There was no single moment of invention
The internet is often described as if someone invented it in a particular year. The record is more tangled than that. It grew out of several research efforts in the United States, Britain and France, a shared set of rules for linking networks, and then a long period of commercial and public adoption. Credit is spread across many people, and historians still disagree about how to weigh individual contributions.
A useful way to read the story is in four stages: the ideas of the 1960s, the working research network of the 1970s, the standardization and growth of the 1980s, and the opening to the public in the 1990s.
The 1960s: ideas before machines
In the early 1960s, computers were large, expensive and usually used one job at a time. Time-sharing systems let several people use one machine from separate terminals, and that raised a natural question: could the machines themselves be connected? J.C.R. Licklider, who worked at the US Defense Department’s Advanced Research Projects Agency (ARPA), wrote memos around 1962 imagining a network of computers that researchers could use to share programs and data.
At the same time, two groups were working on how to move data efficiently. Paul Baran, at the RAND Corporation, studied communications networks that could keep working if parts failed. Donald Davies, at the National Physical Laboratory in Britain, independently developed the idea of cutting messages into small blocks, which he called packets. Leonard Kleinrock’s work on queueing theory gave mathematical support to the approach. The core technique, packet switching, is now the foundation of nearly all data networks.
1969: ARPANET begins
ARPA funded a network linking research computers at universities. The contractor Bolt Beranek and Newman built small dedicated computers, called Interface Message Processors, to sit between each site and the network. The first link connected the University of California, Los Angeles, and the Stanford Research Institute in late October 1969. By the end of that year, four sites were connected, including the University of California, Santa Barbara, and the University of Utah.
The 1970s: more networks, and a common language
ARPANET grew through the early 1970s, and some of its most lasting uses were not planned in detail. Electronic mail appeared in 1971 and soon accounted for a large share of traffic. Other networks appeared as well: a packet radio network, a satellite network, Ethernet at Xerox PARC, and the French Cyclades project led by Louis Pouzin.
Each used different rules, so connecting them was a new problem. Vinton Cerf and Robert Kahn published a design in 1974 for a protocol that let different networks exchange data through gateways. It was later divided into two parts, TCP and IP, which together became known as TCP/IP. The word “internet” was coming into use for such a linked collection of networks. How that design works is the subject of the idea of a network of networks.
The 1980s: standardization and growth
On January 1, 1983, ARPANET switched over to TCP/IP, a date sometimes called its flag day. The Domain Name System, introduced in the mid-1980s, replaced hand-maintained host lists with names such as those ending in .com or .edu. Universities and research bodies built their own networks and joined them to the larger system. The US National Science Foundation funded NSFNET, a faster backbone that linked supercomputing centers and regional networks and became the main academic route by the late 1980s.
Local networks were spreading at the same time. Ethernet became an IEEE standard in the early 1980s, and organizations that had local networks wanted to join them to wide area networks. The original ARPANET was shut down in 1990, by which time it had been replaced in practice by a larger system.
The 1990s: the web and the public
For most of its first two decades, the internet was used mainly by researchers and technical staff. That changed with the World Wide Web, proposed by Tim Berners-Lee at CERN in 1989 and put into operation in 1990 and 1991. Graphical web browsers, especially Mosaic in 1993 and Netscape Navigator in 1994, made it simple for non-specialists to browse pages.
Commercial traffic had been restricted on the NSFNET backbone, but those limits were relaxed in the early 1990s, and the backbone was handed to private providers in 1995. Dial-up service providers sold access to households. Online shopping, search engines and free email services appeared during the second half of the decade, followed by a period of heavy investment and then a sharp correction around 2000.
After 2000: broadband and mobile
Telephone-line modems gave way to broadband over cable, DSL and, increasingly, optical fiber, which carries the long-distance backbone of the modern network. Wi-Fi spread through homes and offices. When smartphones arrived in the late 2000s, the internet became something many people carried and used constantly, and mobile devices now account for a large share of web use.
Later developments built on the same base. Cloud computing moved storage and processing into large data centers, and the Internet of Things attached sensors and household devices to the network. These devices brought new security worries, and the field of cybersecurity expanded as more of daily life depended on connected systems.
Who ran it, and who pays
The network has never had a single operator. Early on, the US government paid for most of it through research agencies. As commercial providers took over in the 1990s, ownership spread among telephone companies, cable companies and specialist carriers, each running part of the infrastructure. Technical coordination is handled by open bodies such as the Internet Engineering Task Force, while names and numbers are managed by organizations including ICANN. The arrangement works through agreements among thousands of independent networks, not through a central command, and it has changed repeatedly as the network grew.
What the history shows
Several choices made early on explain the network’s growth. Data was split into packets rather than carried on dedicated circuits. Standards were published openly, so anyone could build compatible equipment. And the protocols joined networks without needing a central owner, so new networks could attach without permission from a single authority.
Dates in this history are often given more neatly than the evidence allows, and different sources emphasize different milestones. Treat any claim that one person or event created the internet with caution.