The service is possible because it is run by Beijing Telecom, the only fixed-line telecommunication operator in the city. For it, the Internet is a value-added service, allowing it to earn more revenue from subscribers to its phone lines. It charges for Internet access by simply adding another item to a subscriber's monthly bill hours of Internet usage -- and charging for it. This gives it every incentive to boost the numbers of people going on-line, and is part of the reason for the dramatic increase in the numbers already starting to access the Internet.

However, to focus on the Internet is to miss the point. The Internet is neither the starting point nor the finishing point. It is simply the means to an ends. Indeed, to a variety of ends. It is, as we will repeat, simply infrastructure. True, much media coverage focuses on the Internet as starting point of China's embrace of the on-line world. But this is where much outside analysis of the potential impact of the Internet in China goes wrong, for it focuses neither on what the Internet is doing, nor where the Internet is going.

With Internet stocks reaching dizzy heights in the United States, it is perhaps not surprising that the Western media have picked up on China's potential Internet stars. Top of the list are Charles Zhang and Edward Zeng, who head up respectively Sohu, a Chinese-language search engine and Internet portal modelled on the USA's Yahoo!, and Unicom Sparkice, a Sino-Canadian venture with various Internet interests, the most visible of which is a small chain of Internet cafés. Sohu's investors include Intel and Nicholas Negroponte, while visitors to the Internet Café in Beijing's China World complex have included Bill Clinton.

But it is in fact the other way round. The Internet is built on the back of China's national information infrastructure -- a series of information networks rolled out over the last decade. Why have these networks been rolled out by the government? Why did the government create the environment for what now has become the "pandora's box" that is the Internet? Because that is what governments do -- that is, build infrastructure. And it should never be forgotten that China is still a socialist government, with a commitment to building infrastructure, and attempting to connect with all of its citizens, benignly or otherwise.

So why is this bit of infrastructure getting special priority? After all, roads, water, energy and sewers are all equally pressing requirements. Perhaps more so. One reason is demand. Having built it, people are using it, which creates a virtuous circle -- revenues are generated, which both encourages and allows more infrastructure to be built. As has been shown with mobile phones in particular, but also with fixed lines, China's telecom authorities are very happy to reinvest in their networks if they expect revenue to flow in immediately. Another -- and the most important in explaining why so much effort has been invested so far -- is that the government is driving its development, for its administrative purposes.


One of the most useful sources of basic data on the Internet in China are the surveys conducted every six months by ChinaNIC. The latest, carried out in June 1999, received more than 52,000 valid answers. It found that China's Internet users are predominantly young, well-educated and well-off. They spend a lot of time on -- 40% are on the Net for more than 10 hours a month -- and nearly half surf from their homes.

On-Lining Up the Numbers According to ChinaNIC (CNNIC), an organisation under the Chinese Academy of Sciences charged with overseeing the development of the Internet in China, by July 1999 the country had 4.0 million Internet users, up from 2.1 million just six months previously. There are good reasons to think this is a serious underestimate, and the total now could be twice this figure. According to research conducted by Maverick Research Limited in conjunction with New York-based Virtual China Inc, a more credible figure for the end of 1998 was 3.4 million users, set to rise to 5.5 million by mid 1999 and 10 million by mid-2000. The main reason for the difference in estimates appears to the low numbers of users per access assumed by CNNIC. In the past its multipliers have been 1.85 users per dial-up line and 3.96 users per leased line; it has also made no distinction between the different kinds of users -- whether they are in business, government or education. Surveys of different work places, however, have consistently found more people logging on from each account, especially for dial-up and leased lines in businesses and for leased lines used in government offices.

II. Top down: It's All About the Network >>