The History of Modems

from 1837 to today

Epoch I

Telegraph, Theory & the First Signals
1837 – 1950

• 01Home
• 02History
• 03List of modems
• 04Technology
• 05Stories
• 06Info

← Back to History  /  Epoch I

The Telegraph: Information Becomes Electrical

On 24 May 1844, Samuel Morse sat in the U.S. Capitol building in Washington and tapped out four words in his code: What hath God wrought. The message travelled instantly along 61 kilometres of wire to Baltimore. A new age had begun — not just of communication, but of encoding.

Morse's genius was not the wire or the battery. It was the idea that information could be reduced to a small alphabet of symbols — dots, dashes, and silences — and that these symbols could be mapped onto electrical pulses. This is, at its core, exactly what every modem ever built has done. The vocabulary changes; the principle does not.

The commercial telegraph networks that spread across continents in the 1850s and 1860s were the first data networks in history. By 1866, a working transatlantic cable connected Europe and North America. Messages that had taken weeks by ship now arrived in minutes. The world had shrunk, and engineers were already asking: how much faster can we go?

Baudot and the Teletype: Speed Through Encoding

Morse code had a problem: it required a skilled human operator at each end. Speed was limited by human fingers and human ears. In 1870, French telegraph engineer Émile Baudot proposed a different approach. Instead of a variable-length code, he defined a fixed 5-bit binary code for each letter of the alphabet — 32 possible combinations, enough for the letters, numbers, and basic punctuation needed in a telegram.

Baudot's encoding was revolutionary for two reasons. First, it could be sent mechanically, without a trained operator reading and tapping. Second, it introduced the concept of a fixed symbol rate — a certain number of code elements transmitted per second. The unit of symbol rate that bears his name, the baud, is still used today. When you read that a modem operates at 3200 baud, you are using Baudot's unit, over a century after he defined it.

By the early 20th century, the teletype machine — essentially a typewriter connected to a Baudot telegraph line — had become standard infrastructure for news agencies, stock exchanges, military communications, and government departments worldwide. The teletype network was, in a meaningful sense, the first wide-area text network: a global system for transmitting written information electrically at machine speed.

Harry Nyquist and the Bandwidth Theorem (1928)

By the 1920s, telephone networks were expanding rapidly, and engineers at Bell Telephone Laboratories were asking a precise question: given a telegraph or telephone channel of a certain bandwidth, what is the maximum number of symbols per second it can carry? The answer came from a Swedish-born engineer named Harry Nyquist.

Nyquist's theorem set the physical framework within which every modem designer since 1928 has worked. When Hayes engineered the Smartmodem, when Rockwell designed the 56k chipset, when Qualcomm built 5G modems — they were all working inside the boundaries Nyquist described.

Claude Shannon and the Information Limit (1948)

Nyquist told engineers how fast they could send symbols. But he left open a deeper question: given that every real channel has noise — static, interference, random fluctuations — how much useful information can actually get through? This question was answered in 1948 by Claude Shannon, a mathematician at Bell Labs, in a paper that founded the entire field of information theory.

Shannon's 1948 paper, A Mathematical Theory of Communication, is one of the most influential scientific publications of the 20th century. It gave engineers not just a formula but a way of thinking: information is a measurable quantity, noise is its enemy, and there is a precise, calculable limit to how much of it any channel can carry. Every modem is, in a sense, an attempt to get as close to the Shannon limit as possible.

SAGE: The First Real Modem (1958)

Through the 1940s, digital computing was emerging from university laboratories and wartime codebreaking projects. Computers were fast; communication networks were not. The two technologies existed in separate worlds. The event that brought them together was the Cold War.

In the early 1950s, the United States Air Force began constructing SAGE — the Semi-Automatic Ground Environment — a continent-wide air defence system designed to track Soviet bombers in real time. SAGE consisted of large AN/FSQ-7 computers at radar stations across North America, each of which needed to share data with the others continuously. The data had to travel over the only long-distance network that existed: the telephone system.

The device built to do this — designed by MIT's Lincoln Laboratory and manufactured by IBM — was a modem in every functional sense: it converted digital data from the computer into audio tones that could travel over a telephone line, and converted incoming tones back into digital data at the other end. It operated at 1300 bits per second and was the size of a large refrigerator.

The SAGE modem was never a commercial product. It was classified, expensive, and built for a single purpose. But it proved the concept: a computer could communicate with another computer over an ordinary telephone line. Everything that followed — the Bell 103, the Hayes Smartmodem, the 56k modem in your drawer — descends from this proof.