The Internet shake-out is getting brutal. Last Friday, Eve.com, the US online beauty products retailer, closed down. A few days earlier, Boxman, the European compact disc retailer, and Priceline Webhouse Club, the US name-your-price site for grocery and petrol sales, joined the casualty list.

Urbanfetch has closed its online delivery service, and other recent victims, such as HomeGrocer.com and Petstore.com, have been driven into the arms of rivals.

As dot.com companies burn up their cash and fundraising opportunities evaporate, it seems certain the rout will continue. But less clear is where it will stop. Could it end with just one Internet leviathan dominating the business-to-consumer market? A Yahoo.Ebay.Amazon.com that serves all the consumer's needs?

The Internet is often regarded as the most open and entrepreneurial market in history. In theory, there are few barriers to entry: almost anyone can start a business, without the need for large amounts of capital or a big factory. So, as some companies die, others should spring up in their place, offering new features, better service or lower prices.

Yet, as the business-to-consumer market develops, there are signs that it is becoming an online version of the winner-takes-all society, one in which two or three companies at the top take a vastly disproportionate share of the market, making it difficult for others to thrive.

The consolidation is one such sign. Another is the way that advertising revenues for many dot.com companies, the main or only source of income, are disproportionately favouring the most popular sites.

According to latest figures from the Internet Advertising Bureau, an industry group, 71 per cent of US dot.com advertising revenues are going to the top 10 sites, 83 per cent to the top 25 and 91 per cent to the top 50.

The bureau does not say which sites are getting most of the revenues, but it is not much of a mystery. The advertisers want to be where the eyeballs are, which overwhelmingly means the big-name portals such as Yahoo!, MSN and America Online.

"The big are getting bigger," says Tom Hyland, chairman of PricewaterhouseCoopers new media group in New York, which conducts the bureau survey. Sites that attract the most viewers get the most advertising revenues, he says, which in turn means they can offer more services and attract even larger audiences. "It's a natural evolution in the industry that will continue to play out."

In the old economy, the business world has already become familiar with the winner-takes-all phenomenon. It results from the greatly increased competition brought by globalisation.

When markets were small, separate and nation-based, there was limited competition across borders. Even big, multinational companies tended to have separate headquarters and factories in each country, so their size did not give them much of an advantage over smaller companies.

Globalisation, together with improvements in transport and communications, has changed that by allowing big companies to treat large regions, or even the world, as a single market. This has given them economies of scale in production, distribution, marketing and management that smaller companies cannot match.

You see the result on supermarket shelves. There are far too many consumer products for the space available, so retailers give priority to the biggest brands. This further increases the sales of these brands, giving their manufacturers even greater economies of scale and so on, in a virtuous circle.

Often, it is not the best product that becomes the best seller. But scale, once established, is difficult to overthrow. JVC's VHS video recording system was regarded as inferior to Sony's Beta, but Beta is now defunct. And in blind taste tests, most people prefer Pepsi-Cola to CocaCola, but Coke outsells Pepsi.

Recognising this, old-economy companies have been scrambling to establish scale before it is too late. Almost every industry, from advertising to telecommunications, is in a race for global domination. According to Thomson Financial Securities Data, the value of mergers and acquisitions announced last year rose 24 per cent to a record $3,029 billion (#3,602 billion).

If scale is important in the old economy, it is even more significant in the new. This is because the products of the "weightless" economy are more often intangible information or ideas than tangible objects that are manufactured and distributed.

In the old economy, a company can benefit from scale by spreading its fixed costs over a larger output, but its variable costs will increase as sales go up. In the digital economy, a company may incur high costs in developing and marketing an idea, but because its output is intangible, the incremental cost of distributing it to a wider customer base may be zero, yielding vast economies of scale.

Another aspect of the Internet's scalability is the so-called network effect, which holds that products or services become disproportionately more valuable as more people use them.

If people want to go to an auction website, for example, they want the one that attracts the most users: that way, buyers get the widest choice of products, and sellers get the most bidders. So there is a natural tendency towards monopoly and an almost insurmountable barrier to new entrants.

The same could be said of Amazon.com, which makes much of its ability to recommend products based on other people's purchasing habits ("Customers who bought this book also bought . . ."). The more people who use the site, the better these recommendations become, giving Amazon.com an advantage over retailers with fewer customers.

To get down to practicalities, it is becoming increasingly uncontroversial to argue that, in the business-to-consumer sector, only a handful of the very largest dot.com companies have the scale or strength to survive in their present form. Despite squandering billions of dollars on marketing, almost no dot.com companies have succeeded in building an enduring brand and, lacking either that or a profitable business, they face the likelihood of imminent oblivion.

Forrester Research, the Internet research company, has already forecast that most dot.com retailers will be driven out of business by next year. And Mr Robert Lessin, chairman and chief executive of Wit SoundView, a US boutique investment bank that specialises in the Internet and technology sectors, foresees the survival of only three or four dot.com companies in the business-to-consumer world.

According to Mr Lessin, the Internet will overwhelmingly end up as just an alternative channel of distribution for "real" corporations, because they have the brands, the infrastructure, the expertise, the customers, the financial resources and all the other things dot.coms lack.

"The fact that most of these corporations did not move fast enough to position themselves for the Internet is irrelevant, because they now have a second bite at the apple," Mr Lessin says. "It's make versus buy - either they can buy a truly troubled dot.com operation at a very depressed price, or they can build it themselves. But most corporations have lost nothing by not becoming a force on the Internet three years ago."

When the shake-out is over, will there ever be another Internet start-up in the business-to-consumer sector?

"When I invest the first question I ask of a prospective company is `Who is your physical analogy? Who are you like in the real world?' And if they give me an answer to that, I won't invest," Mr Lessin says.

In other words, there may be no money for start-ups unless they are like nothing else on earth. And they will probably need to be unlike anything else on the Internet, too. Who, after all, is going to finance a fight to the death with Yahoo!, Ebay or Amazon.com? Or a Yahoobayzon, if that should emerge?