The Heritage Foundation

Executive Memorandum #255 on Energy and Environment

November 2, 1989

November 2, 1989 | Executive Memorandum on Energy and Environment

A Billion Dollar Supercomputer Boondoggle?

(Archived document, may contain errors)

11/2/89 255


In a report released in September, Director of the White House Office of Science. and,Technol- ogy Policy (OSTP) Alan Bromely calls for a "...fe derally coordinated government, industry, and university..." program to develop a futuristic supercomputer network. This network would link computers in government agencies, universities, and private businesses, allowing information to be obtained and exc h anged at high speed. The OSTP urges creation of a "government-business" partnership to develop a system of powerful new hardware and larger storage capacity, and the software to run the system.The price tag: as much as $1.9 billion by 1996. Such a governm e nt-sponsored network would be a costly mistake. If a system is needed, the private sector can develop it more efficiently, without taxpayer assistance. Worse, government involvement could hinder the private development of such a system.The Bush Administra t ion recently rejected a similar proposal to spend over a billion dollars to develop a high definition television to compete with Japanese products. The Administration should reject this project as well. The OSTP argues that a complex network of supercompu t ers will make the U.S. more competitive by allowing government agencies, and later the private sector, to transmit, receive, analyze, and store massive amounts of information. The idea is for government funds to be provided to such companies as Internatio n al Business Machines Corp. (IBM), Convex Computer Inc., and Cray Research Inc. to develop a new range of supercomputers to establish the system, together with the software needed to link up the system. Under the plan, the government initially will own and operate the system, but eventually it would be privatized, with access offered to customers for a fee. Pre-empting Private Firms. Despite the attractiveness of such a national system, it still is, not- - clear it is needed. Numerous information distributi o n networks and data banks already exist. These are available to subscribers for a fee - and were developed by private enterprises, not by government. For example, the Nexis, Lexis, and Mexis data services, owned by Meade Data Central, Inc., provide comput e rized general, legal, and medical data. While these services are not directly comparable to the proposed OSTP system, they do suggest that if the customer demand is there for a larger system then private companies, or a consortium of hardware and software producers, likely would develop the system on their own. In fact, some companies seem to be interested in such a network. For example, IBM has several projects underway, in cooperation with other private firms, for the research and development of supercom p uter technology. The OSTP plan assumes that government knows better than private businesses what customers want and which companies and technologies can best develop goods and services. But there is no evidence that government bureaucrats know best - and good reason to believe they do not. In

addition, telecommunications firms, including American Telephone and Telegraph Co. (AT&T), MCI Communications Corp., and U.S. Sprint Communications Co., complain that their plans to expand their information services will face unfair competition if the government, using taxpayer funds, subsidizes other firms. Big Brother Computer? The danger also exists that a government-sponsored system would give federal agencies too much access to private information concerning Am e ricans. Attempts by the federal government to create a national data bank have been rejected in the past by policy makers because of this concern. A computer network in government hands could enable government to gather information on citizens, gleaned fr o m the many sources logged into the system. America currently is the world's leader in advanced computer technology and software. If the federal government wishes to promote the development of an advanced computer network, and if there is a genuine public d emand for such a network, removing barriers to entrepreneurship and innovation would be the best way to foster creation of a system. For example, reforming the outdated antitrust laws would allow companies to pool resources to assemble the capital needed f or research and development. Similarly, lowering business taxes would encourage businesses to undertake the commercial risks inherent in such a project. If it is needed, the proposed supercomputer network can be developed by private companies without gove r nment assistance. If it is not needed, the network would be a multi-billion dollar federal boondoggle financed by the American taxpayers. One of the great benefits of America's free market system is that it allows customers, suppliers, and investors to ma ke these decisions. The Bush Administration should let the market reach its own decisionon a supercomputer network. Bryan T. Johnson Research Associate


About the Author