What “commercial” means
A computer built for a single laboratory is one thing. A design that a company builds again and again and offers to customers is another. The second kind appeared around the end of the 1940s and the start of the 1950s, and the question of which one deserves the label “first commercial computer” has no single settled answer. It depends on whether you count orders taken, machines delivered, or machines used for business work rather than science.
Several early contenders
A few machines are commonly mentioned, usually with caveats.
- Ferranti Mark 1 (Britain). Developed with the University of Manchester, based on the university’s earlier prototype work, and delivered to the university in 1951. It is often described as one of the first computers sold from a production run.
- UNIVAC I (United States). Designed by the company founded by J. Presper Eckert and John Mauchly, the people behind ENIAC. The first one went to the U.S. Census Bureau in 1951, and later units went to other government agencies and businesses.
- LEO I (Britain). Built by J. Lyons & Co., a catering and food company, which based the machine on the Cambridge EDSAC design. It began running routine business jobs around 1951. It was unusual in being built by a customer for its own needs, and the company later sold the design.
- BINAC and early special orders. Eckert and Mauchly’s firm delivered a smaller machine around 1949, and European makers such as Zuse in Germany supplied machines to universities in the early 1950s.
Because each of these fits a different definition, many historians prefer to say that commercial computing began in the early 1950s on both sides of the Atlantic rather than naming one machine.
Who bought them
The first customers were mostly organizations with large, repetitive calculations and the money to pay for equipment that was expensive and demanding. Government agencies such as the Census Bureau were natural buyers, as were defense and aviation organizations. Insurance firms and manufacturers were among the early business users, and a few companies, including Lyons, used machines for payroll, stock records and similar office tasks. The idea that a computer could be a general business tool rather than a scientific instrument was still new, and some firms bought mainly to learn what the technology could do.
These machines were not like later office equipment. They filled rooms, needed trained staff and air conditioning or other careful housing, and were programmed in machine code or close to it. The people who ran them were part of a small and new profession, and many were women, as with the early programmers of the electronic era.
Technology of the period
The first commercial computers were built mostly with vacuum tubes. Storage and input varied by design:
- Mercury delay lines, or similar early memory, held working data. Magnetic core memory spread later in the 1950s.
- Magnetic tape drives gave large and relatively fast storage for records. UNIVAC was well known for using tape.
- Punched cards, already familiar from earlier tabulating equipment, handled much of the input and output. Their history overlaps with that of punched cards generally.
Reliability was a constant problem. Machines often spent a share of each day being maintained, and results were checked carefully.
A famous publicity moment
UNIVAC became widely known in the United States in 1952, when it was used by a television network to help forecast the presidential election. Accounts say its early prediction pointed to a lopsided outcome that the broadcasters were reluctant to trust at first. Details vary between retellings, but the episode is often credited with bringing computers to the public’s attention, even though most viewers still had no idea how such a machine worked.
IBM enters
IBM had long sold punched-card equipment and was initially cautious about electronic computers. In the early 1950s it built the IBM 701 for scientific customers, then moved into business-oriented machines. Its IBM 650, introduced in the mid-1950s, became popular partly because it was smaller and less costly than the largest systems. Over time IBM’s existing sales network and customer relationships helped it take a leading position, which led into the era of mainframe computers.
What was still uncertain
Early forecasts about the market were modest. Some people thought only a small number of machines would ever be needed, and the technology was still changing quickly. Cost, reliability and the difficulty of programming slowed adoption. Business computing grew only gradually, helped later in the decade by languages such as COBOL language and FORTRAN, and by transistors replacing tubes.
The 1950s computers were therefore a mix of experiments and early products. The first commercial machines mattered less for their individual specifications than for showing that there was a market, and that organizations would pay to have computation done as a routine service. That shift set the stage for the wider spread of business computing in the following decades.