NATIONAL SERVICE? The point, says Amnon Landan in his guttural Israeli accent, is that joining the army can be the best preparation of all for launching a business.

I think he’s kidding. Not at all, he says unsmilingly. If you make it to officer, you get to lead people, you learn how to motivate them and you gain experience of decision-making in uncertain conditions.

And he should know. A former Israeli paratrooper, he now heads Mercury Interactive, one of America’s hottest technology companies.

Last autumn Forbes magazine voted him America’s entrepreneur of the year. Later this month Mercury will report full-year revenue figures that could reach $500m (£280m) for the first time. Some think Landan’s firm could be the next Oracle.

And you have to give it to him, it’s not every day you meet an entrepreneur who claims that the Israeli Defence Force is as good a business school as Harvard or London. But then again, maybe nothing in technology should surprise us any more.

Landan’s firm makes its money by providing the expertise to test and debug other businesses’ software. It is a booming market. Up to 50% of every business’s IT budget is wasted, says Landan: stuff doesn’t work properly, nothing fits together, and co-ordinating new purchases with all the software and equipment your firm is running is a nightmare. Into the breach steps Mercury, selling tools that automate the testing of new applications before they are run, and measuring performance while they are running. It’s that simple, at least in theory.

For laymen like me, though, this is what you call a head-scratching moment. Shouldn’t the software bought by big companies work in the first place? “It’s not about other people’s mistakes,” says Landan. It’s about the speed and complexity of modern business, the amount of software, middleware and hardware that IT departments have been buying, the way everything is always moving. “Someone somewhere is always plugging in another box. It’s nothing to do with problems in the applications, it’s to do with the complex environment.”

Mercury was set up in America 15 years ago with a research department in Israel, where universities churn out high-calibre computer-science boffins prepared to work for half the pay expected in Silicon Valley. It has cleaned up after spotting its niche — “business technology optimisation” — way before anybody else.

And Landan, hired to run the research and development department before becoming chief executive in 1999, has driven the firm’s phenomenal growth — about 25% a year in revenue and profits, even when the sector is in recession. As a result, he is being touted as a future Larry Ellison (the high-profile boss of Oracle).

Not, you suspect, that all this would impress him. Sitting in a hired interview suite in central London, having jetted in to press the flesh with Mercury’s growing band of European customers, Landan, 45, exudes calm purpose. Compact, stern, with hooded eyes and a tomahawk face that shows little emotion, he shrugs off other people’s plaudits.

“Ach, I can retire now, sure,” he says, “but I don’t want to.” Friends report that he is famously unflashy, renting his house in California until only a few years ago and working out of modest offices. What he does want, though, is size.

“We’ll soon be north of $500m revenues. We will hit a billion in three years, especially with a bit of acquisition on the side. We have $1.3 billion in cash, so we don’t need to raise any. The aggressive goal we have set ourselves is to become a top five software firm.”

That means topping $3 billion in revenues and starting to challenge the giants: Microsoft (“although they are not really a software firm”), Oracle, SAP and Computer Associates. It’s a tall order, but Landan believes the only threat to Mercury is itself.