in the UK, Infogrames, the leading European-based software group, and Sega, Capcom and Namco of Japan.

"We were the first company which went out there and actually asked the developers what they wanted," says Mr Bach. One outcome was that Microsoft chose standard rather than custom components for the Xbox, which makes it easier for the developers to write software for the machine.

In response to demands from the software companies, Microsoft also added online capability to the Xbox from the outset. This added to the cost of manufacturing, but should eventually provide larger and more reliable revenue streams for Microsoft and the software companies.

"The combination of a hard drive and broadband connection allows software companies to offer new levels and options that can be downloaded. It not only provides a more steady cash flow, but also saves on distribution and packaging costs," says Mr Bach.

"The Xbox has been designed with the designers in mind," says John Davison, editor in chief of Electronic Gaming Monthly.

"Our attention to the software companies means that the pipeline for the next 18 months is very full and should offer breadth - a lot of titles - with depth, some really high-quality games," says Mr Bach. In addition, the group has in-house developers to supplement third-party offerings.

Early reviews of games such as Halo, an in-house first-person battle game, and Oddworld: Munch's Oddysee [sic], a holistic, ecologically-friendly title, have been positive, although there has been some disappointment about the rate at which frames refresh, given the high hardware specifications. This should be overcome as the software developers become used to their development kits.

However, even if Xbox's software is compelling, Microsoft as a new entrant will be handicapped by the lack of popular franchises, insists Larry Probst, chairman and chief executive of Electronic Arts, the world's largest independent games developer. "Microsoft just doesn't have anything like Mario and Star Fox - franchises that are system sellers," he explains.

No matter how successful the launch, there is no doubt that Microsoft will initially lose money on the Xbox. Analysts' estimates of losses on each box range from between $100 and $150. Microsoft will not say how much it will initially cost to make the machine, which will retail in the US at $299, but Mr Bach does point out that it contains standard components. There are also economies of scale: if the machine sells in huge volumes then the cost of manufacturing will naturally fall.

When Microsoft expects to break even is moot. Henry Blodget, Merrill Lynch's celebrated internet analyst estimates it could take as long as five years for the company to recoup its investment - essentially the life-cycle of the product.

However, Mr Bach disputes this and says that some of Mr Blodget's assumptions are wrong. Not least, he says that Mr Blodget has not understood the possible revenue streams from online gaming. The Xbox comes with a built-in broadband connection, allowing players to compete with each other over the internet.

"Online gaming is a revolution. It is like when sound and colour were added to cinema," claims J. Allard, general manager for the Xbox and one of the team that persuaded Bill Gates, Microsoft's founder and chairman, to enter the games console market. "Over the life-cycle of the Xbox, online gaming will move from being a novelty to a necessity," he adds.

What is certain, is that Microsoft would have to do heroically to wrest market leadership from Sony. Kazuo Hirai, president and chief operating officer of Sony Computer Entertainment America, is dismissive of his American rival's efforts. "The battle in 2002 is already over and Sony is the winner. The struggle is for the number two spot between Nintendo and Microsoft," he says.

"Bill Gates knows he will not win in this generation of console," agrees Mr Probst at Electronic Arts. "But he wants to be in a position to win in the next round. He told me he will spend as much as it takes."

Mr Bach at Microsoft firmly denies his company does not care about losses. "Microsoft does not go into businesses that make losses for five years. I have my own profit-and-loss account," he insists.

"Xbox will not get 50 per cent market share this cycle," says Robert Kotick, chairman and chief executive of Activision, the second largest games software group in the US. "But Microsoft is a tenacious company. They are a version 2.0 company that eventually gets things right," he adds.

"The next generation of consoles in five years time will be truly revolutionary. It will be a profound shift, because the back office network architecture that will be required for online gaming will be expensive, the complexity unimaginable. And that will play to Microsoft's strengths," adds Mr Kotick.

It may be tempting for Microsoft to have low expectations for the Xbox, given the competition, but the danger is that such predictions can become self-fulfilling. Mr Bach believes he has put together the right strategy. Now he must execute it flawlessly. That would make Microsoft a version 1.0 company, in the games market at least.