It was a volatile mixture from the start. An alliance between the University of Cambridge, with its reputation for academic purity, and Massachusetts Institute of Technology, renowned for the commercial enterprise of its staff and graduates.

Two more factors made the venture inherently risky: the Cambridge-MIT Institute (CMI) was set up with £68 million sterling (€111 million) of British taxpayers' money and also had the personal backing of Gordon Brown, the Chancellor of the Exchequer.

It is not difficult to see why Mr Brown, after visiting MIT's Boston campus in 1998, wanted to bring MIT's expertise to Britain.

A 1997 BankBoston report found that MIT-related companies employed 1.1 million people and enjoyed annual worldwide sales of $232 billion (€259 billion). Companies founded by MIT graduates include Hewlett-Packard, Rockwell International, McDonnell Douglas, Intel and Campbell's Soup.

While Cambridge's own successes at the heart of Silicon Fen are outstanding in British terms, the BankBoston MIT report had described an achievement in another league. If companies founded by MIT staff and graduates pooled their income, they would rank as the world's 24th-largest economy.

The MIT phenomenon chimed perfectly with New Labour's aspirations to build a new enterprise economy.

CMI - now in its second academic year - was set up to enhance the impact of research and teaching on economic success, primarily in Britain. It would conduct research, swap students and run professional practice courses in subjects such as biosciences enterprise. It would act as the gateway for new ideas for Europe. But the risks were as high as the potential prizes.

Inevitably, stories of a clash of cultures have emerged. Cambridge, it was said, had a low view of administrators generally - particularly those tasked with building links to industry. MIT, for its part, found Cambridge's attitude to closer links with business over-fastidious.

"MIT is not afraid to spend money on administrative areas or departments that might be frowned upon elsewhere," says Mr John Vander Sande, executive director of CMI in the US.

MIT's industrial liaison programme began in 1948. "We know what it takes to make the relationships with companies successful. We understand the resources required to do so."

Hence the well-resourced departments in corporate relations, alumni relations and technology transfer.

"The difference is that when they decide there is an administrative job to do, they build an engine that's big enough to do it - with enough dollars to make it work," says Prof Alan Windle, executive director of CMI in Cambridge.

MIT created posts for administrators rather than burdening academics with extra responsibilities. Things are changing fast at Cambridge, he says.

But does the university have a problem with attaching esteem to administrative jobs?

"It's not that people look down on administrators. Esteem comes from the fact that the university says it is so worthwhile, we will put the money in to make it work."

Mr Vander Sande identifies another big difference: government attitudes to funding research. Years of experience have allowed Washington to operate a "hands-off" approach to MIT's research programmes.

"The various US government agencies demand less paperwork from us because they are building on a tried and trusted relationship."

CMI, he says, is a new breed of institution that is being closely monitored by the British government, which has, in effect, acted as a venture capitalist.

He commends US attitudes to breaking down barriers to problem-solving. Government agencies help identify research goals, bring together resources, find the best academics and then fund a programme.

"This may be a model that the UK government could follow in the future - identifying a particular gap that needs to be addressed and then utilising the interface between government, industry and academia to work together."

Prof Windle believes Britain is changing rapidly in the way it seeks to tackle research. He points to co-operation between government and universities and within universities such as Cambridge, and increasing horizontal integration between subjects in research areas supported by the government.

"Take nanotechnology. You can't learn about it by trying to undo nano and technology. It's the greatest silo-breaker of recent years."

Fundamental institutional differences show themselves at the student level, too. Currently, 33 students from Cambridge are at MIT and 27 from MIT at Cambridge. "Another example of cultural differences is our contrasting attitudes towards the independence and maturity of our students," says Mr Vander Sande.

"We check that they are keeping up with the subject on a weekly, if not a daily, basis," he says - an example of a "fair amount of hand-holding".

Prof Windle says the Cambridge tutorial system underpins the university's quality but there have been changes, such as more course work. "Each system is very good. We are not trying to make Cambridge like MIT and they're not trying to make MIT like Cambridge."

But there is cross-fertilisation. MIT students do broad-based degrees in which scientists spend 25 per cent of their time on HAS (humanities, arts and social science) courses. MIT students have complained that Cambridge's courses are too highly specialised.

"We do 100 per cent physics after the third year. That's because we want to turn out the best physicists in the world," says Prof Windle. But Cambridge is designing HAS modules for students as a result of the MIT experience.

Do such differences matter? Both sides of the CMI venture agree that tackling such issues is not incidental but central to the success of the venture. Mr Vander Sande says "cultural differences" were one of the main motivations behind the venture. If the two institutions had been identical, "MIT would not have been interested in getting involved and I don't think Gordon Brown would have offered us a shilling," he says.

"There are differences in the institutions and that's what makes it worthwhile - there wouldn't have been so much point in doing it with a clone of Cambridge," Prof Windle says.

The early signs for CMI are bright. The leadership does seem able to distinguish small-scale differences - such as that of monitoring students - from large-scale factors such as a role for administration that will have a measurable effect on technology transfer.

CMI has also caught Cambridge at an ideal moment. The recent rapid modernisation of the university under Sir Alec Broers, its vice-chancellor, has given it a pivotal position in industry/academia links in Britain.

As such, it was an ideal choice for a British gateway - even if it raised hackles at other research-led institutions.

Today, a CMI summit in Cambridge announces a series of research breakthroughs with wide commercial impact. But on a more fundamental level, good practice is being disseminated to Britain's higher education sector.

Take the role of universities in technology transfer. The MIT experience is that universities should measure themselves on the number of companies started by staff and graduates. "What's been said from the States is: don't try to get rich on a licensing income stream," says Prof Windle.

He says there is a need for universities to develop a "greater-good concept" and hope, long term, to benefit from equity stakes. The culture should be: "We do this not to get rich; we do this because it is part of our mission."

There is also a palpable excitement at CMI. A business plan is already being drawn up for 2006-10. CMI, in some form, will move beyond the stage of needing public money as "seed corn" and will start paying its way.

But Prof Windle says the real buzz is from intellectual co-operation. He recalls a workshop at MIT during which someone excused themselves in order to file the content of the conversation as a patent before returning to the meeting.

"We hadn't even started the project," he says. "No money had been spent except the airfare and out was popping a patent. If you get the right people in the right room, amazing things can happen."