On he shelves of Bruce Dunlevie's office in Sand Hill Road, the epicentre of Silicon Valley's venture capital industry, stands a small forest of plastic plaques, trophies from his career as general partner at Benchmark Capital.

The plaques commemorate scores of initial public offerings by the firm's clients, including what Mr Dunlevie calls Benchmark's "slam-dunks" - eBay, Ariba and Juniper Network.

"There is some kind of alchemy to venture capital," says Mr Dunlevie, gesturing to his mementos. And indeed, for a moment in the late 1990s, Benchmark appeared to have discovered the Philosopher's Stone. Its 30 per cent stake in eBay, the online auction company, cost just $6.7 million (€7.3 million) and generated a return in two years of more than 100,000 per cent - which makes it the most successful Silicon Valley investment ever.

Next to Mr Dunlevie's eBay plaque sits an empty jeroboam of vintage Dom Perignon, but these days the champagne on Sand Hill road has gone flat.

The apparent ability of Benchmark and other venture capital groups to conjure technology into unimaginable wealth prompted investors to increase their funding of US venture-backed companies from just $11.6 billion in 1996 to $98 billion last year, according to Venture Economics, a research company that tracks the industry. Most of that money now seems likely to be lost.

"We are looking at huge capital destruction and human tragedy," argues Chris Nawn, a partner at Technology Crossover Ventures in Palo Alto.

The venture capital industry is facing a cash crunch, with investors increasingly reluctant to put money into the sector, while companies in which they have invested burn cash.

Mr Nawn expects many venture capital groups to go under, the bulk of client companies to fail and investors to lose huge sums.

If that happens, the ramifications could go far beyond the venture-capital industry and the privileged world of Silicon Valley. There is a danger that technological innovation, which provided the productivity growth underpinning the US economy during the 1990s, could falter. Venture capital has provided a bridge between the private and public markets, ushering in new companies and technologies.

The scale of the impending crisis should not be underestimated, warns Tae Hea Nahm, a partner at Storm Ventures in Palo Alto. "We are facing a bloodbath. There will be an $80 billion shortfall in funding during the next 12 months alone. If you look at just one sector, the optical space, there are 300 companies that will need between $20 billion and $40 billion to sustain them. The money is just not out there."

Mr Nawn calls the first quarter of this year "a nuclear winter. Nobody wanted to invest."

Since then, the industry's plight has deteriorated further, according to data released this week by PricewaterhouseCoopers (PwC) and VentureOne.

The figures show investment in venture-backed companies down from $10 billion in the first quarter to $8 billion in the second. That compares with $24 billion invested in the same quarter last year.

Investors who have lost money are increasingly unwilling to commit more capital to start-ups.

"Look, people were setting up dotcoms not just for food, but dog food, and not just dog food, but Scottish terrier dog food. No wonder investors feel disillusioned," says Mr Dunlevie at Benchmark.

"Quality control went right out of the window. It was a function of volume. There was too much money out there," says Mr Jim Seaberg, a principal at consultancy Mckinsey & Co.

The market for initial public offerings is now moribund. Only 13 venture-backed companies have gone public this year, according to Mr Tracy Lefteroff, global managing partner of the venture capital practice at PwC.

"It could be that companies now go to market not after 12 months but after three or four years," says Mr Brad Koenig, managing director of west coast investment banking at Goldman Sachs, "There are no companies out there that fit the profile. The discontinuity is causing real pain among the venture capitalists."

The other way of generating a return - a trade sale - is also blocked. "The technology economy is absolutely in recession," says Mr Philip Sanderson, a general partner at Walden VC, based in San Francisco.

Venture capital groups traditionally tried to bring in late-stage investors to support companies until they were ready to go to market. But such investors are demanding increasingly tough terms. "Negotiations are really fraught," says Mr Shahan Soghikian, a partner at JP Morgan Partners, a late-stage investment group.

Some funds have handed back cash to their investors rather than risk the destruction of further value. In April, Octane Capital Management returned $130 million to investors. But most venture capitalists appear determined to struggle on and are having to cull their portfolios.

"We've spent the past three quarters in triage," admits Mr Rosemary Remacle, a venture capitalist at Sevin Rosen in Palo Alto. "We've been selling, cutting and shutting down."

This carries its own risks for the surviving companies financed by venture capital.

Some venture capitalists talk of companies being closed down, even though they had viable operations, because their cash balances were regarded as being more valuable. "You don't want too little cash but you don't want too much, because you might get raided by your VCs," complains one entrepreneur.

Given the shortage of funds from outside investors, some venture capitalists have taken to investing more of their own money in the companies they think will make it. But that could create risks, says Ms Remacle.

"A $200 million write-off can crater a fund and there are already a few of those out there."

The ramifications of the cash crunch for early-stage companies and innovation are bleak.

"VCs are in crisis mode, spending 70 per cent of their time rescuing their existing companies rather than making new investments," says Mr Seaberg at McKinsey.

That effect could be seen in the PwC/VentureOne data: investment in first-round, early-stage companies fell during the second quarter from $8 billion to only $1 billion.

The implications for many of the 3,000-odd US-based venture capital groups are also grim. For the moment, the sector appears calm because none have to go public with their pain.

"The suffering is under the surface. The industry is fragmented and far from transparent but there is huge carnage in the portfolios and the day of reckoning is coming," says Mr Koenig.

"This year venture capital companies will start to die. They will simply run out of cash," says Mr Nahm at Storm Ventures,

"Eventually, there will be between six and 10 venture capital companies of any size."

However, he is convinced that this is just a readjustment to a market growth rate.

"In the long term, applications drive technology and more applications will continue to be invented. This downturn is a market phenomenon, not a technological phenomenon."

Others are less sanguine. A McKinsey analysis of the top 100 venture capital groups between December 1990 and December 2000 showed a median return on investment of only 13.4 per cent, compared with 12.1 per cent for public equity funds.

For much of the 1990s they under-performed the equity markets, only to pull ahead during the rapid growth of the end of the decade. "That is not much of a higher return given the risks involved," says Mr Seaberg.

These days, the venture capitalists of Sand Hill Road are guarding their cash and hoping a few successful investments will see them through a difficult period. It may have seemed for a short time that some had discovered the secret of alchemy. But the magic has very definitely worn off.