No single person invented the internet. It emerged between 1961 and 1989 through the work of hundreds of researchers, funded largely by the U.S. government, with four contributions that mattered most: J.C.R. Licklider's 1962 vision of a "galactic network," packet switching proven by Donald Davies and Paul Baran, the ARPANET built by Bob Taylor, Larry Roberts, and Bolt Beranek and Newman in 1969, and the TCP/IP protocols Vint Cerf and Bob Kahn designed in 1973. Tim Berners-Lee added the World Wide Web in 1989.

The question "who invented the internet" hides a category error. The internet is not a device with an inventor but an agreement — a set of protocols that let independently owned networks pass data to each other. Understanding how that agreement was reached, and why it required so many people, explains both the history and why the internet still works the way it does.

Packet switching: the mechanism that made networks possible

Before packet switching, telephone networks worked on circuit switching. A call opened a dedicated physical path between two phones and held it for the entire conversation. That design wasted capacity whenever nobody was talking, and it was fragile: destroy one switching office and every call routed through it died.

Packet switching breaks a message into small chunks called packets. Each packet carries a destination address and travels independently. Routers read that address and forward the packet along whatever path is available at that moment. The receiving computer reassembles the packets in order.

The insight that made this work was statistical. Voice traffic is bursty, so a network sized for average load rather than peak load can serve far more users at once. Paul Baran at the RAND Corporation worked out the mathematics of survivable packet networks for the U.S. Air Force starting in 1960. Donald Davies at the UK's National Physical Laboratory independently coined the term "packet" in 1965 and built a small local test network in 1970.

The counterintuitive detail: the internet's core design assumes the network is unreliable. Packets get lost, duplicated, or arrive out of order, and the endpoints — not the routers — fix the problem. This "end-to-end principle" is why the internet scaled to billions of devices without a central controller.

From vision to working network: 1962 to 1969

J.C.R. Licklider, a psychologist at MIT, published a series of memos in 1962 describing an "Intergalactic Computer Network" — a set of computers that could share data and programs. He then funded the research that made it real. As head of the Information Processing Techniques Office at the U.S. Advanced Research Projects Agency (ARPA), he directed money toward time-sharing and networking research.

Bob Taylor took over the office in 1966 and pushed for an actual network connecting ARPA-funded computers. He hired Larry Roberts to manage the project. Roberts wrote the specification, and in 1968 the contract went to Bolt Beranek and Newman (BBN), a Cambridge, Massachusetts firm. BBN engineers built the Interface Message Processors — the first routers — and the first ARPANET link went live between UCLA and Stanford Research Institute in October 1969.

A second node at UC Santa Barbara and a fourth at the University of Utah followed within weeks. By December 1969, four computers were talking. The first message sent over the network was "LO" — the system crashed after the letter "L" and "O" of "LOGIN" went through. The full story is less heroic than the legend.

The protocol layer: why TCP/IP is the real invention

By the early 1970s, ARPANET worked, but it was one network. Connecting different networks — satellite links, radio networks, ground cables — required a common language. Vint Cerf and Bob Kahn published "A Protocol for Packet Network Intercommunication" in 1974, introducing TCP. The design split into two protocols later: TCP handles reliable delivery and ordering, and IP handles addressing and routing.

The 1983 flag day, when ARPANET switched from the older NCP protocol to TCP/IP on January 1, is the date most historians treat as the internet's true birthday. The internet, as distinct from ARPANET, became real when TCP/IP let any network join.

A professional caveat that casual histories miss: the internet's foundational protocols were developed on government contracts with public funding. The commercial internet only opened in 1995, when the National Science Foundation lifted restrictions on commercial traffic on its backbone. The "inventors" were mostly academics and government-funded engineers, not entrepreneurs.

Timeline: how the pieces arrived

Year Event Who
1962 "Galactic network" memos describing networked computers J.C.R. Licklider
1965 Independent packet-switching design and the term "packet" Donald Davies (NPL)
1969 First four ARPANET nodes operational BBN, Roberts, Taylor
1973 TCP design published; initial specification Vint Cerf, Bob Kahn
1983 ARPANET switches to TCP/IP ARPANET community
1989 Proposal for the World Wide Web Tim Berners-Lee (CERN)
1991 First public web server and browser CERN
1995 NSF lifts commercial restrictions on the backbone U.S. government

The pattern across that table is worth stating plainly. Each row solved one layer of a problem the previous row exposed. Vision led to funding, funding led to a network, the network needed a protocol, the protocol needed applications, and applications needed an easy interface — which the web provided.

Common misconceptions, checked against the record

The most persistent myth is that Al Gore claimed to have invented the internet. In a 1999 interview he said he "took the initiative in creating the internet," referring to his legislative work funding high-performance computing and networking in the 1980s and 1990s. He did not claim to have built it. The misquote spread because it made a better story than the funding record.

A second myth credits the internet to a single country or agency. The packet-switching concept appeared independently in the U.S. and UK within a few years. The web was invented at CERN in Switzerland by a British physicist. The internet's technical lineage is multinational by design.

A third myth treats ARPANET and the internet as synonyms. ARPANET was one network. The internet is the set of rules that let many networks interoperate. The distinction matters because it explains why no single organization owns the internet today.

Red flag: any source that names one inventor of the internet with a single date is oversimplifying. Red flag: claims that the internet was built as a military command-and-control system are misleading — ARPA funded it, but survivability was one design goal among several. Red flag: confusing the web with the internet. The web is an application; the internet is the transport layer beneath it. Red flag: attributing the first email, first search engine, or first website to a single person without checking publication dates.

Practical actions: how to verify internet history yourself

If you want to check the claims in this article rather than take them on faith, specific steps will get you there.

Read the primary sources. The Internet Archive hosts the original RFC documents, including RFC 675 (1974) and RFC 791 (1981), which describe TCP and IP in the authors' own words. Check dates against the Internet Society's published timeline and the Computer History Museum's oral histories with Roberts, Cerf, and Kahn. Distinguish the layers. Trace packet switching (1960s research), ARPANET (1969), TCP/IP (1974 design, 1983 deployment), and the web (1989 proposal) as separate milestones. Treat "first" claims with care. Many early networks ran in parallel, and the record often shows independent invention, not a single origin. Avoid undated secondary sources. A history that names no years or sources cannot be checked and should not be cited.

Threshold for trusting a source: it names the primary document, the author, and the year. If a claim fails all three, treat it as folklore.

Limitations and what the record cannot settle

Historians still debate who deserves credit for the first working packet-switched network. The NPL network in the UK started running in 1970, after ARPANET's first link but before its full deployment. Bidirectional credit is more accurate than a ranking.

There is also a gap between invention and adoption. TCP/IP was technically ready in the late 1970s but took until the 1990s to reach ordinary users, and the gap between the 1989 web proposal and mass adoption was roughly five years. Invention and impact are different questions.

A final limitation: the record favors documented, English-language, institutionally funded work. Contributions from early commercial networks, university labs, and non-U.S. researchers are less well archived, so any single narrative understates the true count of contributors.

The internet has no inventor because it was never one invention. It was a layered agreement built over three decades by people who each solved the problem in front of them — and that is the most useful thing to know about it.